Fix spelling of 'Unaffected' in help texts (#838)

Noticed this while setting up pinchflat, figured I'd submit a fix.

.devcontainer/devcontainer.json

@@ -0,0 +1,12 @@
+{
+  "name": "Pinchflat Dev",
+  "dockerComposeFile": "../docker-compose.yml",
+  "service": "phx",
+  "workspaceFolder": "/app",
+  "shutdownAction": "stopCompose",
+  "customizations": {
+    "vscode": {
+      "extensions": ["phoenixframework.phoenix", "JakeBecker.elixir-ls", "esbenp.prettier-vscode"]
+    }
+  }
+}

.formatter.exs

@@ -1,3 +1,5 @@
+# TODO: figure out why my vscode extension doesn't respect the formatter.exs file
+# if it's in a subdirectory
 [
   import_deps: [:ecto, :ecto_sql, :phoenix],
   subdirectories: ["priv/*/migrations"],

.github/ISSUE_TEMPLATE/bug_report.md

@@ -0,0 +1,34 @@
+---
+name: Bug report
+about: Create a report to help us improve
+title: '[Triage] <your title here> '
+labels: triage
+assignees: kieraneglin
+---
+
+**Describe the bug**
+A clear and concise description of what the bug is.
+
+**To Reproduce**
+Steps to reproduce the behavior:
+
+1. Go to '...'
+2. Click on '....'
+3. Scroll down to '....'
+4. See error
+
+**Expected behavior**
+A clear and concise description of what you expected to happen.
+
+**Screenshots**
+If applicable, add screenshots to help explain your problem.
+
+**Diagnostic info**
+
+<!-- Go to Config > App Info > Copy Diagnostic Info and paste that here -->
+
+**Additional context**
+
+<!-- Go to Config > App Info > Download Logs and attach them, if applicable -->
+
+Add any other context about the problem here.

.github/ISSUE_TEMPLATE/feature_request.md

@@ -0,0 +1,19 @@
+---
+name: Feature request
+about: Suggest an idea for this project
+title: '[FR] <your title here>'
+labels: feature request
+assignees: kieraneglin
+---
+
+**Is your feature request related to a problem? Please describe.**
+A clear and concise description of what the problem is. Ex. It's too complicated to [...]
+
+**Describe the solution you'd like**
+A clear and concise description of what you want to happen.
+
+**Describe alternatives you've considered**
+A clear and concise description of any alternative solutions or features you've considered.
+
+**Additional context**
+Add any other context or screenshots about the feature request here.

.github/ISSUE_TEMPLATE/other.md

@@ -0,0 +1,7 @@
+---
+name: Other
+about: For everything else
+title: ''
+labels: ''
+assignees: kieraneglin
+---

.github/pull_request_template.md

@@ -13,3 +13,5 @@ N/A
 ## Any other comments?
 
 N/A
+
+- [ ] I am the original author of this code and I am giving it freely to the community and Pinchflat project maintainers

.github/workflows/build_and_push_docker.yml

@@ -1,32 +0,0 @@
-name: Build and Push Docker Image
-
-on:
-  workflow_dispatch:
-    inputs:
-      image_tag:
-        description: 'Docker Image Tag'
-        required: true
-        default: 'dev'
-
-jobs:
-  docker:
-    runs-on: ubuntu-latest
-    steps:
-      - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3
-
-      - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
-
-      - name: Login to Docker Hub
-        uses: docker/login-action@v3
-        with:
-          username: ${{ secrets.DOCKERHUB_USERNAME }}
-          password: ${{ secrets.DOCKERHUB_TOKEN }}
-
-      - name: Build and push
-        uses: docker/build-push-action@v5
-        with:
-          push: true
-          file: ./selfhosted.Dockerfile
-          tags: 'keglin/pinchflat:${{ github.event.inputs.image_tag }}'

.github/workflows/docker_release.yml

@@ -0,0 +1,86 @@
+name: Docker Release
+
+on:
+  workflow_dispatch:
+    inputs:
+      platforms:
+        type: choice
+        description: 'Build Platforms'
+        required: true
+        default: 'linux/amd64'
+        options:
+          - 'linux/amd64'
+          - 'linux/amd64,linux/arm64'
+      docker_tags:
+        type: string
+        description: 'Docker Tags'
+        required: true
+        default: 'dev'
+
+  push:
+    branches:
+      - master
+  release:
+    types:
+      - published
+
+jobs:
+  build_and_push_docker:
+    runs-on: ubuntu-latest
+    env:
+      dev_arch: 'linux/amd64'
+      release_arch: 'linux/amd64,linux/arm64'
+    permissions:
+      contents: read
+      packages: write
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Docker meta
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: |
+            docker.io/keglin/pinchflat
+            ghcr.io/${{ github.repository }}
+          # All non-release actions will be tagged as `dev` (ie: push, workflow_dispatch)
+          tags: |
+            type=ref,event=tag
+            type=raw,value=dev,enable=${{ github.event_name != 'release' && github.event_name != 'workflow_dispatch' }}
+            type=raw,value=${{ inputs.docker_tags }},enable=${{ github.event_name == 'workflow_dispatch' }}
+          flavor: |
+            latest=auto
+
+      - name: Set up QEMU
+        uses: docker/setup-qemu-action@v3
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Login to Docker Hub
+        uses: docker/login-action@v3
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Login to GitHub Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Build and Push
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: ./docker/selfhosted.Dockerfile
+          # If the event is a release, use the release_arch, otherwise use the
+          # platforms input if present, falling back to dev_arch
+          platforms: ${{ github.event_name == 'release' && env.release_arch || (github.event.inputs.platforms || env.dev_arch) }}
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max

.github/workflows/lint_and_test.yml

@@ -16,24 +16,28 @@ jobs:
     if: "! contains(toJSON(github.event.commits.*.message), '[skip ci]')"
     env:
       COMPOSE_FILE: ./docker-compose.ci.yml
+      MIX_ENV: test
 
     steps:
       - name: Checkout code
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
 
       - name: Pull prebuilt images
         run: docker compose pull
 
-      - name: Setup Docker layer caching
-        uses: jpribyl/action-docker-layer-caching@v0.1.1
-        continue-on-error: true
-        with:
-          key: ci-docker-cache-{hash}
-          restore-keys: |
-            ci-docker-cache-
-            layer-ci-docker-cache-
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
 
-      - name: Build and Run Docker image
+      - name: Build docker image
+        uses: docker/build-push-action@v6
+        with:
+          context: .
+          file: ./docker/dev.Dockerfile
+          load: true
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+
+      - name: Run Docker image
         run: docker compose up --detach
 
       # NOTE: All exec commands use the -T flag to compensate for
@@ -43,8 +47,7 @@ jobs:
       # See https://github.com/actions/runner/issues/241 and https://github.com/docker/compose/issues/8537
       - name: Install Elixir and JS deps
         run: |
-          docker compose exec -T phx yarn install && cd assets && yarn install && cd ..
-          docker compose exec -T phx mix deps.get
+          docker compose exec -T phx mix deps.get && yarn install && cd assets && yarn install && cd ..
 
       - name: Create and Migrate database
         run: |

.iex.exs

@@ -4,8 +4,7 @@ alias Pinchflat.Repo
 alias Pinchflat.Tasks.Task
 alias Pinchflat.Sources.Source
 alias Pinchflat.Media.MediaItem
-alias Pinchflat.Tasks.SourceTasks
-alias Pinchflat.Media.MediaMetadata
+alias Pinchflat.Metadata.MediaMetadata
 alias Pinchflat.Profiles.MediaProfile
 
 alias Pinchflat.Tasks
@@ -14,50 +13,20 @@ alias Pinchflat.Profiles
 alias Pinchflat.Sources
 alias Pinchflat.Settings
 
-alias Pinchflat.MediaClient.MediaDownloader
-alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
-alias Pinchflat.YtDlp.Backend.MediaCollection, as: YtDlpCollection
+alias Pinchflat.Downloading.MediaDownloader
+alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+alias Pinchflat.YtDlp.MediaCollection, as: YtDlpCollection
 
-alias Pinchflat.Api.YoutubeRss
+alias Pinchflat.FastIndexing.YoutubeRss
 alias Pinchflat.Metadata.MetadataFileHelpers
 
-alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer
+alias Pinchflat.SlowIndexing.FileFollowerServer
+
+Pinchflat.Release.check_file_permissions()
 
 defmodule IexHelpers do
-  def playlist_url do
-    "https://www.youtube.com/playlist?list=PLmqC3wPkeL8kSlTCcSMDD63gmSi7evcXS"
-  end
-
-  def channel_url do
-    "https://www.youtube.com/c/TheUselessTrials"
-  end
-
-  def video_url do
-    "https://www.youtube.com/watch?v=bR52O78ZIUw"
-  end
-
-  def last_media_item do
-    Repo.one(from m in MediaItem, limit: 1)
-  end
-
-  def details(type) do
-    source =
-      case type do
-        :playlist -> playlist_url()
-        :channel -> channel_url()
-      end
-
-    YtDlpCollection.get_source_details(source)
-  end
-
-  def ids(type) do
-    source =
-      case type do
-        :playlist -> playlist_url()
-        :channel -> channel_url()
-      end
-
-    YtDlpCollection.get_media_attributes_for_collection(source)
+  def restart do
+    :init.restart()
   end
 end
 

.prettierignore

@@ -1,3 +1 @@
 assets/vendor/
-deps/
-_build/

.sobelow-conf

@@ -9,7 +9,16 @@
   threshold: :low,
   # All of these are ignorable because this app is intended to be single-user and self-hosted.
   # There is an expectation that the user won't intentionally run a FS Traversal on themselves
-  ignore: ["CI.System", "Traversal.FileModule", "Config.HTTPS", "Config.CSP"],
+  ignore:
+    [
+      "CI.System",
+      "Traversal.FileModule",
+      "Config.HTTPS",
+      "Config.CSP",
+      "XSS.ContentType",
+      "Traversal.SendFile",
+      "Traversal.SendDownload"
+    ],
   ignore_files: [],
   version: false
 ]

README.md

@@ -1,39 +1,203 @@
-# Pinchflat (Alpha)
+> [!IMPORTANT]  
+> (2025-02-14) [zakkarry](https://github.com/sponsors/zakkarry), who is a collaborator on [cross-seed](https://github.com/cross-seed/cross-seed) and an extremely helpful community member in general, is facing hard times due to medical debt and family illness. If you're able, please consider [sponsoring him on GitHub](https://github.com/sponsors/zakkarry) or donating via [buymeacoffee](https://tip.ary.dev). Tell him I sent you!
 
-This is alpha software and anything can break at any time. I make not guarantees about the stability of this software, forward-compatibility of updates, or integrity (both related to and independent of Pinchflat). Essentially, use at your own risk and expect there will be rough edges.
+<p align="center">  
+  <img 
+    src="priv/static/images/originals/logo-white-wordmark-with-background.png" 
+    alt="Pinchflat Logo by @hernandito"
+    width="700" 
+  />
+</p>
 
-## What is Pinchflat?
+<p align="center">  
+  <sup>
+    <em>logo by <a href="https://github.com/hernandito" target="_blank">@hernandito</a></em>
+  </sup>
+</p>
 
-TODO: expand on this.
+<div align="center">
 
-Pinchflat is a lightweight self-contained app for downloading YouTube content. For now, it is not intended as a way to consume content, but instead as a way to download content to disk using specified rules and schedules.
+[![](https://img.shields.io/github/license/kieraneglin/pinchflat?style=for-the-badge&color=ee512b)](LICENSE)
+[![](https://img.shields.io/github/v/release/kieraneglin/pinchflat?style=for-the-badge&color=purple)](https://github.com/kieraneglin/pinchflat/releases)
+[![](https://img.shields.io/static/v1?style=for-the-badge&logo=discord&message=Chat&color=5865F2&label=Discord)](https://discord.gg/j7T6dCuwU4)
+[![](https://img.shields.io/github/actions/workflow/status/kieraneglin/pinchflat/lint_and_test.yml?style=for-the-badge)](#)
+[![](https://img.shields.io/static/v1?label=Dev%20Containers&message=Open&color=blue&logo=visualstudiocode&style=for-the-badge)](https://vscode.dev/redirect?url=vscode://ms-vscode-remote.remote-containers/cloneInVolume?url=https://github.com/kieraneglin/pinchflat)
 
-I have plans for more to come, but for now this is the focus. Think of Pinchflat as nothing more than an automated way to get content from YouTube to your disk.
+</div>
+
+# Your next YouTube media manager
+
+## Table of contents:
+
+- [What it does](#what-it-does)
+- [Features](#features)
+- [Screenshots](#screenshots)
+- [Installation](#installation)
+  - [Unraid](#unraid)
+  - [Portainer](#portainer)
+  - [Docker](#docker)
+  - [Environment Variables](#environment-variables)
+  - [A note on reverse proxies](#reverse-proxies)
+- [Username and Password (authentication)](https://github.com/kieraneglin/pinchflat/wiki/Username-and-Password)
+- [Frequently asked questions](https://github.com/kieraneglin/pinchflat/wiki/Frequently-Asked-Questions)
+- [Documentation](https://github.com/kieraneglin/pinchflat/wiki)
+- [EFF donations](#eff-donations)
+- [Pre-release disclaimer](#pre-release-disclaimer)
+- [Development and Contributing](https://github.com/kieraneglin/pinchflat/wiki/Development-and-Contributing)
+
+## What it does
+
+Pinchflat is a self-hosted app for downloading YouTube content built using [yt-dlp](https://github.com/yt-dlp/yt-dlp). It's designed to be lightweight, self-contained, and easy to use. You set up rules for how to download content from YouTube channels or playlists and it'll do the rest, periodically checking for new content. It's perfect for people who want to download content for use in with a media center app (Plex, Jellyfin, Kodi) or for those who want to archive media!
+
+While you can [download individual videos](https://github.com/kieraneglin/pinchflat/wiki/Frequently-Asked-Questions#how-do-i-download-one-off-videos), Pinchflat is best suited for downloading content from channels or playlists. It's also not meant for consuming content in-app - Pinchflat downloads content to disk where you can then watch it with a media center app or VLC.
+
+If it doesn't work for your use case, please make a feature request! You can also check out these great alternatives: [Tube Archivist](https://github.com/tubearchivist/tubearchivist), [ytdl-sub](https://github.com/jmbannon/ytdl-sub), and [TubeSync](https://github.com/meeb/tubesync)
+
+## Features
+
+- Self-contained - just one Docker container with no external dependencies
+- Powerful naming system so content is stored where and how you want it
+- Easy-to-use web interface with presets to get you started right away
+- First-class support for media center apps like Plex, Jellyfin, and Kodi ([docs](https://github.com/kieraneglin/pinchflat/wiki/Frequently-Asked-Questions#how-do-i-get-media-into-plexjellyfinkodi))
+- Supports serving RSS feeds to your favourite podcast app ([docs](https://github.com/kieraneglin/pinchflat/wiki/Podcast-RSS-Feeds))
+- Automatically downloads new content from channels and playlists
+  - Uses a novel approach to download new content more quickly than other apps
+- Supports downloading audio content
+- Custom rules for handling YouTube Shorts and livestreams
+- Apprise support for notifications
+- Allows automatically redownloading new media after a set period
+  - This can help improve the download quality of new content or improve SponsorBlock tags
+- Optionally automatically delete old content ([docs](https://github.com/kieraneglin/pinchflat/wiki/Automatically-Delete-Media))
+- Advanced options like setting cutoff dates and filtering by title ([docs](https://github.com/kieraneglin/pinchflat/wiki/Frequently-Asked-Questions#i-only-want-certain-videos-from-a-source---how-can-i-only-download-those))
+- Reliable hands-off operation
+- Can pass cookies to YouTube to download your private playlists ([docs](https://github.com/kieraneglin/pinchflat/wiki/YouTube-Cookies))
+- Sponsorblock integration
+- \[Advanced\] allows custom `yt-dlp` options ([docs](https://github.com/kieraneglin/pinchflat/wiki/%5BAdvanced%5D-Custom-yt%E2%80%90dlp-options))
+- \[Advanced\] supports running custom scripts when after downloading/deleting media (alpha - [docs](https://github.com/kieraneglin/pinchflat/wiki/%5BAdvanced%5D-Custom-lifecycle-scripts))
+
+## Screenshots
+
+<img src="priv/static/images/app-form-screenshot.jpg" alt="Pinchflat screenshot" width="700" />
+<img src="priv/static/images/app-screenshot.jpg" alt="Pinchflat screenshot" width="700" />
 
 ## Installation
 
-Pinchflat is designed to be self-hosted. I'm building it for my own needs which means it's designed to work well with Unraid, but it should work on any computer/server that can run Docker images.
+### Unraid
 
-I'll update with Unraid instructions once I get something in the Community Apps store. Until then, here's how you build it with Docker:
+Simply search for Pinchflat in the Community Apps store!
 
-```bash
-docker build . --file selfhosted.Dockerfile -t pinchflat:dev
+### Portainer
 
-docker run \
-  -p 8945:8945 \
-  -v /Users/work/Desktop/test_volumes/config:/config \
-  -v /Users/work/Desktop/test_volumes/downloads:/downloads \
-  pinchflat:dev
+> [!IMPORTANT]  
+> See the note below about storing config on a network file share. It's preferred to store the config on a local disk if at all possible.
+
+Docker Compose file:
+
+```yaml
+version: '3'
+services:
+  pinchflat:
+    image: ghcr.io/kieraneglin/pinchflat:latest
+    environment:
+      # Set the timezone to your local timezone
+      - TZ=America/New_York
+    ports:
+      - '8945:8945'
+    volumes:
+      - /host/path/to/config:/config
+      - /host/path/to/downloads:/downloads
 ```
 
-## Authentication
+### Docker
 
-Currently HTTP basic auth is optionally supported. To use it, set the `BASIC_AUTH_USERNAME` and `BASIC_AUTH_PASSWORD` environment variables when starting the container. If you don't set both of these, no authentication will be required.
+1. Create two directories on your host machine: one for storing config and one for storing downloaded media. Make sure they're both writable by the user running the Docker container.
+2. Prepare the docker image in one of the two ways below:
+   - **From GHCR:** `docker pull ghcr.io/kieraneglin/pinchflat:latest`
+     - NOTE: also available on Docker Hub at `keglin/pinchflat:latest`
+   - **Building locally:** `docker build . --file docker/selfhosted.Dockerfile -t ghcr.io/kieraneglin/pinchflat:latest`
+3. Run the container:
+
+```bash
+# Be sure to replace /host/path/to/config and /host/path/to/downloads below with
+# the paths to the directories you created in step 1
+# Be sure to replace America/New_York with your local timezone
+docker run \
+  -e TZ=America/New_York \
+  -p 8945:8945 \
+  -v /host/path/to/config:/config \
+  -v /host/path/to/downloads:/downloads \
+  ghcr.io/kieraneglin/pinchflat:latest
+```
+
+### Podman
+
+The Podman setup is similar to Docker but changes a few flags to run under a User Namespace instead of root. To run Pinchflat under Podman and use the current user's UID/GID for file access run this:
+
+```
+podman run \
+  --security-opt label=disable \
+  --userns=keep-id --user=$UID \
+  -e TZ=America/Los_Angeles \
+  -p 8945:8945 \
+  -v /host/path/to/config:/config:rw \
+  -v /host/path/to/downloads/:/downloads:rw \
+  ghcr.io/kieraneglin/pinchflat:latest
+```
+
+Using this setup consider creating a new `pinchflat` user and giving that user ownership to the config and download directory. See [Podman --userns](https://docs.podman.io/en/v4.6.1/markdown/options/userns.container.html) docs.
+
+### IMPORTANT: File permissions
+
+You _must_ ensure the host directories you've mounted are writable by the user running the Docker container. If you get a permission error follow the steps it suggests. See [#106](https://github.com/kieraneglin/pinchflat/issues/106) for more.
+
+> [!IMPORTANT]
+> It's not recommended to run the container as root. Doing so can create permission issues if other apps need to work with the downloaded media.
+
+### ADVANCED: Storing Pinchflat config directory on a network share
+
+As pointed out in [#137](https://github.com/kieraneglin/pinchflat/issues/137), SQLite doesn't like being run in WAL mode on network shares. If you're running Pinchflat on a network share, you can disable WAL mode by setting the `JOURNAL_MODE` environment variable to `delete`. This will make Pinchflat run in rollback journal mode which is less performant but should work on network shares.
+
+> [!CAUTION]
+> Changing this setting from WAL to `delete` on an existing Pinchflat instance could, conceivably, result in data loss. Only change this setting if you know what you're doing, why this is important, and are okay with possible data loss or DB corruption. Backup your database first!
+
+If you change this setting and it works well for you, please leave a comment on [#137](https://github.com/kieraneglin/pinchflat/issues/137)! Doubly so if it does _not_ work well.
+
+### Environment variables
+
+| Name                        | Required? | Default                   | Notes                                                                                                                                     |
+| --------------------------- | --------- | ------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- |
+| `TZ`                        | No        | `UTC`                     | Must follow IANA TZ format                                                                                                                |
+| `LOG_LEVEL`                 | No        | `debug`                   | Can be set to `info` but `debug` is strongly recommended                                                                                  |
+| `UMASK`                     | No        | `022`                     | Unraid users may want to set this to `000`                                                                                                |
+| `BASIC_AUTH_USERNAME`       | No        |                           | See [authentication docs](https://github.com/kieraneglin/pinchflat/wiki/Username-and-Password)                                            |
+| `BASIC_AUTH_PASSWORD`       | No        |                           | See [authentication docs](https://github.com/kieraneglin/pinchflat/wiki/Username-and-Password)                                            |
+| `EXPOSE_FEED_ENDPOINTS`     | No        | `false`                   | See [RSS feed docs](https://github.com/kieraneglin/pinchflat/wiki/Podcast-RSS-Feeds)                                                      |
+| `ENABLE_IPV6`               | No        | `false`                   | Setting to _any_ non-blank value will enable IPv6                                                                                         |
+| `JOURNAL_MODE`              | No        | `wal`                     | Set to `delete` if your config directory is stored on a network share (not recommended)                                                   |
+| `TZ_DATA_DIR`               | No        | `/etc/elixir_tzdata_data` | The container path where the timezone database is stored                                                                                  |
+| `BASE_ROUTE_PATH`           | No        | `/`                       | The base path for route generation. Useful when running behind certain reverse proxies - prefixes must be stripped.                       |
+| `YT_DLP_WORKER_CONCURRENCY` | No        | `2`                       | The number of concurrent workers that use `yt-dlp` _per queue_. Set to 1 if you're getting IP limited, otherwise don't touch it           |
+| `ENABLE_PROMETHEUS`         | No        | `false`                   | Setting to _any_ non-blank value will enable Prometheus. See [docs](https://github.com/kieraneglin/pinchflat/wiki/Prometheus-and-Grafana) |
+
+### Reverse Proxies
+
+Pinchflat makes heavy use of websockets for real-time updates. If you're running Pinchflat behind a reverse proxy then you'll need to make sure it's configured to support websockets.
+
+## EFF donations
+
+Prior to 2024-05-10, a portion of all donations were given to the [Electronic Frontier Foundation](https://www.eff.org/). Now, the app doesn't accept donations that go to me personally and instead directs you straight to the EFF. [Here](https://github.com/kieraneglin/pinchflat/issues/234) are some people that have generously donated.
+
+The EFF defends your online liberties and [backed](https://github.com/github/dmca/blob/9a85e0f021f7967af80e186b890776a50443f06c/2020/11/2020-11-16-RIAA-reversal-effletter.pdf) `youtube-dl` when Google took them down.
+
+## Stability disclaimer
+
+This software is in active development and anything can break at any time. I make no guarantees about the stability of this software, forward-compatibility of updates, or integrity (both related to and independent of Pinchflat).
 
 ## License
 
 See `LICENSE` file
 
-```
+<!-- Images and links -->
 
-```
+[license-badge]: https://img.shields.io/github/license/kieraneglin/pinchflat?style=for-the-badge&color=ee512b
+[license-badge-url]: LICENSE

assets/css/satoshi.css

@@ -22,8 +22,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-Light.woff2'), url('/fonts/satoshi/Satoshi-Light.woff'),
-    url('/fonts/satoshi/Satoshi-Light.ttf');
+  src: url('../fonts/satoshi/Satoshi-Light.woff2'), url('../fonts/satoshi/Satoshi-Light.woff'),
+    url('../fonts/satoshi/Satoshi-Light.ttf');
   font-weight: 300;
   font-display: swap;
   font-style: normal;
@@ -31,8 +31,9 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-LightItalic.woff2'),
-    url('/fonts/satoshi/Satoshi-LightItalic.woff'), url('/fonts/satoshi/Satoshi-LightItalic.ttf');
+  src: url('../fonts/satoshi/Satoshi-LightItalic.woff2'),
+    url('../fonts/satoshi/Satoshi-LightItalic.woff'),
+    url('../fonts/satoshi/Satoshi-LightItalic.ttf');
   font-weight: 300;
   font-display: swap;
   font-style: italic;
@@ -40,8 +41,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-Regular.woff2'), url('/fonts/satoshi/Satoshi-Regular.woff'),
-    url('/fonts/satoshi/Satoshi-Regular.ttf');
+  src: url('../fonts/satoshi/Satoshi-Regular.woff2'), url('../fonts/satoshi/Satoshi-Regular.woff'),
+    url('../fonts/satoshi/Satoshi-Regular.ttf');
   font-weight: 400;
   font-display: swap;
   font-style: normal;
@@ -49,8 +50,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-Italic.woff2'), url('/fonts/satoshi/Satoshi-Italic.woff'),
-    url('/fonts/satoshi/Satoshi-Italic.ttf');
+  src: url('../fonts/satoshi/Satoshi-Italic.woff2'), url('../fonts/satoshi/Satoshi-Italic.woff'),
+    url('../fonts/satoshi/Satoshi-Italic.ttf');
   font-weight: 400;
   font-display: swap;
   font-style: italic;
@@ -58,8 +59,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-Medium.woff2'), url('/fonts/satoshi/Satoshi-Medium.woff'),
-    url('/fonts/satoshi/Satoshi-Medium.ttf');
+  src: url('../fonts/satoshi/Satoshi-Medium.woff2'), url('../fonts/satoshi/Satoshi-Medium.woff'),
+    url('../fonts/satoshi/Satoshi-Medium.ttf');
   font-weight: 500;
   font-display: swap;
   font-style: normal;
@@ -67,8 +68,9 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-MediumItalic.woff2'),
-    url('/fonts/satoshi/Satoshi-MediumItalic.woff'), url('/fonts/satoshi/Satoshi-MediumItalic.ttf');
+  src: url('../fonts/satoshi/Satoshi-MediumItalic.woff2'),
+    url('../fonts/satoshi/Satoshi-MediumItalic.woff'),
+    url('../fonts/satoshi/Satoshi-MediumItalic.ttf');
   font-weight: 500;
   font-display: swap;
   font-style: italic;
@@ -76,8 +78,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-Bold.woff2'), url('/fonts/satoshi/Satoshi-Bold.woff'),
-    url('/fonts/satoshi/Satoshi-Bold.ttf');
+  src: url('../fonts/satoshi/Satoshi-Bold.woff2'), url('../fonts/satoshi/Satoshi-Bold.woff'),
+    url('../fonts/satoshi/Satoshi-Bold.ttf');
   font-weight: 700;
   font-display: swap;
   font-style: normal;
@@ -85,8 +87,8 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-BoldItalic.woff2'), url('/fonts/satoshi/Satoshi-BoldItalic.woff'),
-    url('/fonts/satoshi/Satoshi-BoldItalic.ttf');
+  src: url('../fonts/satoshi/Satoshi-BoldItalic.woff2'),
+    url('../fonts/satoshi/Satoshi-BoldItalic.woff'), url('../fonts/satoshi/Satoshi-BoldItalic.ttf');
   font-weight: 700;
   font-display: swap;
   font-style: italic;
@@ -94,8 +96,9 @@
 
 @font-face {
   font-family: 'Satoshi';
-  src: url('/fonts/satoshi/Satoshi-BlackItalic.woff2'),
-    url('/fonts/satoshi/Satoshi-BlackItalic.woff'), url('/fonts/satoshi/Satoshi-BlackItalic.ttf');
+  src: url('../fonts/satoshi/Satoshi-BlackItalic.woff2'),
+    url('../fonts/satoshi/Satoshi-BlackItalic.woff'),
+    url('../fonts/satoshi/Satoshi-BlackItalic.ttf');
   font-weight: 900;
   font-display: swap;
   font-style: italic;

assets/js/alpine_helpers.js

@@ -0,0 +1,48 @@
+window.copyTextToClipboard = async (text) => {
+  // Navigator clipboard api needs a secure context (https)
+  if (navigator.clipboard && window.isSecureContext) {
+    await navigator.clipboard.writeText(text)
+  } else {
+    const textArea = document.createElement('textarea')
+    textArea.value = text
+    // Move textarea out of the viewport so it's not visible
+    textArea.style.position = 'absolute'
+    textArea.style.left = '-999999px'
+
+    document.body.prepend(textArea)
+    textArea.select()
+
+    try {
+      document.execCommand('copy')
+    } catch (error) {
+      console.error(error)
+    } finally {
+      textArea.remove()
+    }
+  }
+}
+
+window.copyWithCallbacks = async (text, onCopy, onAfterDelay, delay = 4000) => {
+  await window.copyTextToClipboard(text)
+  onCopy()
+  setTimeout(onAfterDelay, delay)
+}
+
+window.markVersionAsSeen = (versionString) => {
+  localStorage.setItem('seenVersion', versionString)
+}
+
+window.isVersionSeen = (versionString) => {
+  return localStorage.getItem('seenVersion') === versionString
+}
+
+window.dispatchFor = (elementOrId, eventName, detail = {}) => {
+  const element =
+    typeof elementOrId === 'string' ? document.getElementById(elementOrId) : elementOrId
+
+  // This is needed to ensure the DOM has updated before dispatching the event.
+  // Doing so ensures that the latest DOM state is what's sent to the server
+  setTimeout(() => {
+    element.dispatchEvent(new Event(eventName, { bubbles: true, detail }))
+  }, 0)
+}

assets/js/app.js

@@ -22,12 +22,14 @@ import { Socket } from 'phoenix'
 import { LiveSocket } from 'phoenix_live_view'
 import topbar from '../vendor/topbar'
 import Alpine from 'alpinejs'
+import './tabs'
+import './alpine_helpers'
 
 window.Alpine = Alpine
 Alpine.start()
 
 let csrfToken = document.querySelector("meta[name='csrf-token']").getAttribute('content')
-let liveSocket = new LiveSocket('/live', Socket, {
+let liveSocket = new LiveSocket(document.body.dataset.socketPath, Socket, {
   params: { _csrf_token: csrfToken },
   dom: {
     onBeforeElUpdated(from, to) {
@@ -35,6 +37,17 @@ let liveSocket = new LiveSocket('/live', Socket, {
         window.Alpine.clone(from, to)
       }
     }
+  },
+  hooks: {
+    'supress-enter-submission': {
+      mounted() {
+        this.el.addEventListener('keypress', (event) => {
+          if (event.key === 'Enter') {
+            event.preventDefault()
+          }
+        })
+      }
+    }
   }
 })
 

assets/js/tabs.js

@@ -0,0 +1,20 @@
+window.setTabByName = (tabName) => {
+  window.location.hash = `tab-${tabName}`
+
+  return tabName
+}
+
+// The conditionals and currIndex stuff ensures that
+// the tab index is always set to 0 if the hash is empty
+// AND other hash values are ignored
+window.getTabFromHash = (currentTabName, defaultTabName) => {
+  if (window.location.hash === '' || window.location.hash === '#') {
+    return defaultTabName
+  }
+
+  if (window.location.hash.startsWith('#tab-')) {
+    return window.location.hash.replace('#tab-', '')
+  }
+
+  return currentTabName
+}

assets/tailwind.config.js

@@ -347,6 +347,38 @@ module.exports = {
         },
         { values }
       )
+    }),
+    plugin(function ({ matchComponents, theme }) {
+      let iconsDir = path.join(__dirname, './vendor/simple-icons')
+      let values = {}
+
+      fs.readdirSync(iconsDir).forEach((file) => {
+        let name = path.basename(file, '.svg')
+        values[name] = { name, fullPath: path.join(iconsDir, file) }
+      })
+
+      matchComponents(
+        {
+          si: ({ name, fullPath }) => {
+            let content = fs
+              .readFileSync(fullPath)
+              .toString()
+              .replace(/\r?\n|\r/g, '')
+            return {
+              [`--si-${name}`]: `url('data:image/svg+xml;utf8,${content}')`,
+              '-webkit-mask': `var(--si-${name})`,
+              mask: `var(--si-${name})`,
+              'mask-repeat': 'no-repeat',
+              'background-color': 'currentColor',
+              'vertical-align': 'middle',
+              display: 'inline-block',
+              width: theme('spacing.5'),
+              height: theme('spacing.5')
+            }
+          }
+        },
+        { values }
+      )
     })
   ]
 }

assets/vendor/simple-icons/discord.svg

@@ -0,0 +1 @@
+<svg role="img" viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg"><title>Discord</title><path d="M20.317 4.3698a19.7913 19.7913 0 00-4.8851-1.5152.0741.0741 0 00-.0785.0371c-.211.3753-.4447.8648-.6083 1.2495-1.8447-.2762-3.68-.2762-5.4868 0-.1636-.3933-.4058-.8742-.6177-1.2495a.077.077 0 00-.0785-.037 19.7363 19.7363 0 00-4.8852 1.515.0699.0699 0 00-.0321.0277C.5334 9.0458-.319 13.5799.0992 18.0578a.0824.0824 0 00.0312.0561c2.0528 1.5076 4.0413 2.4228 5.9929 3.0294a.0777.0777 0 00.0842-.0276c.4616-.6304.8731-1.2952 1.226-1.9942a.076.076 0 00-.0416-.1057c-.6528-.2476-1.2743-.5495-1.8722-.8923a.077.077 0 01-.0076-.1277c.1258-.0943.2517-.1923.3718-.2914a.0743.0743 0 01.0776-.0105c3.9278 1.7933 8.18 1.7933 12.0614 0a.0739.0739 0 01.0785.0095c.1202.099.246.1981.3728.2924a.077.077 0 01-.0066.1276 12.2986 12.2986 0 01-1.873.8914.0766.0766 0 00-.0407.1067c.3604.698.7719 1.3628 1.225 1.9932a.076.076 0 00.0842.0286c1.961-.6067 3.9495-1.5219 6.0023-3.0294a.077.077 0 00.0313-.0552c.5004-5.177-.8382-9.6739-3.5485-13.6604a.061.061 0 00-.0312-.0286zM8.02 15.3312c-1.1825 0-2.1569-1.0857-2.1569-2.419 0-1.3332.9555-2.4189 2.157-2.4189 1.2108 0 2.1757 1.0952 2.1568 2.419 0 1.3332-.9555 2.4189-2.1569 2.4189zm7.9748 0c-1.1825 0-2.1569-1.0857-2.1569-2.419 0-1.3332.9554-2.4189 2.1569-2.4189 1.2108 0 2.1757 1.0952 2.1568 2.419 0 1.3332-.946 2.4189-2.1568 2.4189Z"/></svg>

assets/vendor/simple-icons/github.svg

@@ -0,0 +1 @@
+<svg role="img" viewBox="0 0 24 24" xmlns="http://www.w3.org/2000/svg"><title>GitHub</title><path d="M12 .297c-6.63 0-12 5.373-12 12 0 5.303 3.438 9.8 8.205 11.385.6.113.82-.258.82-.577 0-.285-.01-1.04-.015-2.04-3.338.724-4.042-1.61-4.042-1.61C4.422 18.07 3.633 17.7 3.633 17.7c-1.087-.744.084-.729.084-.729 1.205.084 1.838 1.236 1.838 1.236 1.07 1.835 2.809 1.305 3.495.998.108-.776.417-1.305.76-1.605-2.665-.3-5.466-1.332-5.466-5.93 0-1.31.465-2.38 1.235-3.22-.135-.303-.54-1.523.105-3.176 0 0 1.005-.322 3.3 1.23.96-.267 1.98-.399 3-.405 1.02.006 2.04.138 3 .405 2.28-1.552 3.285-1.23 3.285-1.23.645 1.653.24 2.873.12 3.176.765.84 1.23 1.91 1.23 3.22 0 4.61-2.805 5.625-5.475 5.92.42.36.81 1.096.81 2.22 0 1.606-.015 2.896-.015 3.286 0 .315.21.69.825.57C20.565 22.092 24 17.592 24 12.297c0-6.627-5.373-12-12-12"/></svg>

config/config.exs

@@ -10,18 +10,29 @@ import Config
 config :pinchflat,
   ecto_repos: [Pinchflat.Repo],
   generators: [timestamp_type: :utc_datetime],
+  env: config_env(),
   # Specifying backend data here makes mocking and local testing SUPER easy
   yt_dlp_executable: System.find_executable("yt-dlp"),
-  yt_dlp_runner: Pinchflat.YtDlp.Backend.CommandRunner,
+  apprise_executable: System.find_executable("apprise"),
+  yt_dlp_runner: Pinchflat.YtDlp.CommandRunner,
+  apprise_runner: Pinchflat.Lifecycle.Notifications.CommandRunner,
   media_directory: "/downloads",
   # The user may or may not store metadata for their needs, but the app will always store its copy
   metadata_directory: "/config/metadata",
+  extras_directory: "/config/extras",
   tmpfile_directory: Path.join([System.tmp_dir!(), "pinchflat", "data"]),
   # Setting BASIC_AUTH_USERNAME and BASIC_AUTH_PASSWORD implies you want to use basic auth.
   # If either is unset, basic auth will not be used.
-  basic_auth_username: System.get_env("BASIC_AUTH_USERNAME"),
-  basic_auth_password: System.get_env("BASIC_AUTH_PASSWORD"),
-  file_watcher_poll_interval: 1000
+  basic_auth_username: "",
+  basic_auth_password: "",
+  expose_feed_endpoints: false,
+  file_watcher_poll_interval: 1000,
+  timezone: "UTC",
+  base_route_path: "/"
+
+config :pinchflat, Pinchflat.Repo,
+  journal_mode: :wal,
+  pool_size: 5
 
 # Configures the endpoint
 config :pinchflat, PinchflatWeb.Endpoint,
@@ -31,25 +42,15 @@ config :pinchflat, PinchflatWeb.Endpoint,
   adapter: Phoenix.Endpoint.Cowboy2Adapter,
   render_errors: [
     formats: [html: PinchflatWeb.ErrorHTML, json: PinchflatWeb.ErrorJSON],
-    layout: false
+    root_layout: {PinchflatWeb.Layouts, :root},
+    layout: {PinchflatWeb.Layouts, :app}
   ],
   pubsub_server: Pinchflat.PubSub,
   live_view: [signing_salt: "/t5878kO"]
 
 config :pinchflat, Oban,
   engine: Oban.Engines.Lite,
-  repo: Pinchflat.Repo,
-  # Keep old jobs for 30 days for display in the UI
-  plugins: [{Oban.Plugins.Pruner, max_age: 30 * 24 * 60 * 60}],
-  # TODO: consider making this an env var or something?
-  queues: [
-    default: 10,
-    fast_indexing: 6,
-    media_indexing: 2,
-    media_collection_indexing: 2,
-    media_fetching: 2,
-    media_local_metadata: 8
-  ]
+  repo: Pinchflat.Repo
 
 # Configures the mailer
 #
@@ -72,7 +73,7 @@ config :esbuild,
 
 # Configure tailwind (the version is required)
 config :tailwind,
-  version: "3.3.2",
+  version: "3.4.3",
   default: [
     args: ~w(
       --config=tailwind.config.js
@@ -83,13 +84,19 @@ config :tailwind,
   ]
 
 # Configures Elixir's Logger
-config :logger, :console,
-  format: "$time $metadata[$level] $message\n",
+config :logger, :default_formatter,
+  format: "$date $time $metadata[$level] | $message\n",
   metadata: [:request_id]
 
 # Use Jason for JSON parsing in Phoenix
 config :phoenix, :json_library, Jason
 
+config :pinchflat, Pinchflat.PromEx,
+  disabled: true,
+  manual_metrics_start_delay: :no_delay,
+  drop_metrics_groups: [],
+  metrics_server: :disabled
+
 # Import environment specific config. This must remain at the bottom
 # of this file so it overrides the configuration defined above.
 import_config "#{config_env()}.exs"

config/dev.exs

@@ -3,6 +3,7 @@ import Config
 config :pinchflat,
   media_directory: Path.join([File.cwd!(), "tmp", "media"]),
   metadata_directory: Path.join([File.cwd!(), "tmp", "metadata"]),
+  extras_directory: Path.join([File.cwd!(), "tmp", "extras"]),
   tmpfile_directory: Path.join([File.cwd!(), "tmp", "tmpfiles"])
 
 # Configure your database
@@ -66,7 +67,7 @@ config :pinchflat, PinchflatWeb.Endpoint,
 config :pinchflat, dev_routes: true
 
 # Do not include metadata nor timestamps in development logs
-config :logger, :console, format: "[$level] $message\n"
+config :logger, :default_formatter, format: "[$level] $message\n"
 
 # Set a higher stacktrace during development. Avoid configuring such
 # in production as building large stacktraces may be expensive.
@@ -80,3 +81,5 @@ config :phoenix_live_view, :debug_heex_annotations, true
 
 # Disable swoosh api client as it is only required for production adapters.
 config :swoosh, :api_client, false
+
+config :pinchflat, Pinchflat.PromEx, disabled: false

config/runtime.exs

@@ -25,37 +25,93 @@ config :pinchflat,
   basic_auth_username: System.get_env("BASIC_AUTH_USERNAME"),
   basic_auth_password: System.get_env("BASIC_AUTH_PASSWORD")
 
-if config_env() == :prod do
-  config_path =
-    System.get_env("CONFIG_PATH") ||
-      raise """
-      environment variable CONFIG_PATH is missing.
-      For example: /etc/pinchflat/config
-      """
+arch_string = to_string(:erlang.system_info(:system_architecture))
 
+system_arch =
+  cond do
+    String.contains?(arch_string, "arm") -> "arm"
+    String.contains?(arch_string, "aarch") -> "arm"
+    String.contains?(arch_string, "x86") -> "x86"
+    true -> "unknown"
+  end
+
+config :pinchflat, Pinchflat.Repo,
+  load_extensions: [
+    Path.join([:code.priv_dir(:pinchflat), "repo", "extensions", "sqlean-linux-#{system_arch}", "sqlean"])
+  ]
+
+# Some users may want to increase the number of workers that use yt-dlp to improve speeds
+# Others may want to decrease the number of these workers to lessen the chance of an IP ban
+{yt_dlp_worker_count, _} = Integer.parse(System.get_env("YT_DLP_WORKER_CONCURRENCY", "2"))
+# Used to set the cron for the yt-dlp update worker. The reason for this is
+# to avoid all instances of PF updating yt-dlp at the same time, which 1)
+# could result in rate limiting and 2) gives me time to react if an update
+# breaks something
+%{hour: current_hour, minute: current_minute} = DateTime.utc_now()
+
+config :pinchflat, Oban,
+  queues: [
+    default: 10,
+    fast_indexing: yt_dlp_worker_count,
+    media_collection_indexing: yt_dlp_worker_count,
+    media_fetching: yt_dlp_worker_count,
+    remote_metadata: yt_dlp_worker_count,
+    local_data: 8
+  ],
+  plugins: [
+    # Keep old jobs for 30 days for display in the UI
+    {Oban.Plugins.Pruner, max_age: 30 * 24 * 60 * 60},
+    {Oban.Plugins.Cron,
+     crontab: [
+       {"#{current_minute} #{current_hour} * * *", Pinchflat.YtDlp.UpdateWorker},
+       {"0 1 * * *", Pinchflat.Downloading.MediaRetentionWorker},
+       {"0 2 * * *", Pinchflat.Downloading.MediaQualityUpgradeWorker}
+     ]}
+  ]
+
+if config_env() == :prod do
+  # Various paths. These ones shouldn't be tweaked if running in Docker
+  media_path = System.get_env("MEDIA_PATH", "/downloads")
+  config_path = System.get_env("CONFIG_PATH", "/config")
   db_path = System.get_env("DATABASE_PATH", Path.join([config_path, "db", "pinchflat.db"]))
   log_path = System.get_env("LOG_PATH", Path.join([config_path, "logs", "pinchflat.log"]))
   metadata_path = System.get_env("METADATA_PATH", Path.join([config_path, "metadata"]))
+  extras_path = System.get_env("EXTRAS_PATH", Path.join([config_path, "extras"]))
+  tmpfile_path = System.get_env("TMPFILE_PATH", Path.join([System.tmp_dir!(), "pinchflat", "data"]))
+  # This one can be changed if you want
+  tz_data_path = System.get_env("TZ_DATA_PATH", Path.join([extras_path, "elixir_tz_data"]))
+  # For running PF as a podcast host on self-hosted environments
+  expose_feed_endpoints = String.length(System.get_env("EXPOSE_FEED_ENDPOINTS", "")) > 0
+  # For testing alternate journal modes (see issue #137)
+  journal_mode = String.to_existing_atom(System.get_env("JOURNAL_MODE", "wal"))
+  # For running PF in a subdirectory via a reverse proxy
+  base_route_path = System.get_env("BASE_ROUTE_PATH", "/")
+  enable_ipv6 = String.length(System.get_env("ENABLE_IPV6", "")) > 0
+  enable_prometheus = String.length(System.get_env("ENABLE_PROMETHEUS", "")) > 0
 
-  # We want to force _some_ level of useful logging in production
-  acceptable_log_levels = ~w(debug info)a
-  log_level = String.to_existing_atom(System.get_env("LOG_LEVEL", "info"))
-
-  if log_level in acceptable_log_levels do
-    config :logger, level: log_level
-  else
-    Logger.error("Invalid log level: #{log_level}. Defaulting to info.")
-    config :logger, level: :info
-  end
+  config :logger, level: String.to_existing_atom(System.get_env("LOG_LEVEL", "debug"))
 
   config :pinchflat,
     yt_dlp_executable: System.find_executable("yt-dlp"),
+    apprise_executable: System.find_executable("apprise"),
+    media_directory: media_path,
     metadata_directory: metadata_path,
-    dns_cluster_query: System.get_env("DNS_CLUSTER_QUERY")
+    extras_directory: extras_path,
+    tmpfile_directory: tmpfile_path,
+    dns_cluster_query: System.get_env("DNS_CLUSTER_QUERY"),
+    expose_feed_endpoints: expose_feed_endpoints,
+    # This is configured in application.ex
+    timezone: "UTC",
+    log_path: log_path,
+    base_route_path: base_route_path
+
+  config :tzdata, :data_dir, tz_data_path
 
   config :pinchflat, Pinchflat.Repo,
     database: db_path,
-    pool_size: String.to_integer(System.get_env("POOL_SIZE") || "5")
+    journal_mode: journal_mode
+
+  config :pinchflat, Pinchflat.PromEx, disabled: !enable_prometheus
 
   # The secret key base is used to sign/encrypt cookies and other secrets.
   # A default value is used in config/dev.exs and config/test.exs but you
@@ -89,9 +145,10 @@ if config_env() == :prod do
       # Set it to  {0, 0, 0, 0, 0, 0, 0, 1} for local network only access.
       # See the documentation on https://hexdocs.pm/plug_cowboy/Plug.Cowboy.html
       # for details about using IPv6 vs IPv4 and loopback vs public addresses.
-      ip: {0, 0, 0, 0, 0, 0, 0, 0},
+      ip: if(enable_ipv6, do: {0, 0, 0, 0, 0, 0, 0, 0}, else: {0, 0, 0, 0}),
       port: String.to_integer(System.get_env("PORT") || "4000")
     ],
+    url: [path: base_route_path],
     secret_key_base: secret_key_base
 
   config :pinchflat, :logger, [
@@ -108,54 +165,4 @@ if config_env() == :prod do
        formatter: Logger.Formatter.new()
      }}
   ]
-
-  # ## SSL Support
-  #
-  # To get SSL working, you will need to add the `https` key
-  # to your endpoint configuration:
-  #
-  #     config :pinchflat, PinchflatWeb.Endpoint,
-  #       https: [
-  #         ...,
-  #         port: 443,
-  #         cipher_suite: :strong,
-  #         keyfile: System.get_env("SOME_APP_SSL_KEY_PATH"),
-  #         certfile: System.get_env("SOME_APP_SSL_CERT_PATH")
-  #       ]
-  #
-  # The `cipher_suite` is set to `:strong` to support only the
-  # latest and more secure SSL ciphers. This means old browsers
-  # and clients may not be supported. You can set it to
-  # `:compatible` for wider support.
-  #
-  # `:keyfile` and `:certfile` expect an absolute path to the key
-  # and cert in disk or a relative path inside priv, for example
-  # "priv/ssl/server.key". For all supported SSL configuration
-  # options, see https://hexdocs.pm/plug/Plug.SSL.html#configure/1
-  #
-  # We also recommend setting `force_ssl` in your endpoint, ensuring
-  # no data is ever sent via http, always redirecting to https:
-  #
-  #     config :pinchflat, PinchflatWeb.Endpoint,
-  #       force_ssl: [hsts: true]
-  #
-  # Check `Plug.SSL` for all available options in `force_ssl`.
-
-  # ## Configuring the mailer
-  #
-  # In production you need to configure the mailer to use a different adapter.
-  # Also, you may need to configure the Swoosh API client of your choice if you
-  # are not using SMTP. Here is an example of the configuration:
-  #
-  #     config :pinchflat, Pinchflat.Mailer,
-  #       adapter: Swoosh.Adapters.Mailgun,
-  #       api_key: System.get_env("MAILGUN_API_KEY"),
-  #       domain: System.get_env("MAILGUN_DOMAIN")
-  #
-  # For this example you need include a HTTP client required by Swoosh API client.
-  # Swoosh supports Hackney and Finch out of the box:
-  #
-  #     config :swoosh, :api_client, Swoosh.ApiClient.Hackney
-  #
-  # See https://hexdocs.pm/swoosh/Swoosh.html#module-installation for details.
 end

config/test.exs

@@ -3,6 +3,7 @@ import Config
 config :pinchflat,
   # Specifying backend data here makes mocking and local testing SUPER easy
   yt_dlp_executable: Path.join([File.cwd!(), "/test/support/scripts/yt-dlp-mocks/repeater.sh"]),
+  apprise_executable: Path.join([File.cwd!(), "/test/support/scripts/yt-dlp-mocks/repeater.sh"]),
   media_directory: Path.join([System.tmp_dir!(), "test", "media"]),
   metadata_directory: Path.join([System.tmp_dir!(), "test", "metadata"]),
   tmpfile_directory: Path.join([System.tmp_dir!(), "test", "tmpfiles"]),
@@ -17,7 +18,6 @@ config :pinchflat, Oban, testing: :manual
 # Run `mix help test` for more information.
 config :pinchflat, Pinchflat.Repo,
   database: Path.expand("../priv/repo/pinchflat_test.db", Path.dirname(__ENV__.file)),
-  pool_size: 5,
   pool: Ecto.Adapters.SQL.Sandbox
 
 # We don't run a server during test. If one is required,
@@ -33,8 +33,7 @@ config :pinchflat, Pinchflat.Mailer, adapter: Swoosh.Adapters.Test
 # Disable swoosh api client as it is only required for production adapters.
 config :swoosh, :api_client, false
 
-# Print only warnings and errors during test
-config :logger, level: :warning
+config :logger, level: :critical
 
 # Initialize plugs at runtime for faster test compilation
 config :phoenix, :plug_init_mode, :runtime

dev.Dockerfile

@@ -1,35 +0,0 @@
-# Extend from the official Elixir image.
-FROM elixir:latest
-
-# Install debian packages
-RUN apt-get update -qq
-RUN apt-get install -y inotify-tools ffmpeg \
-  python3 python3-pip python3-setuptools python3-wheel python3-dev
-
-# Install nodejs
-RUN curl -sL https://deb.nodesource.com/setup_20.x -o nodesource_setup.sh
-RUN bash nodesource_setup.sh
-RUN apt-get install nodejs
-RUN npm install -g yarn
-
-# Install baseline Elixir packages
-RUN mix local.hex --force
-RUN mix local.rebar --force
-
-# Download YT-DLP
-# NOTE: If you're seeing weird issues, consider using the FFMPEG released by yt-dlp
-RUN python3 -m pip install -U --pre yt-dlp --break-system-packages
-
-# Create app directory and copy the Elixir projects into it.
-WORKDIR /app
-COPY . ./
-
-# Needs permissions to be updated AFTER the copy step
-RUN chmod +x ./docker-run.dev.sh
-
-# Install Elixir deps
-RUN mix deps.get
-# Gives us iex shell history
-ENV ERL_AFLAGS="-kernel shell_history enabled"
-
-EXPOSE 4008

docker-compose.ci.yml

@@ -1,13 +1,15 @@
-version: '3'
 services:
   phx:
     build:
       context: .
-      dockerfile: dev.Dockerfile
+      dockerfile: ./docker/dev.Dockerfile
     environment:
       - MIX_ENV=test
     volumes:
       - '.:/app'
+      # These lines ensure the deps can be saved as build artifacts for caching
+      - '/app/deps'
+      - '/app/_build'
     ports:
       - '4008:4008'
     command: tail -F /dev/null

docker-compose.yml

@@ -1,14 +1,12 @@
-version: '3'
 services:
   phx:
     build:
       context: .
-      dockerfile: dev.Dockerfile
+      dockerfile: ./docker/dev.Dockerfile
     volumes:
       - '.:/app'
     ports:
       - '4008:4008'
-    command:
-      - ./docker-run.dev.sh
+    command: bash -c "chmod +x docker/docker-run.dev.sh && docker/docker-run.dev.sh"
     stdin_open: true
     tty: true

docker/dev.Dockerfile

@@ -0,0 +1,72 @@
+ARG ELIXIR_VERSION=1.18.4
+ARG OTP_VERSION=27.2.4
+ARG DEBIAN_VERSION=bookworm-20250428-slim
+
+ARG DEV_IMAGE="hexpm/elixir:${ELIXIR_VERSION}-erlang-${OTP_VERSION}-debian-${DEBIAN_VERSION}"
+
+FROM ${DEV_IMAGE}
+
+ARG TARGETPLATFORM
+RUN echo "Building for ${TARGETPLATFORM:?}"
+
+# Install debian packages
+RUN apt-get update -qq && \
+  apt-get install -y inotify-tools curl git openssh-client jq \
+    python3 python3-setuptools python3-wheel python3-dev pipx \
+    python3-mutagen locales procps build-essential graphviz zsh unzip
+
+# Install ffmpeg
+RUN export FFMPEG_DOWNLOAD=$(case ${TARGETPLATFORM:-linux/amd64} in \
+    "linux/amd64")   echo "https://github.com/yt-dlp/FFmpeg-Builds/releases/download/latest/ffmpeg-master-latest-linux64-gpl.tar.xz"   ;; \
+    "linux/arm64")   echo "https://github.com/yt-dlp/FFmpeg-Builds/releases/download/latest/ffmpeg-master-latest-linuxarm64-gpl.tar.xz" ;; \
+    *)               echo ""        ;; esac) && \
+    curl -L ${FFMPEG_DOWNLOAD} --output /tmp/ffmpeg.tar.xz && \
+    tar -xf /tmp/ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/bin/ "ffmpeg" && \
+    tar -xf /tmp/ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/bin/ "ffprobe"
+
+# Install nodejs and Yarn
+RUN curl -sL https://deb.nodesource.com/setup_20.x -o nodesource_setup.sh && \
+  bash nodesource_setup.sh && \
+  apt-get install -y nodejs && \
+  npm install -g yarn && \
+  # Install baseline Elixir packages
+  mix local.hex --force && \
+  mix local.rebar --force && \
+  # Install Deno - required for YouTube downloads (See yt-dlp#14404)
+  curl -fsSL https://deno.land/install.sh | DENO_INSTALL=/usr/local sh -s -- -y --no-modify-path && \
+  # Download and update YT-DLP
+  export YT_DLP_DOWNLOAD=$(case ${TARGETPLATFORM:-linux/amd64} in \
+  "linux/amd64")   echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux"   ;; \
+  "linux/arm64")   echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux_aarch64" ;; \
+  *)               echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux"        ;; esac) && \
+  curl -L ${YT_DLP_DOWNLOAD} -o /usr/local/bin/yt-dlp && \
+  chmod a+rx /usr/local/bin/yt-dlp && \
+  yt-dlp -U && \
+  # Install Apprise
+  export PIPX_HOME=/opt/pipx && \
+  export PIPX_BIN_DIR=/usr/local/bin && \
+  pipx install apprise && \
+  # Set up ZSH tools
+  chsh -s $(which zsh) && \
+  sh -c "$(curl -fsSL https://raw.githubusercontent.com/ohmyzsh/ohmyzsh/master/tools/install.sh)"
+
+# Set the locale
+RUN sed -i '/en_US.UTF-8/s/^# //g' /etc/locale.gen && locale-gen
+ENV LANG=en_US.UTF-8
+ENV LANGUAGE=en_US:en
+ENV LC_ALL=en_US.UTF-8
+
+WORKDIR /app
+
+COPY mix.exs mix.lock ./
+# Install Elixir deps
+# NOTE: this has to be before the bulk copy to ensure that deps are cached
+RUN MIX_ENV=dev mix deps.get && MIX_ENV=dev mix deps.compile
+RUN MIX_ENV=test mix deps.get && MIX_ENV=test mix deps.compile
+
+COPY . ./
+
+# Gives us iex shell history
+ENV ERL_AFLAGS="-kernel shell_history enabled"
+
+EXPOSE 4008

docker/docker-run.dev.sh

@@ -2,12 +2,12 @@
 
 set -e
 
-# Ensure the app's deps are installed
+echo "\nInstalling Elixir deps..."
 mix deps.get
 
-# Install JS deps
-echo "\nInstalling JS..."
-cd assets && yarn install
+# Install both project-level and assets-level JS dependencies
+echo "\nInstalling JS deps..."
+yarn install && cd assets && yarn install
 cd ..
 
 # Potentially Set up the database

docker/selfhosted.Dockerfile

@@ -0,0 +1,146 @@
+# Find eligible builder and runner images on Docker Hub. We use Ubuntu/Debian
+# instead of Alpine to avoid DNS resolution issues in production.
+ARG ELIXIR_VERSION=1.18.4
+ARG OTP_VERSION=27.2.4
+ARG DEBIAN_VERSION=bookworm-20250428-slim
+
+ARG BUILDER_IMAGE="hexpm/elixir:${ELIXIR_VERSION}-erlang-${OTP_VERSION}-debian-${DEBIAN_VERSION}"
+ARG RUNNER_IMAGE="debian:${DEBIAN_VERSION}"
+
+FROM ${BUILDER_IMAGE} AS builder
+
+ARG TARGETPLATFORM
+RUN echo "Building for ${TARGETPLATFORM:?}"
+
+# install build dependencies
+RUN apt-get update -y && \
+    # System packages
+    apt-get install -y \
+      build-essential \
+      git \
+      curl && \
+    # Node.js and Yarn
+    curl -sL https://deb.nodesource.com/setup_20.x -o nodesource_setup.sh && \
+    bash nodesource_setup.sh && \
+    apt-get install -y nodejs && \
+    npm install -g yarn && \
+    # Hex and Rebar
+    mix local.hex --force && \
+    mix local.rebar --force && \
+    # FFmpeg (latest build that doesn't cause an illegal instruction error for some users - see #347)
+    export FFMPEG_DOWNLOAD=$(case ${TARGETPLATFORM:-linux/amd64} in \
+    "linux/amd64")   echo "https://github.com/yt-dlp/FFmpeg-Builds/releases/download/autobuild-2024-07-30-14-10/ffmpeg-N-116468-g0e09f6d690-linux64-gpl.tar.xz"   ;; \
+    "linux/arm64")   echo "https://github.com/yt-dlp/FFmpeg-Builds/releases/download/autobuild-2024-07-30-14-10/ffmpeg-N-116468-g0e09f6d690-linuxarm64-gpl.tar.xz" ;; \
+    *)               echo ""        ;; esac) && \
+    curl -L ${FFMPEG_DOWNLOAD} --output /tmp/ffmpeg.tar.xz && \
+    tar -xf /tmp/ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/local/bin/ "ffmpeg" && \
+    tar -xf /tmp/ffmpeg.tar.xz --strip-components=2 --no-anchored -C /usr/local/bin/ "ffprobe" && \
+    # Cleanup
+    apt-get clean && \
+    rm -f /var/lib/apt/lists/*_*
+
+# prepare build dir
+WORKDIR /app
+
+# set build ENV
+ENV MIX_ENV="prod"
+ENV ERL_FLAGS="+JPperf true"
+
+# install mix dependencies
+COPY mix.exs mix.lock ./
+RUN mix deps.get --only $MIX_ENV && mkdir config
+
+# copy compile-time config files before we compile dependencies
+# to ensure any relevant config change will trigger the dependencies
+# to be re-compiled.
+COPY config/config.exs config/${MIX_ENV}.exs config/
+RUN mix deps.compile
+
+COPY priv priv
+COPY lib lib
+COPY assets assets
+
+# Compile assets
+RUN yarn --cwd assets install && mix assets.deploy && mix compile
+
+# Changes to config/runtime.exs don't require recompiling the code
+COPY config/runtime.exs config/
+
+COPY rel rel
+RUN mix release
+
+## -- Release Stage --
+
+FROM ${RUNNER_IMAGE}
+
+ARG TARGETPLATFORM
+ARG PORT=8945
+
+COPY --from=builder ./usr/local/bin/ffmpeg /usr/bin/ffmpeg
+COPY --from=builder ./usr/local/bin/ffprobe /usr/bin/ffprobe
+
+RUN apt-get update -y && \
+    # System packages
+    apt-get install -y \
+      libstdc++6 \
+      openssl \
+      libncurses5 \
+      locales \
+      ca-certificates \
+      python3-mutagen \
+      curl \
+      zip \
+      openssh-client \
+      nano \
+      python3 \
+      pipx \
+      jq \
+      # unzip is needed for Deno
+      unzip \
+      procps && \
+    # Install Deno - required for YouTube downloads (See yt-dlp#14404)
+    curl -fsSL https://deno.land/install.sh | DENO_INSTALL=/usr/local sh -s -- -y --no-modify-path && \
+    # Apprise
+    export PIPX_HOME=/opt/pipx && \
+    export PIPX_BIN_DIR=/usr/local/bin && \
+    pipx install apprise && \
+    # yt-dlp
+    export YT_DLP_DOWNLOAD=$(case ${TARGETPLATFORM:-linux/amd64} in \
+    "linux/amd64")   echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux"   ;; \
+    "linux/arm64")   echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux_aarch64" ;; \
+    *)               echo "https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp_linux"        ;; esac) && \
+    curl -L ${YT_DLP_DOWNLOAD} -o /usr/local/bin/yt-dlp && \
+    chmod a+rx /usr/local/bin/yt-dlp && \
+    yt-dlp -U && \
+    # Set the locale
+    sed -i '/en_US.UTF-8/s/^# //g' /etc/locale.gen && locale-gen && \
+    # Clean up
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+# More locale setup
+ENV LANG=en_US.UTF-8
+ENV LANGUAGE=en_US:en
+ENV LC_ALL=en_US.UTF-8
+
+WORKDIR "/app"
+
+# Set up data volumes
+RUN mkdir -p /config /downloads /etc/elixir_tzdata_data /etc/yt-dlp/plugins && \ 
+  chmod ugo+rw /etc/elixir_tzdata_data /etc/yt-dlp /etc/yt-dlp/plugins /usr/local/bin /usr/local/bin/yt-dlp
+
+# set runner ENV
+ENV MIX_ENV="prod"
+ENV PORT=${PORT}
+ENV RUN_CONTEXT="selfhosted"
+ENV UMASK=022
+EXPOSE ${PORT}
+
+# Only copy the final release from the build stage
+COPY --from=builder /app/_build/${MIX_ENV}/rel/pinchflat ./
+
+HEALTHCHECK --interval=30s --start-period=15s \
+  CMD curl --fail http://localhost:${PORT}/healthcheck || exit 1
+
+# Start the app
+CMD ["/app/bin/docker_start"]

ideas.md

@@ -1,6 +0,0 @@
-- Write media datbase ID as metadata/to file/whatever so it gives us an option to retroactively match media to the DB down the line. Useful if someone moves the media without informing the UI
-  - Use a UUID for the media database ID (or at least alongside it)
-- Look into this and its recommended plugins https://hexdocs.pm/ex_check/readme.html
-- Add output template option for the source's friendly name
-- TODO: Install Elixir 1.16.2 when available to fix bug with `ex_check` https://github.com/karolsluszniak/ex_check/issues/41#issuecomment-1921390413
-  - delete `{:elixir_tests, "mix test"}` and formatting check

lib/pinchflat/application.ex

@@ -4,32 +4,35 @@ defmodule Pinchflat.Application do
   @moduledoc false
 
   use Application
+  require Logger
 
   @impl true
   def start(_type, _args) do
-    children = [
+    check_and_update_timezone()
+    attach_oban_telemetry()
+    Logger.add_handlers(:pinchflat)
+
+    # See https://hexdocs.pm/elixir/Supervisor.html
+    # for other strategies and supported options
+    [
+      Pinchflat.PromEx,
       PinchflatWeb.Telemetry,
       Pinchflat.Repo,
       # Must be before startup tasks
+      Pinchflat.Boot.PreJobStartupTasks,
       {Oban, Application.fetch_env!(:pinchflat, Oban)},
-      Pinchflat.StartupTasks,
+      Pinchflat.Boot.PostJobStartupTasks,
       {DNSCluster, query: Application.get_env(:pinchflat, :dns_cluster_query) || :ignore},
       {Phoenix.PubSub, name: Pinchflat.PubSub},
       # Start the Finch HTTP client for sending emails
       {Finch, name: Pinchflat.Finch},
       # Start a worker by calling: Pinchflat.Worker.start_link(arg)
       # {Pinchflat.Worker, arg},
-      # Start to serve requests, typically the last entry
-      PinchflatWeb.Endpoint
+      # Start to serve requests, typically the last entry (except for the post-boot tasks)
+      PinchflatWeb.Endpoint,
+      Pinchflat.Boot.PostBootStartupTasks
     ]
-
-    :ok = Oban.Telemetry.attach_default_logger()
-    Logger.add_handlers(:pinchflat)
-
-    # See https://hexdocs.pm/elixir/Supervisor.html
-    # for other strategies and supported options
-    opts = [strategy: :one_for_one, name: Pinchflat.Supervisor]
-    Supervisor.start_link(children, opts)
+    |> Supervisor.start_link(strategy: :one_for_one, name: Pinchflat.Supervisor)
   end
 
   # Tell Phoenix to update the endpoint configuration
@@ -39,4 +42,27 @@ defmodule Pinchflat.Application do
     PinchflatWeb.Endpoint.config_change(changed, removed)
     :ok
   end
+
+  defp attach_oban_telemetry do
+    events = [[:oban, :job, :start], [:oban, :job, :stop], [:oban, :job, :exception]]
+
+    :ok = Oban.Telemetry.attach_default_logger()
+    :telemetry.attach_many("job-telemetry-broadcast", events, &PinchflatWeb.Telemetry.job_state_change_broadcast/4, [])
+  end
+
+  # This has to be here (rather than runtime.exs) since the `tzdata` application
+  # has to be started before we can check the timezone
+  defp check_and_update_timezone do
+    attempted_timezone = System.get_env("TIMEZONE") || System.get_env("TZ") || "UTC"
+
+    valid_timezone =
+      if Tzdata.zone_exists?(attempted_timezone) do
+        attempted_timezone
+      else
+        Logger.warning("Invalid timezone #{attempted_timezone}, defaulting to UTC")
+        "UTC"
+      end
+
+    Application.put_env(:pinchflat, :timezone, valid_timezone)
+  end
 end

lib/pinchflat/boot/post_boot_startup_tasks.ex

@@ -0,0 +1,46 @@
+defmodule Pinchflat.Boot.PostBootStartupTasks do
+  @moduledoc """
+  This module is responsible for running startup tasks on app boot
+  AFTER all other boot steps have taken place and the app is ready to serve requests.
+
+  It's a GenServer because that plays REALLY nicely with the existing
+  Phoenix supervision tree.
+  """
+
+  alias Pinchflat.YtDlp.UpdateWorker, as: YtDlpUpdateWorker
+
+  # restart: :temporary means that this process will never be restarted (ie: will run once and then die)
+  use GenServer, restart: :temporary
+  import Ecto.Query, warn: false
+
+  def start_link(opts \\ []) do
+    GenServer.start_link(__MODULE__, %{env: Application.get_env(:pinchflat, :env)}, opts)
+  end
+
+  @doc """
+  Runs post-boot application startup tasks.
+
+  Any code defined here will run every time the application starts. You must
+  make sure that the code is idempotent and safe to run multiple times.
+
+  This is a good place to set up default settings, create initial records, stuff like that.
+  Should be fast - anything with the potential to be slow should be kicked off as a job instead.
+  """
+  @impl true
+  def init(%{env: :test} = state) do
+    # Do nothing _as part of the app bootup process_.
+    # Since bootup calls `start_link` and that's where the `env` state is injected,
+    # you can still call `.init()` manually to run these tasks for testing purposes
+    {:ok, state}
+  end
+
+  def init(state) do
+    update_yt_dlp()
+
+    {:ok, state}
+  end
+
+  defp update_yt_dlp do
+    YtDlpUpdateWorker.kickoff()
+  end
+end

lib/pinchflat/boot/post_job_startup_tasks.ex

@@ -1,6 +1,7 @@
-defmodule Pinchflat.StartupTasks do
+defmodule Pinchflat.Boot.PostJobStartupTasks do
   @moduledoc """
-  This module is responsible for running startup tasks on app boot.
+  This module is responsible for running startup tasks on app boot
+  AFTER the job runner has initialized.
 
   It's a GenServer because that plays REALLY nicely with the existing
   Phoenix supervision tree.
@@ -10,12 +11,8 @@ defmodule Pinchflat.StartupTasks do
   use GenServer, restart: :temporary
   import Ecto.Query, warn: false
 
-  alias Pinchflat.Repo
-  alias Pinchflat.Settings
-  alias Pinchflat.Workers.DataBackfillWorker
-
   def start_link(opts \\ []) do
-    GenServer.start_link(__MODULE__, %{}, opts)
+    GenServer.start_link(__MODULE__, %{env: Application.get_env(:pinchflat, :env)}, opts)
   end
 
   @doc """
@@ -28,23 +25,16 @@ defmodule Pinchflat.StartupTasks do
   Should be fast - anything with the potential to be slow should be kicked off as a job instead.
   """
   @impl true
-  def init(state) do
-    apply_default_settings()
-    enqueue_backfill_worker()
-
+  def init(%{env: :test} = state) do
+    # Do nothing _as part of the app bootup process_.
+    # Since bootup calls `start_link` and that's where the `env` state is injected,
+    # you can still call `.init()` manually to run these tasks for testing purposes
     {:ok, state}
   end
 
-  defp apply_default_settings do
-    Settings.fetch!(:onboarding, true)
-    Settings.fetch!(:pro_enabled, false)
-  end
+  def init(state) do
+    # Nothing at the moment!
 
-  defp enqueue_backfill_worker do
-    DataBackfillWorker.cancel_pending_backfill_jobs()
-
-    %{}
-    |> DataBackfillWorker.new()
-    |> Repo.insert_unique_job()
+    {:ok, state}
   end
 end

lib/pinchflat/boot/pre_job_startup_tasks.ex

@@ -0,0 +1,121 @@
+defmodule Pinchflat.Boot.PreJobStartupTasks do
+  @moduledoc """
+  This module is responsible for running startup tasks on app boot
+  BEFORE the job runner has initiallized.
+
+  It's a GenServer because that plays REALLY nicely with the existing
+  Phoenix supervision tree.
+  """
+
+  # restart: :temporary means that this process will never be restarted (ie: will run once and then die)
+  use GenServer, restart: :temporary
+  import Ecto.Query, warn: false
+  require Logger
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Settings
+  alias Pinchflat.Utils.FilesystemUtils
+
+  alias Pinchflat.Lifecycle.UserScripts.CommandRunner, as: UserScriptRunner
+
+  def start_link(opts \\ []) do
+    GenServer.start_link(__MODULE__, %{env: Application.get_env(:pinchflat, :env)}, opts)
+  end
+
+  @doc """
+  Runs application startup tasks.
+
+  Any code defined here will run every time the application starts. You must
+  make sure that the code is idempotent and safe to run multiple times.
+
+  This is a good place to set up default settings, create initial records, stuff like that.
+  Should be fast - anything with the potential to be slow should be kicked off as a job instead.
+  """
+  @impl true
+  def init(%{env: :test} = state) do
+    # Do nothing _as part of the app bootup process_.
+    # Since bootup calls `start_link` and that's where the `env` state is injected,
+    # you can still call `.init()` manually to run these tasks for testing purposes
+    {:ok, state}
+  end
+
+  def init(state) do
+    ensure_tmpfile_directory()
+    reset_executing_jobs()
+    create_blank_yt_dlp_files()
+    create_blank_user_script_file()
+    apply_default_settings()
+    run_app_init_script()
+
+    {:ok, state}
+  end
+
+  defp ensure_tmpfile_directory do
+    tmpfile_dir = Application.get_env(:pinchflat, :tmpfile_directory)
+
+    if !File.exists?(tmpfile_dir) do
+      File.mkdir_p!(tmpfile_dir)
+    end
+  end
+
+  # If a node cannot gracefully shut down, the currently executing jobs get stuck
+  # in the "executing" state. This is a problem because the job runner will not
+  # pick them up again
+  defp reset_executing_jobs do
+    {count, _} =
+      Oban.Job
+      |> where(state: "executing")
+      |> Repo.update_all(set: [state: "retryable"])
+
+    Logger.info("Reset #{count} executing jobs")
+  end
+
+  defp create_blank_yt_dlp_files do
+    files = ["cookies.txt", "yt-dlp-configs/base-config.txt"]
+    base_dir = Application.get_env(:pinchflat, :extras_directory)
+
+    Enum.each(files, fn file ->
+      filepath = Path.join(base_dir, file)
+
+      if !File.exists?(filepath) do
+        Logger.info("Creating blank file: #{filepath}")
+
+        FilesystemUtils.write_p!(filepath, "")
+      end
+    end)
+  end
+
+  defp create_blank_user_script_file do
+    base_dir = Application.get_env(:pinchflat, :extras_directory)
+    filepath = Path.join([base_dir, "user-scripts", "lifecycle"])
+
+    if !File.exists?(filepath) do
+      Logger.info("Creating blank file and making it executable: #{filepath}")
+
+      FilesystemUtils.write_p!(filepath, "")
+      File.chmod(filepath, 0o755)
+    end
+  end
+
+  defp apply_default_settings do
+    {:ok, yt_dlp_version} = yt_dlp_runner().version()
+    {:ok, apprise_version} = apprise_runner().version()
+
+    Settings.set(yt_dlp_version: yt_dlp_version)
+    Settings.set(apprise_version: apprise_version)
+  end
+
+  defp run_app_init_script do
+    runner = Application.get_env(:pinchflat, :user_script_runner, UserScriptRunner)
+
+    runner.run(:app_init, %{})
+  end
+
+  defp yt_dlp_runner do
+    Application.get_env(:pinchflat, :yt_dlp_runner)
+  end
+
+  defp apprise_runner do
+    Application.get_env(:pinchflat, :apprise_runner)
+  end
+end

lib/pinchflat/downloading/download_option_builder.ex

@@ -0,0 +1,238 @@
+defmodule Pinchflat.Downloading.DownloadOptionBuilder do
+  @moduledoc """
+  Builds the options for yt-dlp to download media based on the given media profile.
+  """
+
+  alias Pinchflat.Sources
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.Downloading.OutputPathBuilder
+  alias Pinchflat.Downloading.QualityOptionBuilder
+
+  alias Pinchflat.Utils.FilesystemUtils, as: FSUtils
+
+  @doc """
+  Builds the options for yt-dlp to download media based on the given media's profile.
+
+  Returns {:ok, [Keyword.t()]}
+  """
+  def build(%MediaItem{} = media_item_with_preloads, override_opts \\ []) do
+    media_profile = media_item_with_preloads.source.media_profile
+
+    built_options =
+      default_options(override_opts) ++
+        subtitle_options(media_profile) ++
+        thumbnail_options(media_item_with_preloads) ++
+        metadata_options(media_profile) ++
+        quality_options(media_profile) ++
+        sponsorblock_options(media_profile) ++
+        output_options(media_item_with_preloads) ++
+        config_file_options(media_item_with_preloads)
+
+    {:ok, built_options}
+  end
+
+  @doc """
+  Builds the output path for yt-dlp to download media based on the given source's
+  or media_item's media profile. Uses the source's override output path template if it exists.
+
+  Accepts a %MediaItem{} or %Source{} struct. If a %Source{} struct is passed, it
+  will use a default %MediaItem{} struct with the given source.
+
+  Returns binary()
+  """
+  def build_output_path_for(%Source{} = source_with_preloads) do
+    build_output_path_for(%MediaItem{source: source_with_preloads})
+  end
+
+  def build_output_path_for(%MediaItem{} = media_item_with_preloads) do
+    output_path_template = Sources.output_path_template(media_item_with_preloads.source)
+
+    build_output_path(output_path_template, media_item_with_preloads)
+  end
+
+  @doc """
+  Builds the quality options for yt-dlp to download media based on the given source's
+  or media_item's media profile. Useful for helping predict final filepath of downloaded
+  media.
+
+  returns [Keyword.t()]
+  """
+  def build_quality_options_for(%Source{} = source_with_preloads) do
+    build_quality_options_for(%MediaItem{source: source_with_preloads})
+  end
+
+  def build_quality_options_for(%MediaItem{} = media_item_with_preloads) do
+    media_profile = media_item_with_preloads.source.media_profile
+
+    quality_options(media_profile)
+  end
+
+  defp default_options(override_opts) do
+    overwrite_behaviour = Keyword.get(override_opts, :overwrite_behaviour, :force_overwrites)
+
+    [
+      :no_progress,
+      overwrite_behaviour,
+      # This makes the date metadata conform to what jellyfin expects
+      parse_metadata: "%(upload_date>%Y-%m-%d)s:(?P<meta_date>.+)"
+    ]
+  end
+
+  defp subtitle_options(media_profile) do
+    mapped_struct = Map.from_struct(media_profile)
+
+    Enum.reduce(mapped_struct, [], fn attr, acc ->
+      case {attr, media_profile} do
+        {{:download_subs, true}, _} ->
+          # Force SRT for now - MAY provide as an option in the future
+          acc ++ [:write_subs, convert_subs: "srt"]
+
+        {{:download_auto_subs, true}, %{download_subs: true}} ->
+          acc ++ [:write_auto_subs]
+
+        {{:download_auto_subs, true}, %{embed_subs: true}} ->
+          acc ++ [:write_auto_subs]
+
+        {{:embed_subs, true}, %{preferred_resolution: pr}} when pr != :audio ->
+          acc ++ [:embed_subs]
+
+        {{:sub_langs, sub_langs}, %{download_subs: true}} ->
+          acc ++ [sub_langs: sub_langs]
+
+        {{:sub_langs, sub_langs}, %{embed_subs: true}} ->
+          acc ++ [sub_langs: sub_langs]
+
+        _ ->
+          acc
+      end
+    end)
+  end
+
+  defp thumbnail_options(media_item_with_preloads) do
+    media_profile = media_item_with_preloads.source.media_profile
+    mapped_struct = Map.from_struct(media_profile)
+
+    Enum.reduce(mapped_struct, [], fn attr, acc ->
+      case attr do
+        {:download_thumbnail, true} ->
+          thumbnail_save_location = determine_thumbnail_location(media_item_with_preloads)
+
+          acc ++ [:write_thumbnail, convert_thumbnail: "jpg", output: "thumbnail:#{thumbnail_save_location}"]
+
+        {:embed_thumbnail, true} ->
+          acc ++ [:embed_thumbnail, convert_thumbnail: "jpg"]
+
+        _ ->
+          acc
+      end
+    end)
+  end
+
+  defp metadata_options(media_profile) do
+    mapped_struct = Map.from_struct(media_profile)
+
+    Enum.reduce(mapped_struct, [], fn attr, acc ->
+      case attr do
+        {:download_metadata, true} -> acc ++ [:write_info_json, :clean_info_json]
+        {:embed_metadata, true} -> acc ++ [:embed_metadata]
+        _ -> acc
+      end
+    end)
+  end
+
+  defp quality_options(media_profile) do
+    QualityOptionBuilder.build(media_profile)
+  end
+
+  defp sponsorblock_options(media_profile) do
+    categories = media_profile.sponsorblock_categories
+    behaviour = media_profile.sponsorblock_behaviour
+
+    case {behaviour, categories} do
+      {_, []} -> []
+      {:remove, _} -> [sponsorblock_remove: Enum.join(categories, ",")]
+      {:mark, _} -> [sponsorblock_mark: Enum.join(categories, ",")]
+      {:disabled, _} -> []
+    end
+  end
+
+  # This is put here instead of the CommandRunner module because it should only
+  # be applied to downloading - if it were in CommandRunner it would apply to
+  # all yt-dlp commands (like indexing)
+  defp config_file_options(media_item) do
+    base_dir = Path.join(Application.get_env(:pinchflat, :extras_directory), "yt-dlp-configs")
+    # Ordered by priority - the first file has the highest priority
+    filenames = [
+      "media-item-#{media_item.id}-config.txt",
+      "source-#{media_item.source_id}-config.txt",
+      "media-profile-#{media_item.source.media_profile_id}-config.txt",
+      "base-config.txt"
+    ]
+
+    config_filepaths =
+      Enum.reduce(filenames, [], fn filename, acc ->
+        filepath = Path.join(base_dir, filename)
+
+        if FSUtils.exists_and_nonempty?(filepath) do
+          [filepath | acc]
+        else
+          acc
+        end
+      end)
+
+    Enum.map(config_filepaths, fn filepath -> {:config_locations, filepath} end)
+  end
+
+  defp output_options(media_item_with_preloads) do
+    [
+      output: build_output_path_for(media_item_with_preloads)
+    ]
+  end
+
+  defp build_output_path(string, media_item_with_preloads) do
+    additional_options_map = output_options_map(media_item_with_preloads)
+    {:ok, output_path} = OutputPathBuilder.build(string, additional_options_map)
+
+    Path.join(base_directory(), output_path)
+  end
+
+  defp output_options_map(media_item_with_preloads) do
+    source = media_item_with_preloads.source
+
+    %{
+      "media_item_id" => to_string(media_item_with_preloads.id),
+      "source_id" => to_string(source.id),
+      "media_profile_id" => to_string(source.media_profile_id),
+      "source_custom_name" => source.custom_name,
+      "source_collection_id" => source.collection_id,
+      "source_collection_name" => source.collection_name,
+      "source_collection_type" => to_string(source.collection_type),
+      "media_playlist_index" => pad_int(media_item_with_preloads.playlist_index),
+      "media_upload_date_index" => pad_int(media_item_with_preloads.upload_date_index)
+    }
+  end
+
+  # I don't love the string manipulation here, but what can ya' do.
+  # It's dependent on the output_path_template being a string ending `.{{ ext }}`
+  # (or equivalent), but that's validated by the MediaProfile schema.
+  defp determine_thumbnail_location(media_item_with_preloads) do
+    output_path_template = Sources.output_path_template(media_item_with_preloads.source)
+
+    output_path_template
+    |> String.split(~r{\.}, include_captures: true)
+    |> List.insert_at(-3, "-thumb")
+    |> Enum.join()
+    |> build_output_path(media_item_with_preloads)
+  end
+
+  defp pad_int(integer, count \\ 2, padding \\ "0") do
+    integer
+    |> to_string()
+    |> String.pad_leading(count, padding)
+  end
+
+  defp base_directory do
+    Application.get_env(:pinchflat, :media_directory)
+  end
+end

lib/pinchflat/downloading/downloading_helpers.ex

@@ -0,0 +1,102 @@
+defmodule Pinchflat.Downloading.DownloadingHelpers do
+  @moduledoc """
+  Methods for helping download media
+
+  Many of these methods are made to be kickoff or be consumed by workers.
+  """
+
+  require Logger
+
+  use Pinchflat.Media.MediaQuery
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.Downloading.MediaDownloadWorker
+
+  @doc """
+  Starts tasks for downloading media for any of a sources _pending_ media items.
+  Jobs are not enqueued if the source is set to not download media. This will return :ok.
+
+  NOTE: this starts a download for each media item that is pending,
+  not just the ones that were indexed in this job run. This should ensure
+  that any stragglers are caught if, for some reason, they weren't enqueued
+  or somehow got de-queued.
+
+  Returns :ok
+  """
+  def enqueue_pending_download_tasks(source, job_opts \\ [])
+
+  def enqueue_pending_download_tasks(%Source{download_media: true} = source, job_opts) do
+    source
+    |> Media.list_pending_media_items_for()
+    |> Enum.each(&MediaDownloadWorker.kickoff_with_task(&1, %{}, job_opts))
+  end
+
+  def enqueue_pending_download_tasks(%Source{download_media: false}, _job_opts) do
+    :ok
+  end
+
+  @doc """
+  Deletes ALL pending tasks for a source's media items.
+
+  Returns :ok
+  """
+  def dequeue_pending_download_tasks(%Source{} = source) do
+    source
+    |> Media.list_pending_media_items_for()
+    |> Enum.each(&Tasks.delete_pending_tasks_for/1)
+  end
+
+  @doc """
+  Takes a single media item and enqueues a download job if the media should be
+  downloaded, based on the source's download settings and whether media is
+  considered pending.
+
+  Returns {:ok, %Task{}} | {:error, :should_not_download} | {:error, any()}
+  """
+  def kickoff_download_if_pending(%MediaItem{} = media_item, job_opts \\ []) do
+    media_item = Repo.preload(media_item, :source)
+
+    if media_item.source.download_media && Media.pending_download?(media_item) do
+      Logger.info("Kicking off download for media item ##{media_item.id} (#{media_item.media_id})")
+
+      MediaDownloadWorker.kickoff_with_task(media_item, %{}, job_opts)
+    else
+      {:error, :should_not_download}
+    end
+  end
+
+  @doc """
+  For a given source, enqueues download jobs for all media items _that have already been downloaded_.
+
+  This is useful for when a source's download settings have changed and you want to run through all
+  existing media and retry the download. For instance, if the source didn't originally download thumbnails
+  and you've changed the source to download them, you can use this to download all the thumbnails for
+  existing media items.
+
+  NOTE: does not delete existing files whatsoever. Does not overwrite the existing media file if it exists
+  at the location it expects. Will cause a full redownload of everything if the output template has changed
+
+  NOTE: unrelated to the MediaQualityUpgradeWorker, which is for redownloading media items for quality upgrades
+  or improved sponsorblock segments
+
+  Returns [{:ok, %Task{}} | {:error, any()}]
+  """
+  def kickoff_redownload_for_existing_media(%Source{} = source) do
+    MediaQuery.new()
+    |> MediaQuery.require_assoc(:media_profile)
+    |> where(
+      ^dynamic(
+        [m, s, mp],
+        ^MediaQuery.for_source(source) and
+          ^MediaQuery.downloaded() and
+          not (^MediaQuery.download_prevented())
+      )
+    )
+    |> Repo.all()
+    |> Enum.map(&MediaDownloadWorker.kickoff_with_task/1)
+  end
+end

lib/pinchflat/downloading/media_download_worker.ex

@@ -0,0 +1,154 @@
+defmodule Pinchflat.Downloading.MediaDownloadWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :media_fetching,
+    priority: 5,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable, :executing]],
+    tags: ["media_item", "media_fetching", "show_in_dashboard"]
+
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Tasks
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Media.FileSyncing
+  alias Pinchflat.Downloading.MediaDownloader
+
+  alias Pinchflat.Lifecycle.UserScripts.CommandRunner, as: UserScriptRunner
+
+  @doc """
+  Starts the media_item media download worker and creates a task for the media_item.
+
+  Returns {:ok, %Task{}} | {:error, :duplicate_job} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff_with_task(media_item, job_args \\ %{}, job_opts \\ []) do
+    %{id: media_item.id}
+    |> Map.merge(job_args)
+    |> MediaDownloadWorker.new(job_opts)
+    |> Tasks.create_job_with_task(media_item)
+  end
+
+  @doc """
+  For a given media item, download the media alongside any options.
+  Does not download media if its source is set to not download media
+  (unless forced).
+
+  Options:
+    - `force`: force download even if the source is set to not download media. Fully
+      re-downloads media, including the video
+    - `quality_upgrade?`: re-downloads media, including the video. Does not force download
+      if the source is set to not download media
+
+  Returns :ok | {:error, any, ...any}
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => media_item_id} = args}) do
+    should_force = Map.get(args, "force", false)
+    is_quality_upgrade = Map.get(args, "quality_upgrade?", false)
+
+    media_item = fetch_and_run_prevent_download_user_script(media_item_id)
+
+    if should_download_media?(media_item, should_force, is_quality_upgrade) do
+      download_media_and_schedule_jobs(media_item, is_quality_upgrade, should_force)
+    else
+      :ok
+    end
+  rescue
+    Ecto.NoResultsError -> Logger.info("#{__MODULE__} discarded: media item #{media_item_id} not found")
+    Ecto.StaleEntryError -> Logger.info("#{__MODULE__} discarded: media item #{media_item_id} stale")
+  end
+
+  # If this is a quality upgrade, only check if the source is set to download media
+  # or that the media item's download hasn't been prevented
+  defp should_download_media?(media_item, should_force, true = _is_quality_upgrade) do
+    (media_item.source.download_media && !media_item.prevent_download) || should_force
+  end
+
+  # If it's not a quality upgrade, additionally check if the media item is pending download
+  defp should_download_media?(media_item, should_force, _is_quality_upgrade) do
+    source = media_item.source
+    is_pending = Media.pending_download?(media_item)
+
+    (is_pending && source.download_media && !media_item.prevent_download) || should_force
+  end
+
+  # If a user script exists and, when run, returns a non-zero exit code, prevent this and all future downloads
+  # of the media item.
+  defp fetch_and_run_prevent_download_user_script(media_item_id) do
+    media_item = Media.get_media_item!(media_item_id)
+
+    {:ok, media_item} =
+      case run_user_script(:media_pre_download, media_item) do
+        {:ok, _, exit_code} when exit_code != 0 -> Media.update_media_item(media_item, %{prevent_download: true})
+        _ -> {:ok, media_item}
+      end
+
+    Repo.preload(media_item, :source)
+  end
+
+  defp download_media_and_schedule_jobs(media_item, is_quality_upgrade, should_force) do
+    overwrite_behaviour = if should_force || is_quality_upgrade, do: :force_overwrites, else: :no_force_overwrites
+    override_opts = [overwrite_behaviour: overwrite_behaviour]
+
+    case MediaDownloader.download_for_media_item(media_item, override_opts) do
+      {:ok, downloaded_media_item} ->
+        {:ok, updated_media_item} =
+          Media.update_media_item(downloaded_media_item, %{
+            media_size_bytes: compute_media_filesize(downloaded_media_item),
+            media_redownloaded_at: get_redownloaded_at(is_quality_upgrade)
+          })
+
+        :ok = FileSyncing.delete_outdated_files(media_item, updated_media_item)
+        run_user_script(:media_downloaded, updated_media_item)
+
+        :ok
+
+      {:recovered, _media_item, _message} ->
+        {:error, :retry}
+
+      {:error, :unsuitable_for_download, _message} ->
+        {:ok, :non_retry}
+
+      {:error, _error_atom, message} ->
+        action_on_error(message)
+    end
+  end
+
+  defp compute_media_filesize(media_item) do
+    case File.stat(media_item.media_filepath) do
+      {:ok, %{size: size}} -> size
+      _ -> nil
+    end
+  end
+
+  defp get_redownloaded_at(true), do: DateTime.utc_now()
+  defp get_redownloaded_at(_), do: nil
+
+  defp action_on_error(message) do
+    # This will attempt re-download at the next indexing, but it won't be retried
+    # immediately as part of job failure logic
+    non_retryable_errors = [
+      "Video unavailable",
+      "Sign in to confirm",
+      "This video is available to this channel's members"
+    ]
+
+    if String.contains?(to_string(message), non_retryable_errors) do
+      Logger.error("yt-dlp download will not be retried: #{inspect(message)}")
+
+      {:ok, :non_retry}
+    else
+      {:error, :download_failed}
+    end
+  end
+
+  # NOTE: I like this pattern of using the default value so that I don't have to
+  # define it in config.exs (and friends). Consider using this elsewhere.
+  defp run_user_script(event, media_item) do
+    runner = Application.get_env(:pinchflat, :user_script_runner, UserScriptRunner)
+
+    runner.run(event, media_item)
+  end
+end

lib/pinchflat/downloading/media_downloader.ex

@@ -0,0 +1,216 @@
+defmodule Pinchflat.Downloading.MediaDownloader do
+  @moduledoc """
+  This is the integration layer for actually downloading media.
+  It takes into account the media profile's settings in order
+  to download the media with the desired options.
+  """
+
+  require Logger
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Sources
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.Utils.StringUtils
+  alias Pinchflat.Metadata.NfoBuilder
+  alias Pinchflat.Metadata.MetadataParser
+  alias Pinchflat.Metadata.MetadataFileHelpers
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Downloading.DownloadOptionBuilder
+
+  alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+
+  @doc """
+  Downloads media for a media item, updating the media item based on the metadata
+  returned by yt-dlp. Encountered errors are saved to the Media Item record. Saves
+  the entire metadata response to the associated media_metadata record.
+
+  NOTE: related methods (like the download worker) won't download if Pthe media item's source
+  is set to not download media. However, I'm not enforcing that here since I need this for testing.
+  This may change in the future but I'm not stressed.
+
+  Returns {:ok, %MediaItem{}} | {:error, atom(), String.t()} | {:recovered, %MediaItem{}, String.t()}
+  """
+  def download_for_media_item(%MediaItem{} = media_item, override_opts \\ []) do
+    case attempt_download_and_update_for_media_item(media_item, override_opts) do
+      {:ok, media_item} ->
+        # Returns {:ok, %MediaItem{}}
+        Media.update_media_item(media_item, %{last_error: nil})
+
+      {:error, error_atom, message} ->
+        Media.update_media_item(media_item, %{last_error: StringUtils.wrap_string(message)})
+
+        {:error, error_atom, message}
+
+      {:recovered, media_item, message} ->
+        {:ok, updated_media_item} = Media.update_media_item(media_item, %{last_error: StringUtils.wrap_string(message)})
+
+        {:recovered, updated_media_item, message}
+    end
+  end
+
+  # Looks complicated, but here's the key points:
+  # - download_with_options runs a pre-check to see if the media item is suitable for download.
+  # - If the media item fails the precheck, it returns {:error, :unsuitable_for_download, message}
+  #   - However, if the precheck fails in a way that we think can be fixed by using cookies, we retry with cookies
+  #     and return the result of that
+  # - If the precheck passes but the download fails, it normally returns {:error, :download_failed, message}
+  #   - However, there are some errors we can recover from (eg: failure to communicate with SponsorBlock).
+  #     In this case, we attempt the download anyway and update the media item with what details we do have.
+  #     This case returns {:recovered, updated_media_item, message}
+  #   - If we attempt a retry but it fails, we return {:error, :unrecoverable, message}
+  # - If there is an unknown error unrelated to the above, we return {:error, :unknown, message}
+  # - Finally, if there is no error, we update the media item with the parsed JSON and return {:ok, updated_media_item}
+  #
+  # Restated, here are the return values for each case:
+  # - On success: {:ok, updated_media_item}
+  # - On initial failure but successfully recovered: {:recovered, updated_media_item, message}
+  # - On error: {:error, error_atom, message} where error_atom is one of:
+  #   - `:unsuitable_for_download` if the media item fails the precheck
+  #   - `:unrecoverable` if there was an initial failure and the recovery attempt failed
+  #   - `:download_failed` for all other yt-dlp-related downloading errors
+  #   - `:unknown` for any other errors, including those not related to yt-dlp
+  # - If we retry using cookies, all of the above return values apply. The cookie retry
+  #   logic is handled transparently as far as the caller is concerned
+  defp attempt_download_and_update_for_media_item(media_item, override_opts) do
+    output_filepath = FilesystemUtils.generate_metadata_tmpfile(:json)
+    media_with_preloads = Repo.preload(media_item, [:metadata, source: :media_profile])
+
+    case download_with_options(media_item.original_url, media_with_preloads, output_filepath, override_opts) do
+      {:ok, parsed_json} ->
+        update_media_item_from_parsed_json(media_with_preloads, parsed_json)
+
+      {:error, :unsuitable_for_download} ->
+        message =
+          "Media item ##{media_with_preloads.id} isn't suitable for download yet. May be an active or processing live stream"
+
+        Logger.warning(message)
+
+        {:error, :unsuitable_for_download, message}
+
+      {:error, message, _exit_code} ->
+        Logger.error("yt-dlp download error for media item ##{media_with_preloads.id}: #{inspect(message)}")
+
+        if String.contains?(to_string(message), recoverable_errors()) do
+          attempt_recovery_from_error(media_with_preloads, output_filepath, message)
+        else
+          {:error, :download_failed, message}
+        end
+
+      err ->
+        Logger.error("Unknown error downloading media item ##{media_with_preloads.id}: #{inspect(err)}")
+
+        {:error, :unknown, "Unknown error: #{inspect(err)}"}
+    end
+  end
+
+  defp attempt_recovery_from_error(media_with_preloads, output_filepath, error_message) do
+    with {:ok, contents} <- File.read(output_filepath),
+         {:ok, parsed_json} <- Phoenix.json_library().decode(contents) do
+      Logger.info("""
+      Recovery from yt-dlp error seems possible. Updating media item ##{media_with_preloads.id}
+      with parsed JSON from partial download attempt. Full download will be re-attemted in future
+      anyway
+      """)
+
+      {:ok, updated_media_item} = update_media_item_from_parsed_json(media_with_preloads, parsed_json)
+      {:recovered, updated_media_item, error_message}
+    else
+      err ->
+        Logger.error("Unable to recover error for media item ##{media_with_preloads.id}: #{inspect(err)}")
+
+        {:error, :unrecoverable, error_message}
+    end
+  end
+
+  defp update_media_item_from_parsed_json(media_with_preloads, parsed_json) do
+    parsed_attrs =
+      parsed_json
+      |> MetadataParser.parse_for_media_item()
+      |> Map.merge(%{
+        media_downloaded_at: DateTime.utc_now(),
+        culled_at: nil,
+        nfo_filepath: determine_nfo_filepath(media_with_preloads, parsed_json),
+        metadata: %{
+          # IDEA: might be worth kicking off a job for this since thumbnail fetching
+          # could fail and I want to handle that in isolation
+          metadata_filepath: MetadataFileHelpers.compress_and_store_metadata_for(media_with_preloads, parsed_json),
+          thumbnail_filepath: MetadataFileHelpers.download_and_store_thumbnail_for(media_with_preloads)
+        }
+      })
+
+    # Don't forgor to use preloaded associations or updates to
+    # associations won't work!
+    Media.update_media_item(media_with_preloads, parsed_attrs)
+  end
+
+  defp determine_nfo_filepath(media_item, parsed_json) do
+    if media_item.source.media_profile.download_nfo do
+      filepath = Path.rootname(parsed_json["filepath"]) <> ".nfo"
+
+      NfoBuilder.build_and_store_for_media_item(filepath, parsed_json)
+    else
+      nil
+    end
+  end
+
+  defp download_with_options(url, item_with_preloads, output_filepath, override_opts) do
+    {:ok, options} = DownloadOptionBuilder.build(item_with_preloads, override_opts)
+    force_use_cookies = Keyword.get(override_opts, :force_use_cookies, false)
+    source_uses_cookies = Sources.use_cookies?(item_with_preloads.source, :downloading)
+    should_use_cookies = force_use_cookies || source_uses_cookies
+
+    runner_opts = [output_filepath: output_filepath, use_cookies: should_use_cookies]
+
+    case {YtDlpMedia.get_downloadable_status(url, use_cookies: should_use_cookies), should_use_cookies} do
+      {{:ok, :downloadable}, _} ->
+        YtDlpMedia.download(url, options, runner_opts)
+
+      {{:ok, :ignorable}, _} ->
+        {:error, :unsuitable_for_download}
+
+      {{:error, _message, _exit_code} = err, false} ->
+        # If there was an error and we don't have cookies, this method will retry with cookies
+        # if doing so would help AND the source allows. Otherwise, it will return the error as-is
+        maybe_retry_with_cookies(url, item_with_preloads, output_filepath, override_opts, err)
+
+      # This gets hit if cookies are enabled which, importantly, also covers the case where we
+      # retry a download with cookies and it fails again
+      {{:error, message, exit_code}, true} ->
+        {:error, message, exit_code}
+
+      {err, _} ->
+        err
+    end
+  end
+
+  defp maybe_retry_with_cookies(url, item_with_preloads, output_filepath, override_opts, err) do
+    {:error, message, _} = err
+    source = item_with_preloads.source
+    message_contains_cookie_error = String.contains?(to_string(message), recoverable_cookie_errors())
+
+    if Sources.use_cookies?(source, :error_recovery) && message_contains_cookie_error do
+      download_with_options(
+        url,
+        item_with_preloads,
+        output_filepath,
+        Keyword.put(override_opts, :force_use_cookies, true)
+      )
+    else
+      err
+    end
+  end
+
+  defp recoverable_errors do
+    [
+      "Unable to communicate with SponsorBlock"
+    ]
+  end
+
+  defp recoverable_cookie_errors do
+    [
+      "Sign in to confirm",
+      "This video is available to this channel's members"
+    ]
+  end
+end

lib/pinchflat/downloading/media_quality_upgrade_worker.ex

@@ -0,0 +1,33 @@
+defmodule Pinchflat.Downloading.MediaQualityUpgradeWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :media_fetching,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable, :executing]],
+    tags: ["media_item", "media_fetching", "show_in_dashboard"]
+
+  require Logger
+
+  alias Pinchflat.Media
+  alias Pinchflat.Downloading.MediaDownloadWorker
+
+  @doc """
+  Redownloads media items that are eligible for redownload for the purpose
+  of upgrading the quality of the media or improving things like sponsorblock
+  segments.
+
+  This worker is scheduled to run daily via the Oban Cron plugin
+  and it should run _after_ the retention worker.
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{}) do
+    upgradable_media = Media.list_upgradeable_media_items()
+    Logger.info("Redownloading #{length(upgradable_media)} media items")
+
+    Enum.each(upgradable_media, fn media_item ->
+      MediaDownloadWorker.kickoff_with_task(media_item, %{quality_upgrade?: true})
+    end)
+  end
+end

lib/pinchflat/downloading/media_retention_worker.ex

@@ -0,0 +1,75 @@
+defmodule Pinchflat.Downloading.MediaRetentionWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :local_data,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable, :executing]],
+    tags: ["media_item", "local_data"]
+
+  use Pinchflat.Media.MediaQuery
+
+  require Logger
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+
+  @doc """
+  Deletes media items that are past their retention date and prevents
+  them from being re-downloaded.
+
+  This worker is scheduled to run daily via the Oban Cron plugin.
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{}) do
+    cull_cullable_media_items()
+    delete_media_items_from_before_cutoff()
+
+    :ok
+  end
+
+  defp cull_cullable_media_items do
+    cullable_media =
+      MediaQuery.new()
+      |> MediaQuery.require_assoc(:source)
+      |> where(^MediaQuery.cullable())
+      |> Repo.all()
+
+    Logger.info("Culling #{length(cullable_media)} media items past their retention date")
+
+    Enum.each(cullable_media, fn media_item ->
+      # Setting `prevent_download` does what it says on the tin, but `culled_at` is purely informational.
+      # We don't actually do anything with that in terms of queries and it gets set to nil if the media item
+      # gets re-downloaded.
+      Media.delete_media_files(media_item, %{
+        prevent_download: true,
+        culled_at: DateTime.utc_now()
+      })
+    end)
+  end
+
+  # NOTE: Since this is a date and not a datetime, we can't add logic to have to-the-minute
+  # comparison like we can with retention periods. We can only compare to the day.
+  defp delete_media_items_from_before_cutoff do
+    deletable_media =
+      MediaQuery.new()
+      |> MediaQuery.require_assoc(:source)
+      |> where(^MediaQuery.deletable_based_on_source_cutoff())
+      |> Repo.all()
+
+    Logger.info("Deleting #{length(deletable_media)} media items that are from before the source cutoff")
+
+    Enum.each(deletable_media, fn media_item ->
+      # Note that I'm not setting `prevent_download` on the media_item here.
+      # That's because cutoff_date can easily change and it's a valid behavior to re-download older
+      # media items if the cutoff_date changes.
+      # Download is ultimately prevented because `MediaQuery.pending()` only returns media items
+      # from after the cutoff date (among other things), so it's not like the media will just immediately
+      # be re-downloaded.
+      Media.delete_media_files(media_item, %{
+        culled_at: DateTime.utc_now()
+      })
+    end)
+  end
+end

lib/pinchflat/downloading/output_path/base.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.RenderedString.Base do
+defmodule Pinchflat.Downloading.OutputPath.Base do
   @moduledoc """
   A base module for parsing rendered strings, designed as a macro to be used
   in other modules. See https://elixirforum.com/t/help-to-parse-a-template-with-nimbleparsec/47980
@@ -6,7 +6,7 @@ defmodule Pinchflat.RenderedString.Base do
   NOTE: if the needs here get any more complicated, look into using a Liquid
   template parser. No need to reinvent the wheel any more than I already have.
 
-  NOTE: this is effectively tested by the `Pinchflat.RenderedString.Parser`'s tests
+  NOTE: this is effectively tested by the `Pinchflat.Downloading.OutputPath.Parser`'s tests
   """
 
   defmacro __using__(_opts) do

lib/pinchflat/downloading/output_path/parser.ex

@@ -1,11 +1,11 @@
-defmodule Pinchflat.RenderedString.Parser do
+defmodule Pinchflat.Downloading.OutputPath.Parser do
   @moduledoc """
   Parses liquid-ish-style strings into a rendered string
 
   Used for turning filepath templates into real filepaths
   """
 
-  use Pinchflat.RenderedString.Base
+  use Pinchflat.Downloading.OutputPath.Base
 
   @doc """
   Parses a string into a rendered string, using the provided variables. Optionally

lib/pinchflat/downloading/output_path_builder.ex

@@ -1,21 +1,31 @@
-defmodule Pinchflat.Profiles.OutputPathBuilder do
+defmodule Pinchflat.Downloading.OutputPathBuilder do
   @moduledoc """
   Builds yt-dlp-friendly output paths for downloaded media
   """
 
-  alias Pinchflat.RenderedString.Parser, as: TemplateParser
+  alias Pinchflat.Downloading.OutputPath.Parser, as: TemplateParser
 
   @doc """
   Builds the actual final filepath from a given template. Optionally, you can pass in
   a map of additional options to be used in the template.
 
+  Custom options are recursively expanded _once_ so you can nest custom options
+  one-deep if needed.
+
   Translates liquid-style templates into yt-dlp-style templates,
   leaving yt-dlp syntax intact.
   """
   def build(template_string, additional_template_options \\ %{}) do
     combined_options = Map.merge(custom_yt_dlp_option_map(), additional_template_options)
 
-    TemplateParser.parse(template_string, combined_options, &identifier_fn/2)
+    expanded_options =
+      Enum.map(combined_options, fn {key, value} ->
+        {:ok, parse_result} = TemplateParser.parse(value, combined_options, &identifier_fn/2)
+
+        {key, parse_result}
+      end)
+
+    TemplateParser.parse(template_string, Map.new(expanded_options), &identifier_fn/2)
   end
 
   # The `nil` case simply wraps the identifier in yt-dlp-style syntax. This assumes that
@@ -39,7 +49,17 @@ defmodule Pinchflat.Profiles.OutputPathBuilder do
       # Individual parts of the upload date
       "upload_year" => "%(upload_date>%Y)S",
       "upload_month" => "%(upload_date>%m)S",
-      "upload_day" => "%(upload_date>%d)S"
+      "upload_day" => "%(upload_date>%d)S",
+      "upload_yyyy_mm_dd" => "%(upload_date>%Y-%m-%d)S",
+      "season_from_date" => "%(upload_date>%Y)S",
+      "season_episode_from_date" => "s%(upload_date>%Y)Se%(upload_date>%m%d)S",
+      "season_episode_index_from_date" => "s%(upload_date>%Y)Se%(upload_date>%m%d)S{{ media_upload_date_index }}",
+      "artist_name" => "%(artist,creator,uploader,uploader_id)S",
+      "static_season__episode_by_index" => "Season 1/s01e{{ media_playlist_index }}",
+      "static_season__episode_by_date" => "Season 1/s01e%(upload_date>%y%m%d)S",
+      "season_by_year__episode_by_date" => "Season %(upload_date>%Y)S/s%(upload_date>%Y)Se%(upload_date>%m%d)S",
+      "season_by_year__episode_by_date_and_index" =>
+        "Season %(upload_date>%Y)S/s%(upload_date>%Y)Se%(upload_date>%m%d)S{{ media_upload_date_index }}"
     }
   end
 end

lib/pinchflat/downloading/quality_option_builder.ex

@@ -0,0 +1,66 @@
+defmodule Pinchflat.Downloading.QualityOptionBuilder do
+  @moduledoc """
+  A standalone builder module for building quality-related options for yt-dlp to download media.
+
+  Currently exclusively used in DownloadOptionBuilder since this logic is too complex to just
+  place in the main module.
+  """
+
+  alias Pinchflat.Settings
+  alias Pinchflat.Profiles.MediaProfile
+
+  @doc """
+  Builds the quality-related options for yt-dlp to download media based on the given media profile
+
+  Includes things like container, preferred format/codec, and audio track options.
+  """
+  def build(%MediaProfile{preferred_resolution: :audio, media_container: container} = media_profile) do
+    acodec = Settings.get!(:audio_codec_preference)
+
+    [
+      :extract_audio,
+      format_sort: "+acodec:#{acodec}",
+      audio_format: container || "best",
+      format: build_format_string(media_profile)
+    ]
+  end
+
+  def build(%MediaProfile{preferred_resolution: resolution_atom, media_container: container} = media_profile) do
+    vcodec = Settings.get!(:video_codec_preference)
+    acodec = Settings.get!(:audio_codec_preference)
+    {resolution_string, _} = resolution_atom |> Atom.to_string() |> Integer.parse()
+
+    [
+      # Since Plex doesn't support reading metadata from MKV
+      remux_video: container || "mp4",
+      format_sort: "res:#{resolution_string},+codec:#{vcodec}:#{acodec}",
+      format: build_format_string(media_profile)
+    ]
+  end
+
+  defp build_format_string(%MediaProfile{preferred_resolution: :audio, audio_track: audio_track}) do
+    if audio_track do
+      "bestaudio[#{build_format_modifier(audio_track)}]/bestaudio/best"
+    else
+      "bestaudio/best"
+    end
+  end
+
+  defp build_format_string(%MediaProfile{audio_track: audio_track}) do
+    if audio_track do
+      "bestvideo+bestaudio[#{build_format_modifier(audio_track)}]/bestvideo*+bestaudio/best"
+    else
+      "bestvideo*+bestaudio/best"
+    end
+  end
+
+  # Reminder to self: this conflicts with `--extractor-args "youtube:lang=<LANG>"`
+  # since that will translate the format_notes as well, which means they may not match.
+  # At least that's what happens now - worth a re-check if I have to come back to this
+  defp build_format_modifier("original"), do: "format_note*=original"
+  defp build_format_modifier("default"), do: "format_note*='(default)'"
+  # This uses the carat to anchor the language to the beginning of the string
+  # since that's what's needed to match `en` to `en-US` and `en-GB`, etc. The user
+  # can always specify the full language code if they want.
+  defp build_format_modifier(language_code), do: "language^=#{language_code}"
+end

lib/pinchflat/fast_indexing/fast_indexing_helpers.ex

@@ -0,0 +1,109 @@
+defmodule Pinchflat.FastIndexing.FastIndexingHelpers do
+  @moduledoc """
+  Methods for performing fast indexing tasks and managing the fast indexing process.
+
+  Many of these methods are made to be kickoff or be consumed by workers.
+  """
+
+  require Logger
+
+  use Pinchflat.Media.MediaQuery
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.FastIndexing.YoutubeRss
+  alias Pinchflat.FastIndexing.YoutubeApi
+  alias Pinchflat.Downloading.DownloadingHelpers
+  alias Pinchflat.FastIndexing.FastIndexingWorker
+  alias Pinchflat.Downloading.DownloadOptionBuilder
+
+  alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+
+  @doc """
+  Kicks off a new fast indexing task for a source. This will delete any existing fast indexing
+  tasks for the source before starting a new one.
+
+  Returns {:ok, %Task{}}
+  """
+  def kickoff_indexing_task(%Source{} = source) do
+    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker", include_executing: true)
+    FastIndexingWorker.kickoff_with_task(source)
+  end
+
+  @doc """
+  Fetches new media IDs for a source from YT's API or RSS, indexes them, and kicks off downloading
+  tasks for any pending media items. See comments in `FastIndexingWorker` for more info on the
+  order of operations and how this fits into the indexing process.
+
+  Returns [%MediaItem{}] where each item is a new media item that was created _but not necessarily
+  downloaded_.
+  """
+  def index_and_kickoff_downloads(%Source{} = source) do
+    # The media_profile is needed to determine the quality options to _then_ determine a more
+    # accurate predicted filepath
+    source = Repo.preload(source, [:media_profile])
+
+    {:ok, media_ids} = get_recent_media_ids(source)
+    existing_media_items = list_media_items_by_media_id_for(source, media_ids)
+    new_media_ids = media_ids -- Enum.map(existing_media_items, & &1.media_id)
+
+    maybe_new_media_items =
+      Enum.map(new_media_ids, fn media_id ->
+        case create_media_item_from_media_id(source, media_id) do
+          {:ok, media_item} ->
+            DownloadingHelpers.kickoff_download_if_pending(media_item, priority: 0)
+            media_item
+
+          err ->
+            Logger.error("Error creating media item '#{media_id}' from URL: #{inspect(err)}")
+            nil
+        end
+      end)
+
+    # Pick up any stragglers. Intentionally has a lower priority than the per-media item
+    # kickoff above
+    DownloadingHelpers.enqueue_pending_download_tasks(source, priority: 1)
+
+    Enum.filter(maybe_new_media_items, & &1)
+  end
+
+  # If possible, use the YouTube API to fetch media IDs. If that fails, fall back to the RSS feed.
+  # If the YouTube API isn't set up, just use the RSS feed.
+  defp get_recent_media_ids(source) do
+    with true <- YoutubeApi.enabled?(),
+         {:ok, media_ids} <- YoutubeApi.get_recent_media_ids(source) do
+      {:ok, media_ids}
+    else
+      _ -> YoutubeRss.get_recent_media_ids(source)
+    end
+  end
+
+  defp list_media_items_by_media_id_for(source, media_ids) do
+    MediaQuery.new()
+    |> where(^dynamic([mi], ^MediaQuery.for_source(source) and mi.media_id in ^media_ids))
+    |> Repo.all()
+  end
+
+  defp create_media_item_from_media_id(source, media_id) do
+    url = "https://www.youtube.com/watch?v=#{media_id}"
+    # This is set to :metadata instead of :indexing since this happens _after_ the
+    # actual indexing process. In reality, slow indexing is the only thing that
+    # should be using :indexing.
+    should_use_cookies = Sources.use_cookies?(source, :metadata)
+
+    command_opts =
+      [output: DownloadOptionBuilder.build_output_path_for(source)] ++
+        DownloadOptionBuilder.build_quality_options_for(source)
+
+    case YtDlpMedia.get_media_attributes(url, command_opts, use_cookies: should_use_cookies) do
+      {:ok, media_attrs} ->
+        Media.create_media_item_from_backend_attrs(source, media_attrs)
+
+      err ->
+        err
+    end
+  end
+end

lib/pinchflat/fast_indexing/fast_indexing_worker.ex

@@ -0,0 +1,86 @@
+defmodule Pinchflat.FastIndexing.FastIndexingWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :fast_indexing,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
+    tags: ["media_source", "fast_indexing", "show_in_dashboard"]
+
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Tasks
+  alias Pinchflat.Media
+  alias Pinchflat.Sources
+  alias Pinchflat.Settings
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.FastIndexing.FastIndexingHelpers
+  alias Pinchflat.Lifecycle.Notifications.SourceNotifications
+
+  @doc """
+  Starts the source fast indexing worker and creates a task for the source.
+
+  Returns {:ok, %Task{}} | {:error, :duplicate_job} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff_with_task(source, opts \\ []) do
+    %{id: source.id}
+    |> FastIndexingWorker.new(opts)
+    |> Tasks.create_job_with_task(source)
+  end
+
+  @doc """
+  Similar to `MediaCollectionIndexingWorker`, but for working with RSS feeds.
+  `MediaCollectionIndexingWorker` should be preferred in general, but this is
+  useful for downloading small batches of media items via fast indexing.
+
+  Only kicks off downloads for media that _should_ be downloaded
+  (ie: the source is set to download and the media matches the profile's format preferences)
+
+  Order of operations:
+    1. FastIndexingWorker (this module) periodically checks the YouTube RSS feed for new media.
+       with `FastIndexingHelpers.index_and_kickoff_downloads`
+    2. If the above `index_and_kickoff_downloads` finds new media items in the RSS feed,
+       it indexes them with a yt-dlp call to create the media item records then kicks off downloading
+       tasks (MediaDownloadWorker) for any new media items _that should be downloaded_.
+    3. Once downloads are kicked off, this worker sends a notification to the apprise server if applicable
+       then reschedules itself to run again in the future.
+
+  Returns :ok | {:ok, :job_exists} | {:ok, %Task{}}
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+    source = Sources.get_source!(source_id)
+
+    if source.fast_index do
+      perform_indexing_and_send_notification(source)
+      reschedule_indexing(source)
+    else
+      :ok
+    end
+  rescue
+    Ecto.NoResultsError -> Logger.info("#{__MODULE__} discarded: source #{source_id} not found")
+    Ecto.StaleEntryError -> Logger.info("#{__MODULE__} discarded: source #{source_id} stale")
+  end
+
+  defp perform_indexing_and_send_notification(source) do
+    apprise_server = Settings.get!(:apprise_server)
+
+    new_media_items =
+      source
+      |> FastIndexingHelpers.index_and_kickoff_downloads()
+      |> Enum.filter(&Media.pending_download?(&1))
+
+    if source.download_media do
+      SourceNotifications.send_new_media_notification(apprise_server, source, length(new_media_items))
+    end
+  end
+
+  defp reschedule_indexing(source) do
+    next_run_in = Source.fast_index_frequency() * 60
+
+    case kickoff_with_task(source, schedule_in: next_run_in) do
+      {:ok, task} -> {:ok, task}
+      {:error, :duplicate_job} -> {:ok, :job_exists}
+    end
+  end
+end

lib/pinchflat/fast_indexing/youtube_api.ex

@@ -0,0 +1,131 @@
+defmodule Pinchflat.FastIndexing.YoutubeApi do
+  @moduledoc """
+  Methods for interacting with the YouTube API for fast indexing
+  """
+
+  require Logger
+
+  alias Pinchflat.Settings
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Utils.FunctionUtils
+  alias Pinchflat.FastIndexing.YoutubeBehaviour
+
+  @behaviour YoutubeBehaviour
+
+  @agent_name {:global, __MODULE__.KeyIndex}
+
+  @doc """
+  Determines if the YouTube API is enabled for fast indexing by checking
+  if the user has an API key set
+
+  Returns boolean()
+  """
+  @impl YoutubeBehaviour
+  def enabled?, do: Enum.any?(api_keys())
+
+  @doc """
+  Fetches the recent media IDs from the YouTube API for a given source.
+
+  Returns {:ok, [binary()]} | {:error, binary()}
+  """
+  @impl YoutubeBehaviour
+  def get_recent_media_ids(%Source{} = source) do
+    api_response =
+      source
+      |> determine_playlist_id()
+      |> do_api_request()
+
+    case api_response do
+      {:ok, parsed_json} -> get_media_ids_from_response(parsed_json)
+      {:error, reason} -> {:error, reason}
+    end
+  end
+
+  # The UC prefix is for channels which won't work with this API endpoint. Swapping
+  # the prefix to UU will get us the playlist that represents the channel's uploads
+  defp determine_playlist_id(%{collection_id: c_id}) do
+    String.replace_prefix(c_id, "UC", "UU")
+  end
+
+  defp do_api_request(playlist_id) do
+    Logger.debug("Fetching recent media IDs from YouTube API for playlist: #{playlist_id}")
+
+    playlist_id
+    |> construct_api_endpoint()
+    |> http_client().get(accept: "application/json")
+    |> case do
+      {:ok, response} ->
+        Phoenix.json_library().decode(response)
+
+      {:error, reason} ->
+        Logger.error("Failed to fetch YouTube API: #{inspect(reason)}")
+        {:error, reason}
+    end
+  end
+
+  defp get_media_ids_from_response(parsed_json) do
+    parsed_json
+    |> Map.get("items", [])
+    |> Enum.map(fn item ->
+      item
+      |> Map.get("contentDetails", %{})
+      |> Map.get("videoId", nil)
+    end)
+    |> Enum.reject(&is_nil/1)
+    |> Enum.uniq()
+    |> FunctionUtils.wrap_ok()
+  end
+
+  defp api_keys do
+    case Settings.get!(:youtube_api_key) do
+      nil ->
+        []
+
+      keys ->
+        keys
+        |> String.split(",")
+        |> Enum.map(&String.trim/1)
+        |> Enum.reject(&(&1 == ""))
+    end
+  end
+
+  defp get_or_start_api_key_agent do
+    case Agent.start(fn -> 0 end, name: @agent_name) do
+      {:ok, pid} -> pid
+      {:error, {:already_started, pid}} -> pid
+    end
+  end
+
+  # Gets the next API key in round-robin fashion
+  defp next_api_key do
+    keys = api_keys()
+
+    case keys do
+      [] ->
+        nil
+
+      keys ->
+        pid = get_or_start_api_key_agent()
+
+        current_index =
+          Agent.get_and_update(pid, fn current ->
+            {current, rem(current + 1, length(keys))}
+          end)
+
+        Logger.debug("Using YouTube API key: #{Enum.at(keys, current_index)}")
+        Enum.at(keys, current_index)
+    end
+  end
+
+  defp construct_api_endpoint(playlist_id) do
+    api_base = "https://youtube.googleapis.com/youtube/v3/playlistItems"
+    property_type = "contentDetails"
+    max_results = 50
+
+    "#{api_base}?part=#{property_type}&maxResults=#{max_results}&playlistId=#{playlist_id}&key=#{next_api_key()}"
+  end
+
+  defp http_client do
+    Application.get_env(:pinchflat, :http_client, Pinchflat.HTTP.HTTPClient)
+  end
+end

lib/pinchflat/fast_indexing/youtube_behaviour.ex

@@ -0,0 +1,11 @@
+defmodule Pinchflat.FastIndexing.YoutubeBehaviour do
+  @moduledoc """
+  This module defines the behaviour for clients that interface with YouTube
+  for the purpose of fast indexing.
+  """
+
+  alias Pinchflat.Sources.Source
+
+  @callback enabled?() :: boolean()
+  @callback get_recent_media_ids(%Source{}) :: {:ok, [String.t()]} | {:error, String.t()}
+end

lib/pinchflat/fast_indexing/youtube_rss.ex

@@ -1,18 +1,31 @@
-defmodule Pinchflat.Api.YoutubeRss do
+defmodule Pinchflat.FastIndexing.YoutubeRss do
   @moduledoc """
-  Methods for interacting with YouTube RSS feeds
+  Methods for interacting with YouTube RSS feeds for fast indexing
   """
 
   require Logger
 
   alias Pinchflat.Sources.Source
+  alias Pinchflat.FastIndexing.YoutubeBehaviour
+
+  @behaviour YoutubeBehaviour
+
+  @doc """
+  Determines if the YouTube RSS feed is enabled for fast indexing. Used to satisfy
+  the `YoutubeBehaviour` behaviour.
+
+  Returns true
+  """
+  @impl YoutubeBehaviour
+  def enabled?(), do: true
 
   @doc """
   Fetches the recent media IDs from a YouTube RSS feed for a given source.
 
   Returns {:ok, [binary()]} | {:error, binary()}
   """
-  def get_recent_media_ids_from_rss(%Source{} = source) do
+  @impl YoutubeBehaviour
+  def get_recent_media_ids(%Source{} = source) do
     Logger.debug("Fetching recent media IDs from YouTube RSS feed for source: #{source.collection_id}")
 
     case http_client().get(rss_url_for_source(source)) do

lib/pinchflat/http/http_client.ex

@@ -21,9 +21,11 @@ defmodule Pinchflat.HTTP.HTTPClient do
   """
   @impl HTTPBehaviour
   def get(url, headers \\ [], opts \\ []) do
+    headers = parse_headers(headers)
+
     case :httpc.request(:get, {url, headers}, [], opts) do
       {:ok, {{_version, 200, _reason_phrase}, _headers, body}} ->
-        {:ok, body}
+        {:ok, to_string(body)}
 
       {:ok, {{_version, status_code, reason_phrase}, _headers, _body}} ->
         {:error, "HTTP request failed with status code #{status_code}: #{reason_phrase}"}
@@ -32,4 +34,8 @@ defmodule Pinchflat.HTTP.HTTPClient do
         {:error, "HTTP request failed: #{reason}"}
     end
   end
+
+  defp parse_headers(headers) do
+    Enum.map(headers, fn {k, v} -> {to_charlist(k), to_charlist(v)} end)
+  end
 end

lib/pinchflat/lifecycle/notifications/apprise_command_runner.ex

@@ -0,0 +1,12 @@
+defmodule Pinchflat.Lifecycle.Notifications.AppriseCommandRunner do
+  @moduledoc """
+  A behaviour for running CLI commands against a notification backend (apprise).
+
+  Used so we can implement Mox for testing without actually running the
+  apprise command.
+  """
+
+  @callback run(binary(), keyword()) :: :ok | {:error, binary()}
+  @callback run(List.t(), keyword()) :: :ok | {:error, binary()}
+  @callback version() :: {:ok, binary()} | {:error, binary()}
+end

lib/pinchflat/lifecycle/notifications/command_runner.ex

@@ -0,0 +1,63 @@
+defmodule Pinchflat.Lifecycle.Notifications.CommandRunner do
+  @moduledoc """
+  Runs apprise commands using the `System.cmd/3` function
+  """
+
+  require Logger
+
+  alias Pinchflat.Utils.CliUtils
+  alias Pinchflat.Utils.FunctionUtils
+  alias Pinchflat.Lifecycle.Notifications.AppriseCommandRunner
+
+  @behaviour AppriseCommandRunner
+
+  @doc """
+  Runs an apprise command and returns the string output.
+  Can take a single server string or a list of servers as well as additional
+  arguments to pass to the command.
+
+  Returns {:ok, binary()} | {:error, :no_servers} | {:error, binary()}
+  """
+  @impl AppriseCommandRunner
+  def run(nil, _), do: {:error, :no_servers}
+  def run("", _), do: {:error, :no_servers}
+  def run([], _), do: {:error, :no_servers}
+
+  def run(endpoints, command_opts) do
+    endpoints = List.wrap(endpoints)
+    default_opts = [:verbose]
+    parsed_opts = CliUtils.parse_options(default_opts ++ command_opts)
+
+    {output, exit_code} = CliUtils.wrap_cmd(backend_executable(), parsed_opts ++ endpoints)
+
+    case exit_code do
+      0 -> {:ok, String.trim(output)}
+      _ -> {:error, String.trim(output)}
+    end
+  end
+
+  @doc """
+  Returns the version of apprise as a string.
+
+  Returns {:ok, binary()} | {:error, binary()}
+  """
+  @impl AppriseCommandRunner
+  def version do
+    case CliUtils.wrap_cmd(backend_executable(), ["--version"]) do
+      {output, 0} ->
+        output
+        |> String.split(~r{\r?\n})
+        |> List.first()
+        |> String.replace("Apprise", "")
+        |> String.trim()
+        |> FunctionUtils.wrap_ok()
+
+      {output, _} ->
+        {:error, output}
+    end
+  end
+
+  defp backend_executable do
+    Application.get_env(:pinchflat, :apprise_executable)
+  end
+end

lib/pinchflat/lifecycle/notifications/source_notifications.ex

@@ -0,0 +1,81 @@
+defmodule Pinchflat.Lifecycle.Notifications.SourceNotifications do
+  @moduledoc """
+  Contains utilities for sending notifications about sources
+  """
+
+  require Logger
+
+  use Pinchflat.Media.MediaQuery
+
+  alias Pinchflat.Repo
+
+  @doc """
+  Wraps a function that may change the number of pending  or downloaded
+  media items for a source, sending an apprise notification if
+  the count changes.
+
+  Returns the return value of the provided function
+  """
+  def wrap_new_media_notification(servers, source, func) do
+    before_count = relevant_media_item_count(source)
+    retval = func.()
+    after_count = relevant_media_item_count(source)
+
+    send_new_media_notification(servers, source, after_count - before_count)
+
+    retval
+  end
+
+  @doc """
+  Sends a notification if the count of new media items has changed
+
+  Returns :ok
+  """
+  def send_new_media_notification(_, _, count) when count <= 0, do: :ok
+
+  def send_new_media_notification(servers, source, changed_count) do
+    opts = [
+      title: "[Pinchflat] New media found",
+      body: "Found #{changed_count} new media item(s) for #{source.custom_name}. Downloading them now"
+    ]
+
+    case backend_runner().run(servers, opts) do
+      {:ok, _} ->
+        Logger.info("Sent new media notification for source #{source.id}")
+
+      {:error, :no_servers} ->
+        Logger.info("No notification servers provided for source #{source.id}")
+
+      {:error, err} ->
+        Logger.error("Failed to send new media notification for source #{source.id}: #{err}")
+    end
+
+    :ok
+  end
+
+  defp relevant_media_item_count(source) do
+    if source.download_media do
+      pending_media_item_count(source) + downloaded_media_item_count(source)
+    else
+      0
+    end
+  end
+
+  defp pending_media_item_count(source) do
+    MediaQuery.new()
+    |> MediaQuery.require_assoc(:media_profile)
+    |> where(^dynamic(^MediaQuery.for_source(source) and ^MediaQuery.pending()))
+    |> Repo.aggregate(:count)
+  end
+
+  defp downloaded_media_item_count(source) do
+    MediaQuery.new()
+    |> where(^dynamic(^MediaQuery.for_source(source) and ^MediaQuery.downloaded()))
+    |> Repo.aggregate(:count)
+  end
+
+  defp backend_runner do
+    # This approach lets us mock the command for testing
+    Application.get_env(:pinchflat, :apprise_runner)
+  end
+end

lib/pinchflat/lifecycle/user_scripts/command_runner.ex

@@ -0,0 +1,73 @@
+defmodule Pinchflat.Lifecycle.UserScripts.CommandRunner do
+  @moduledoc """
+  Runs custom user commands commands using the `System.cmd/3` function
+  """
+
+  require Logger
+
+  alias Pinchflat.Utils.CliUtils
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Lifecycle.UserScripts.UserScriptCommandRunner
+
+  @behaviour UserScriptCommandRunner
+
+  @event_types [
+    :app_init,
+    :media_pre_download,
+    :media_downloaded,
+    :media_deleted
+  ]
+
+  @doc """
+  Runs the user script command for the given event type. Passes the event
+  and the encoded data to the user script command.
+
+  This function will succeed in almost all cases, even if the user script command
+  failed - this is because I don't want bad scripts to stop the whole process.
+  If something fails, it'll be logged and returned BUT the tuple will always
+  start with {:ok, ...}.
+
+  The only things that can cause a true failure are passing in an invalid event
+  type or if the passed data cannot be encoded into JSON - both indicative of
+  failures in the development process.
+
+  Returns {:ok, :no_executable} | {:ok, output, exit_code}
+  """
+  @impl UserScriptCommandRunner
+  def run(event_type, encodable_data) when event_type in @event_types do
+    case backend_executable() do
+      {:ok, :no_executable} ->
+        {:ok, :no_executable}
+
+      {:ok, executable_path} ->
+        {:ok, encoded_data} = Phoenix.json_library().encode(encodable_data)
+
+        {output, exit_code} =
+          CliUtils.wrap_cmd(
+            executable_path,
+            [to_string(event_type), encoded_data],
+            [],
+            logging_arg_override: "[suppressed]"
+          )
+
+        {:ok, output, exit_code}
+    end
+  end
+
+  def run(event_type, _encodable_data) do
+    raise ArgumentError, "Invalid event type: #{inspect(event_type)}"
+  end
+
+  defp backend_executable do
+    base_dir = Application.get_env(:pinchflat, :extras_directory)
+    filepath = Path.join([base_dir, "user-scripts", "lifecycle"])
+
+    if FilesystemUtils.exists_and_nonempty?(filepath) do
+      {:ok, filepath}
+    else
+      Logger.info("User scripts lifecyle file either not present or is empty. Skipping.")
+
+      {:ok, :no_executable}
+    end
+  end
+end

lib/pinchflat/lifecycle/user_scripts/user_script_command_runner.ex

@@ -0,0 +1,10 @@
+defmodule Pinchflat.Lifecycle.UserScripts.UserScriptCommandRunner do
+  @moduledoc """
+  A behaviour for running custom user scripts on certain events.
+
+  Used so we can implement Mox for testing without actually running the
+  user's command.
+  """
+
+  @callback run(atom(), map()) :: :ok | {:error, binary()}
+end

lib/pinchflat/media.ex

@@ -1,307 +0,0 @@
-defmodule Pinchflat.Media do
-  @moduledoc """
-  The Media context.
-  """
-
-  import Ecto.Query, warn: false
-
-  alias Pinchflat.Repo
-  alias Pinchflat.Tasks
-  alias Pinchflat.Media.MediaItem
-  alias Pinchflat.Sources.Source
-  alias Pinchflat.Media.MediaMetadata
-
-  @doc """
-  Returns the list of media_items.
-
-  Returns [%MediaItem{}, ...].
-  """
-  def list_media_items do
-    Repo.all(MediaItem)
-  end
-
-  @doc """
-  Returns a list of media_items for a given source.
-
-  Returns [%MediaItem{}, ...].
-  """
-  def list_media_items_for(%Source{} = source) do
-    MediaItem
-    |> where([mi], mi.source_id == ^source.id)
-    |> Repo.all()
-  end
-
-  @doc """
-  Fetches all media items belonging to a given source that have a media_id in the given list.
-  Useful for determining the what media items we DON'T already have for fast indexing.
-
-  NOTE: These queries are getting a little tedious. When I have the time, I should see about
-  implementing a query pattern and having these compose queries from a common base. This would
-  also let me compose simple queries in the module using them for one-off methods
-
-  Returns [%MediaItem{}, ...].
-  """
-  def list_media_items_by_media_id_for(%Source{} = source, media_ids) do
-    MediaItem
-    |> where([mi], mi.source_id == ^source.id and mi.media_id in ^media_ids)
-    |> Repo.all()
-  end
-
-  @doc """
-  Returns a list of pending media_items for a given source, where
-  pending means the `media_filepath` is `nil` AND the media_item
-  matches the format selection rules of the parent media_profile.
-
-  See `build_format_clauses` but tl;dr is it _may_ filter based
-  on shorts or livestreams depending on the media_profile settings.
-
-  Returns [%MediaItem{}, ...].
-  """
-  def list_pending_media_items_for(%Source{} = source, opts \\ []) do
-    limit = Keyword.get(opts, :limit, nil)
-    media_profile = Repo.preload(source, :media_profile).media_profile
-
-    MediaItem
-    |> where([mi], mi.source_id == ^source.id and is_nil(mi.media_filepath))
-    |> where(^build_format_clauses(media_profile))
-    |> where(^maybe_apply_cutoff_date(source))
-    |> Repo.maybe_limit(limit)
-    |> Repo.all()
-  end
-
-  @doc """
-  Returns a list of downloaded media_items for a given source.
-
-  Returns [%MediaItem{}, ...].
-  """
-  def list_downloaded_media_items_for(%Source{} = source, opts \\ []) do
-    limit = Keyword.get(opts, :limit, nil)
-
-    MediaItem
-    |> where([mi], mi.source_id == ^source.id and not is_nil(mi.media_filepath))
-    |> Repo.maybe_limit(limit)
-    |> Repo.all()
-  end
-
-  @doc """
-  For a given media_item, tells you if it is pending download. This is defined as
-  the media_item having a `media_filepath` of `nil` and matching the format selection
-  rules of the parent media_profile.
-
-  Intentionally does not take the `download_media` setting of the source into account.
-
-  Returns boolean()
-  """
-  def pending_download?(%MediaItem{} = media_item) do
-    media_item = Repo.preload(media_item, source: :media_profile)
-
-    MediaItem
-    |> where([mi], mi.id == ^media_item.id and is_nil(mi.media_filepath))
-    |> where(^build_format_clauses(media_item.source.media_profile))
-    |> where(^maybe_apply_cutoff_date(media_item.source))
-    |> Repo.exists?()
-  end
-
-  @doc """
-  Returns a list of media_items that match the search term. Adds a `matching_search_term`
-  virtual field to the result set.
-
-  Has explit handling for blank search terms because SQLite doesn't like empty MATCH clauses.
-
-  Returns [%MediaItem{}, ...].
-  """
-  def search(_search_term, _opts \\ [])
-  def search("", _opts), do: []
-  def search(nil, _opts), do: []
-
-  def search(search_term, opts) do
-    limit = Keyword.get(opts, :limit, 50)
-
-    from(mi in MediaItem,
-      join: mi_search_index in assoc(mi, :media_items_search_index),
-      where: fragment("media_items_search_index MATCH ?", ^search_term),
-      select_merge: %{
-        matching_search_term:
-          fragment("""
-            coalesce(snippet(media_items_search_index, 0, '[PF_HIGHLIGHT]', '[/PF_HIGHLIGHT]', '...', 20), '') ||
-            ' ' ||
-            coalesce(snippet(media_items_search_index, 1, '[PF_HIGHLIGHT]', '[/PF_HIGHLIGHT]', '...', 20), '')
-          """)
-      },
-      order_by: [desc: fragment("rank")],
-      limit: ^limit
-    )
-    |> Repo.all()
-  end
-
-  @doc """
-  Gets a single media_item.
-
-  Returns %MediaItem{}. Raises `Ecto.NoResultsError` if the Media item does not exist.
-  """
-  def get_media_item!(id), do: Repo.get!(MediaItem, id)
-
-  @doc """
-  Produces a flat list of the filesystem paths for a media_item's downloaded files
-
-  Returns [binary()]
-  """
-  def media_filepaths(media_item) do
-    mapped_struct = Map.from_struct(media_item)
-
-    MediaItem.filepath_attributes()
-    |> Enum.map(fn
-      :subtitle_filepaths = field -> Enum.map(mapped_struct[field], fn [_, filepath] -> filepath end)
-      field -> List.wrap(mapped_struct[field])
-    end)
-    |> List.flatten()
-    |> Enum.filter(&is_binary/1)
-  end
-
-  @doc """
-  Produces a flat list of the filesystem paths for a media_item's metadata files.
-  Returns an empty list if the media_item has no metadata.
-
-  Returns [binary()] | []
-  """
-  def metadata_filepaths(media_item) do
-    metadata = Repo.preload(media_item, :metadata).metadata || %MediaMetadata{}
-    mapped_struct = Map.from_struct(metadata)
-
-    MediaMetadata.filepath_attributes()
-    |> Enum.map(fn field -> mapped_struct[field] end)
-    |> Enum.filter(&is_binary/1)
-  end
-
-  @doc """
-  Creates a media_item.
-
-  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
-  """
-  def create_media_item(attrs) do
-    %MediaItem{}
-    |> MediaItem.changeset(attrs)
-    |> Repo.insert()
-  end
-
-  @doc """
-  Creates a media item from the attributes returned by the video backend
-  (read: yt-dlp).
-
-  Unlike `create_media_item`, this will attempt an update if the media_item
-  already exists. This is so that future indexing can pick up attributes that
-  we may not have asked for in the past (eg: upload_date)
-
-  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
-  """
-  def create_media_item_from_backend_attrs(source, media_attrs_struct) do
-    attrs = Map.merge(%{source_id: source.id}, Map.from_struct(media_attrs_struct))
-
-    %MediaItem{}
-    |> MediaItem.changeset(attrs)
-    |> Repo.insert(
-      on_conflict: [
-        set: Map.to_list(attrs)
-      ],
-      conflict_target: [:source_id, :media_id]
-    )
-  end
-
-  @doc """
-  Updates a media_item.
-
-  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
-  """
-  def update_media_item(%MediaItem{} = media_item, attrs) do
-    media_item
-    |> MediaItem.changeset(attrs)
-    |> Repo.update()
-  end
-
-  @doc """
-  Deletes a media_item and its associated tasks.
-  Can optionally delete the media_item's files.
-
-  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
-  """
-  def delete_media_item(%MediaItem{} = media_item, opts \\ []) do
-    delete_files = Keyword.get(opts, :delete_files, false)
-
-    if delete_files do
-      {:ok, _} = delete_all_attachments(media_item)
-    end
-
-    Tasks.delete_tasks_for(media_item)
-    Repo.delete(media_item)
-  end
-
-  @doc """
-  Returns an `%Ecto.Changeset{}` for tracking media_item changes.
-  """
-  def change_media_item(%MediaItem{} = media_item, attrs \\ %{}) do
-    MediaItem.changeset(media_item, attrs)
-  end
-
-  defp delete_all_attachments(media_item) do
-    media_item = Repo.preload(media_item, :metadata)
-
-    media_item
-    |> media_filepaths()
-    |> Enum.concat(metadata_filepaths(media_item))
-    |> Enum.each(&File.rm/1)
-
-    # rmdir will attempt to delete the directory, but only if it is empty
-    if media_item.media_filepath do
-      File.rmdir(Path.dirname(media_item.media_filepath))
-    end
-
-    if media_item.metadata && media_item.metadata.metadata_filepath do
-      File.rmdir(Path.dirname(media_item.metadata.metadata_filepath))
-    end
-
-    {:ok, media_item}
-  end
-
-  defp maybe_apply_cutoff_date(source) do
-    if source.download_cutoff_date do
-      dynamic([mi], mi.upload_date >= ^source.download_cutoff_date)
-    else
-      dynamic(true)
-    end
-  end
-
-  defp build_format_clauses(media_profile) do
-    mapped_struct = Map.from_struct(media_profile)
-
-    Enum.reduce(mapped_struct, dynamic(true), fn attr, dynamic ->
-      case {attr, media_profile} do
-        {{:shorts_behaviour, :only}, %{livestream_behaviour: :only}} ->
-          dynamic(
-            [mi],
-            ^dynamic and (mi.livestream == true or mi.short_form_content == true)
-          )
-
-        # Technically redundant, but makes the other clauses easier to parse
-        # (redundant because this condition is the same as the condition above, just flipped)
-        {{:livestream_behaviour, :only}, %{shorts_behaviour: :only}} ->
-          dynamic
-
-        {{:shorts_behaviour, :only}, _} ->
-          dynamic([mi], ^dynamic and mi.short_form_content == true)
-
-        {{:livestream_behaviour, :only}, _} ->
-          dynamic([mi], ^dynamic and mi.livestream == true)
-
-        {{:shorts_behaviour, :exclude}, %{livestream_behaviour: lb}} when lb != :only ->
-          dynamic([mi], ^dynamic and mi.short_form_content == false)
-
-        {{:livestream_behaviour, :exclude}, %{shorts_behaviour: sb}} when sb != :only ->
-          # return records with livestream: false
-          dynamic([mi], ^dynamic and mi.livestream == false)
-
-        _ ->
-          dynamic
-      end
-    end)
-  end
-end

lib/pinchflat/media/file_syncing.ex

@@ -0,0 +1,93 @@
+defmodule Pinchflat.Media.FileSyncing do
+  @moduledoc """
+  Functions for ensuring file state is accurately reflected in the database.
+  """
+
+  alias Pinchflat.Media
+  alias Pinchflat.Utils.MapUtils
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.Utils.FilesystemUtils, as: FSUtils
+
+  @doc """
+  Deletes files that are no longer needed by a media item.
+
+  This means that if a media item has been updated, the old and new versions
+  can be passed and any files that are no longer needed will be deleted.
+
+  An example is a video that gets its quality upgraded and its name changes
+  between original download and re-download. The old file will exist on-disk
+  with the old name but the database entry will point to the new file. This
+  function can be used to delete the old file in this case.
+
+  Returns :ok
+  """
+  def delete_outdated_files(old_media_item, new_media_item) do
+    non_subtitle_keys = MediaItem.filepath_attributes() -- [:subtitle_filepaths]
+
+    old_non_subtitles = Map.take(old_media_item, non_subtitle_keys)
+    old_subtitles = MapUtils.from_nested_list(old_media_item.subtitle_filepaths)
+    new_non_subtitles = Map.take(new_media_item, non_subtitle_keys)
+    new_subtitles = MapUtils.from_nested_list(new_media_item.subtitle_filepaths)
+
+    handle_file_deletion(old_non_subtitles, new_non_subtitles)
+    handle_file_deletion(old_subtitles, new_subtitles)
+
+    :ok
+  end
+
+  @doc """
+  Nillifies any media item filepaths that don't exist on disk for a list of media items
+
+  returns [%MediaItem{}]
+  """
+  def sync_file_presence_on_disk(media_items) do
+    Enum.map(media_items, fn media_item ->
+      new_attributes = sync_media_item_files(media_item)
+      # Doing this one-by-one instead of batching since this process
+      # can take time and a batch could let MediaItem state get out of sync
+      {:ok, updated_media_item} = Media.update_media_item(media_item, new_attributes)
+
+      updated_media_item
+    end)
+  end
+
+  defp handle_file_deletion(old_attributes, new_attributes) do
+    # The logic:
+    #   - A file should only be deleted if it exists and the new file is different
+    #   - The new attributes are the ones we're interested in keeping
+    #   - If the old attributes have a key that doesn't exist in the new attributes, don't touch it.
+    #     This is good for archiving but may be unpopular for other users so this may change.
+
+    Enum.each(new_attributes, fn {key, new_filepath} ->
+      old_filepath = Map.get(old_attributes, key)
+      files_have_changed = old_filepath && new_filepath && old_filepath != new_filepath
+      files_exist_on_disk = files_have_changed && File.exists?(old_filepath) && File.exists?(new_filepath)
+
+      if files_exist_on_disk && !FSUtils.filepaths_reference_same_file?(old_filepath, new_filepath) do
+        FSUtils.delete_file_and_remove_empty_directories(old_filepath)
+      end
+    end)
+  end
+
+  defp sync_media_item_files(media_item) do
+    non_subtitle_keys = MediaItem.filepath_attributes() -- [:subtitle_filepaths]
+    subtitle_keys = MapUtils.from_nested_list(media_item.subtitle_filepaths)
+    non_subtitles = Map.take(media_item, non_subtitle_keys)
+
+    # This one is checking for the negative (ie: only update if the file doesn't exist)
+    new_non_subtitle_attrs =
+      Enum.reduce(non_subtitles, %{}, fn {key, filepath}, acc ->
+        if filepath && File.exists?(filepath), do: acc, else: Map.put(acc, key, nil)
+      end)
+
+    # This one is checking for the positive (ie: only update if the file exists)
+    # This is because subtitles, being an array type in the DB, are most easily updated
+    # by a full replacement rather than finding the actual diff
+    new_subtitle_attrs =
+      Enum.reduce(subtitle_keys, [], fn {key, filepath}, acc ->
+        if filepath && File.exists?(filepath), do: acc ++ [[key, filepath]], else: acc
+      end)
+
+    Map.put(new_non_subtitle_attrs, :subtitle_filepaths, new_subtitle_attrs)
+  end
+end

lib/pinchflat/media/file_syncing_worker.ex

@@ -0,0 +1,38 @@
+defmodule Pinchflat.Media.FileSyncingWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :local_data,
+    tags: ["sources", "local_data"]
+
+  alias __MODULE__
+  alias Pinchflat.Repo
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Media.FileSyncing
+
+  @doc """
+  Starts the source file syncing worker.
+
+  Returns {:ok, %Task{}} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff_with_task(source, opts \\ []) do
+    %{id: source.id}
+    |> FileSyncingWorker.new(opts)
+    |> Tasks.create_job_with_task(source)
+  end
+
+  @doc """
+  Deletes a profile and optionally deletes its files
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+    source = Repo.preload(Sources.get_source!(source_id), :media_items)
+
+    FileSyncing.sync_file_presence_on_disk(source.media_items)
+
+    :ok
+  end
+end

lib/pinchflat/media/media.ex

@@ -0,0 +1,242 @@
+defmodule Pinchflat.Media do
+  @moduledoc """
+  The Media context.
+  """
+
+  import Ecto.Query, warn: false
+  use Pinchflat.Media.MediaQuery
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Metadata.MediaMetadata
+
+  alias Pinchflat.Lifecycle.UserScripts.CommandRunner, as: UserScriptRunner
+
+  # Some fields should only be set on insert and not on update.
+  @fields_to_drop_on_update [:playlist_index]
+
+  @doc """
+  Returns the list of media_items.
+
+  Returns [%MediaItem{}, ...].
+  """
+  def list_media_items do
+    Repo.all(MediaItem)
+  end
+
+  @doc """
+  Returns a list of media_items that are upgradeable based on the redownload delay
+  of the media_profile their source belongs to. In this context, upgradeable means
+  that it's been long enough since upload that the video may be in a higher quality
+  or have better sponsorblock segments (or similar).
+
+  The logic is that a media_item is past_redownload_delay if the media_item's uploaded_at is
+  at least redownload_delay_days ago AND `media_downloaded_at` - `redownload_delay_days`
+  is before the media_item's `uploaded_at`.
+
+  This logic grabs media that we've recently downloaded AND is recently uploaded, but
+  doesn't grab media that we've recently downloaded and was uploaded a long time ago.
+  This also makes things work as expected when downloading media from a source for the
+  first time.
+
+  Returns [%MediaItem{}, ...]
+  """
+  def list_upgradeable_media_items do
+    MediaQuery.new()
+    |> MediaQuery.require_assoc(:media_profile)
+    |> where(^MediaQuery.upgradeable())
+    |> Repo.all()
+  end
+
+  @doc """
+  Returns a list of pending media_items for a given source, where
+  pending means the media_item satisfies `MediaQuery.pending`. You
+  should really check out that function if you need to know more
+  because it has a lot going on.
+
+  Returns [%MediaItem{}, ...].
+  """
+  def list_pending_media_items_for(%Source{} = source) do
+    MediaQuery.new()
+    |> MediaQuery.require_assoc(:media_profile)
+    |> where(^dynamic(^MediaQuery.for_source(source) and ^MediaQuery.pending()))
+    |> Repo.all()
+  end
+
+  @doc """
+  For a given media_item, tells you if it is pending download. This is defined as
+  the media_item satisfying `MediaQuery.pending` which you should really check out.
+
+  Intentionally does not take the `download_media` setting of the source into account.
+
+  Returns boolean()
+  """
+  def pending_download?(%MediaItem{} = media_item) do
+    media_item = Repo.preload(media_item, source: :media_profile)
+
+    MediaQuery.new()
+    |> MediaQuery.require_assoc(:media_profile)
+    |> where(^dynamic([m, s, mp], m.id == ^media_item.id and ^MediaQuery.pending()))
+    |> Repo.exists?()
+  end
+
+  @doc """
+  Returns a list of media_items that match the search term. Adds a `matching_search_term`
+  virtual field to the result set.
+
+  Has explit handling for blank search terms because SQLite doesn't like empty MATCH clauses.
+
+  Returns [%MediaItem{}, ...].
+  """
+  def search(_search_term, _opts \\ [])
+  def search("", _opts), do: []
+  def search(nil, _opts), do: []
+
+  def search(search_term, opts) do
+    limit = Keyword.get(opts, :limit, 50)
+
+    MediaQuery.new()
+    |> MediaQuery.matching_search_term(search_term)
+    |> Repo.maybe_limit(limit)
+    |> Repo.all()
+  end
+
+  @doc """
+  Gets a single media_item.
+
+  Returns %MediaItem{}. Raises `Ecto.NoResultsError` if the Media item does not exist.
+  """
+  def get_media_item!(id), do: Repo.get!(MediaItem, id)
+
+  @doc """
+  Creates a media_item.
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def create_media_item(attrs) do
+    %MediaItem{}
+    |> MediaItem.changeset(attrs)
+    |> Repo.insert()
+  end
+
+  @doc """
+  Creates a media item from the attributes returned by the video backend
+  (read: yt-dlp).
+
+  Unlike `create_media_item`, this will attempt an update if the media_item
+  already exists. This is so that future indexing can pick up attributes that
+  we may not have asked for in the past (eg: uploaded_at)
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def create_media_item_from_backend_attrs(source, media_attrs_struct) do
+    attrs = Map.merge(%{source_id: source.id}, Map.from_struct(media_attrs_struct))
+
+    %MediaItem{}
+    |> MediaItem.changeset(attrs)
+    |> Repo.insert(
+      on_conflict: [
+        set:
+          attrs
+          |> Map.drop(@fields_to_drop_on_update)
+          |> Map.to_list()
+      ],
+      conflict_target: [:source_id, :media_id]
+    )
+  end
+
+  @doc """
+  Updates a media_item.
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def update_media_item(%MediaItem{} = media_item, attrs) do
+    update_attrs = Map.drop(attrs, @fields_to_drop_on_update)
+
+    media_item
+    |> MediaItem.changeset(update_attrs)
+    |> Repo.update()
+  end
+
+  @doc """
+  Deletes a media_item, its associated tasks, and our internal metadata files.
+  Can optionally delete the media_item's media files (media, thumbnail, subtitles, etc).
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def delete_media_item(%MediaItem{} = media_item, opts \\ []) do
+    delete_files = Keyword.get(opts, :delete_files, false)
+
+    Tasks.delete_tasks_for(media_item)
+
+    if delete_files do
+      {:ok, _} = do_delete_media_files(media_item)
+      run_user_script(:media_deleted, media_item)
+    end
+
+    # Should delete these no matter what
+    delete_internal_metadata_files(media_item)
+    Repo.delete(media_item)
+  end
+
+  @doc """
+  Deletes the tasks and media files associated with a media_item but leaves the
+  media_item in the database. Does not delete anything to do with associated metadata.
+
+  Optionally accepts a second argument `addl_attrs` which will be merged into the
+  media_item before it is updated. Useful for setting things like `prevent_download`
+  and `culled_at`, if wanted
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def delete_media_files(%MediaItem{} = media_item, addl_attrs \\ %{}) do
+    filepath_attrs = MediaItem.filepath_attribute_defaults()
+
+    Tasks.delete_tasks_for(media_item)
+    {:ok, _} = do_delete_media_files(media_item)
+    run_user_script(:media_deleted, media_item)
+
+    update_media_item(media_item, Map.merge(filepath_attrs, addl_attrs))
+  end
+
+  @doc """
+  Returns an `%Ecto.Changeset{}` for tracking media_item changes.
+  """
+  def change_media_item(%MediaItem{} = media_item, attrs \\ %{}) do
+    MediaItem.changeset(media_item, attrs)
+  end
+
+  defp do_delete_media_files(media_item) do
+    mapped_struct = Map.from_struct(media_item)
+
+    MediaItem.filepath_attributes()
+    |> Enum.map(fn
+      :subtitle_filepaths = field -> Enum.map(mapped_struct[field], fn [_, filepath] -> filepath end)
+      field -> List.wrap(mapped_struct[field])
+    end)
+    |> List.flatten()
+    |> Enum.filter(&is_binary/1)
+    |> Enum.each(&FilesystemUtils.delete_file_and_remove_empty_directories/1)
+
+    {:ok, media_item}
+  end
+
+  defp delete_internal_metadata_files(media_item) do
+    metadata = Repo.preload(media_item, :metadata).metadata || %MediaMetadata{}
+    mapped_struct = Map.from_struct(metadata)
+
+    MediaMetadata.filepath_attributes()
+    |> Enum.map(fn field -> mapped_struct[field] end)
+    |> Enum.filter(&is_binary/1)
+    |> Enum.each(&FilesystemUtils.delete_file_and_remove_empty_directories/1)
+  end
+
+  defp run_user_script(event, media_item) do
+    runner = Application.get_env(:pinchflat, :user_script_runner, UserScriptRunner)
+
+    runner.run(event, media_item)
+  end
+end

lib/pinchflat/media/media_item.ex

@@ -4,14 +4,22 @@ defmodule Pinchflat.Media.MediaItem do
   """
 
   use Ecto.Schema
-  import Ecto.Changeset
+  use Pinchflat.Media.MediaQuery
 
+  import Ecto.Changeset
+  import Pinchflat.Utils.ChangesetUtils
+
+  alias __MODULE__
+  alias Pinchflat.Repo
+  alias Pinchflat.Sources
   alias Pinchflat.Tasks.Task
   alias Pinchflat.Sources.Source
-  alias Pinchflat.Media.MediaMetadata
-  alias Pinchflat.Media.MediaItemSearchIndex
+  alias Pinchflat.Metadata.MediaMetadata
+  alias Pinchflat.Media.MediaItemsSearchIndex
 
   @allowed_fields [
+    # these fields are only captured on index
+    :playlist_index,
     # these fields are captured on indexing (and again on download)
     :title,
     :media_id,
@@ -20,27 +28,43 @@ defmodule Pinchflat.Media.MediaItem do
     :livestream,
     :source_id,
     :short_form_content,
-    :upload_date,
+    :uploaded_at,
+    :upload_date_index,
+    :duration_seconds,
+    :predicted_media_filepath,
     # these fields are captured only on download
     :media_downloaded_at,
     :media_filepath,
     :media_size_bytes,
     :subtitle_filepaths,
     :thumbnail_filepath,
-    :metadata_filepath
+    :metadata_filepath,
+    :nfo_filepath,
+    :last_error,
+    # These are user or system controlled fields
+    :prevent_download,
+    :prevent_culling,
+    :culled_at,
+    :media_redownloaded_at
   ]
   # Pretty much all the fields captured at index are required.
   @required_fields ~w(
+    uuid
     title
     original_url
     livestream
     media_id
     source_id
-    upload_date
+    uploaded_at
     short_form_content
-    )a
+  )a
 
   schema "media_items" do
+    # This is _not_ used as the primary key or internally in the database
+    # relations. This is only used to prevent an enumeration attack on the streaming
+    # and RSS feed endpoints since those _must_ be public (ie: no basic auth)
+    field :uuid, Ecto.UUID
+
     field :title, :string
     field :media_id, :string
     field :description, :string
@@ -48,23 +72,34 @@ defmodule Pinchflat.Media.MediaItem do
     field :livestream, :boolean, default: false
     field :short_form_content, :boolean, default: false
     field :media_downloaded_at, :utc_datetime
-    field :upload_date, :date
+    field :media_redownloaded_at, :utc_datetime
+    field :uploaded_at, :utc_datetime
+    field :upload_date_index, :integer, default: 0
+    field :duration_seconds, :integer
+    field :playlist_index, :integer, default: 0
 
+    field :predicted_media_filepath, :string
     field :media_filepath, :string
     field :media_size_bytes, :integer
     field :thumbnail_filepath, :string
     field :metadata_filepath, :string
+    field :nfo_filepath, :string
     # This is an array of [iso-2 language, filepath] pairs. Probably could
     # be an associated record, but I don't see the benefit right now.
     # Will very likely revisit because I can't leave well-enough alone.
     field :subtitle_filepaths, {:array, {:array, :string}}, default: []
 
+    field :last_error, :string
+    field :prevent_download, :boolean, default: false
+    field :prevent_culling, :boolean, default: false
+    field :culled_at, :utc_datetime
+
     field :matching_search_term, :string, virtual: true
 
     belongs_to :source, Source
 
     has_one :metadata, MediaMetadata, on_replace: :update
-    has_one :media_items_search_index, MediaItemSearchIndex, foreign_key: :id
+    has_one :media_items_search_index, MediaItemsSearchIndex, foreign_key: :id
 
     has_many :tasks, Task
 
@@ -76,12 +111,88 @@ defmodule Pinchflat.Media.MediaItem do
     media_item
     |> cast(attrs, @allowed_fields)
     |> cast_assoc(:metadata, with: &MediaMetadata.changeset/2, required: false)
+    |> dynamic_default(:uuid, fn _ -> Ecto.UUID.generate() end)
+    |> update_upload_date_index()
     |> validate_required(@required_fields)
+    # Validate that the title does NOT start with "youtube video #" since that indicates a restriction by YouTube.
+    # See issue #549 for more information.
+    |> validate_format(:title, ~r/^(?!youtube video #)/)
     |> unique_constraint([:media_id, :source_id])
   end
 
   @doc false
   def filepath_attributes do
-    ~w(media_filepath thumbnail_filepath metadata_filepath subtitle_filepaths)a
+    ~w(media_filepath thumbnail_filepath metadata_filepath subtitle_filepaths nfo_filepath)a
+  end
+
+  @doc false
+  def filepath_attribute_defaults do
+    filepath_attributes()
+    |> Enum.map(fn
+      :subtitle_filepaths -> {:subtitle_filepaths, []}
+      field -> {field, nil}
+    end)
+    |> Enum.into(%{})
+  end
+
+  @doc false
+  def json_exluded_fields do
+    ~w(__meta__ __struct__ metadata tasks media_items_search_index)a
+  end
+
+  # Run it on new records no matter what. The method we delegate to
+  # will handle the case where `uploaded_at` is `nil`
+  defp update_upload_date_index(%{data: %{id: nil}} = changeset) do
+    do_update_upload_date_index(changeset)
+  end
+
+  # For the update case, we only want to recalculate if the day itself has changed.
+  # For instance, this is useful in the migration from `upload_date` to `uploaded_at`
+  defp update_upload_date_index(%{changes: changes} = changeset) when is_map_key(changes, :uploaded_at) do
+    old_uploaded_at = changeset.data.uploaded_at
+    new_uploaded_at = get_change(changeset, :uploaded_at)
+    upload_dates_match = DateTime.to_date(old_uploaded_at) == DateTime.to_date(new_uploaded_at)
+
+    if upload_dates_match do
+      changeset
+    else
+      do_update_upload_date_index(changeset)
+    end
+  end
+
+  # If the record is persisted and the `uploaded_at` field is not being changed,
+  # we don't need to recalculate the index.
+  defp update_upload_date_index(changeset), do: changeset
+
+  defp do_update_upload_date_index(%{changes: changes} = changeset) when is_map_key(changes, :uploaded_at) do
+    source_id = get_field(changeset, :source_id)
+    source = Sources.get_source!(source_id)
+    # Channels should count down from 99, playlists should count up from 0
+    # This reflects the fact that channels prepend new videos to the top of the list
+    # and playlists append new videos to the bottom of the list.
+    default_index = if source.collection_type == :channel, do: 99, else: 0
+    aggregator = if source.collection_type == :channel, do: :min, else: :max
+    change_direction = if source.collection_type == :channel, do: -1, else: 1
+
+    current_max =
+      MediaQuery.new()
+      |> where(^dynamic([mi], ^MediaQuery.upload_date_matches(changes.uploaded_at) and ^MediaQuery.for_source(source)))
+      |> Repo.aggregate(aggregator, :upload_date_index)
+
+    case current_max do
+      nil -> put_change(changeset, :upload_date_index, default_index)
+      max -> put_change(changeset, :upload_date_index, max + change_direction)
+    end
+  end
+
+  defp do_update_upload_date_index(changeset), do: changeset
+
+  defimpl Jason.Encoder, for: MediaItem do
+    def encode(value, opts) do
+      value
+      |> Repo.preload(:source)
+      |> Map.drop(MediaItem.json_exluded_fields())
+      |> Jason.Encode.map(opts)
+    end
   end
 end

lib/pinchflat/media/media_items_search_index.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.Media.MediaItemSearchIndex do
+defmodule Pinchflat.Media.MediaItemsSearchIndex do
   @moduledoc """
   The MediaItem fts5 search index. Not made to be directly interacted with,
   but I figured it'd be better to have it in-app so it's not a mystery.

lib/pinchflat/media/media_query.ex

@@ -0,0 +1,226 @@
+defmodule Pinchflat.Media.MediaQuery do
+  @moduledoc """
+  Query helpers for the Media context.
+
+  These methods are made to be one-ish liners used
+  to compose queries. Each method should strive to do
+  _one_ thing. These don't need to be tested as
+  they are just building blocks for other functionality
+  which, itself, will be tested.
+  """
+  import Ecto.Query, warn: false
+
+  alias Pinchflat.Media.MediaItem
+
+  # This allows the module to be aliased and query methods to be used
+  # all in one go
+  # usage: use Pinchflat.Media.MediaQuery
+  defmacro __using__(_opts) do
+    quote do
+      import Ecto.Query, warn: false
+
+      alias unquote(__MODULE__)
+    end
+  end
+
+  def new do
+    MediaItem
+  end
+
+  def for_source(source_id) when is_integer(source_id), do: dynamic([mi], mi.source_id == ^source_id)
+  def for_source(source), do: dynamic([mi], mi.source_id == ^source.id)
+
+  def downloaded, do: dynamic([mi], not is_nil(mi.media_filepath))
+  def download_prevented, do: dynamic([mi], mi.prevent_download == true)
+  def culling_prevented, do: dynamic([mi], mi.prevent_culling == true)
+  def redownloaded, do: dynamic([mi], not is_nil(mi.media_redownloaded_at))
+  def upload_date_matches(other_date), do: dynamic([mi], fragment("date(?) = date(?)", mi.uploaded_at, ^other_date))
+
+  def upload_date_after_source_cutoff do
+    dynamic(
+      [mi, source],
+      is_nil(source.download_cutoff_date) or
+        fragment("date(?) >= ?", mi.uploaded_at, source.download_cutoff_date)
+    )
+  end
+
+  def format_matching_profile_preference do
+    dynamic(
+      [mi, source, media_profile],
+      fragment("""
+        CASE
+          WHEN shorts_behaviour = 'only' AND livestream_behaviour = 'only' THEN
+            livestream = true OR short_form_content = true
+          WHEN shorts_behaviour = 'only' THEN
+            short_form_content = true
+          WHEN livestream_behaviour = 'only' THEN
+            livestream = true
+          WHEN shorts_behaviour = 'exclude' AND livestream_behaviour = 'exclude' THEN
+            short_form_content = false AND livestream = false
+          WHEN shorts_behaviour = 'exclude' THEN
+            short_form_content = false
+          WHEN livestream_behaviour = 'exclude' THEN
+            livestream = false
+          ELSE
+            true
+        END
+      """)
+    )
+  end
+
+  def matches_source_title_regex do
+    dynamic(
+      [mi, source],
+      is_nil(source.title_filter_regex) or fragment("regexp_like(?, ?)", mi.title, source.title_filter_regex)
+    )
+  end
+
+  def meets_min_and_max_duration do
+    dynamic(
+      [mi, source],
+      (is_nil(source.min_duration_seconds) or fragment("duration_seconds >= ?", source.min_duration_seconds)) and
+        (is_nil(source.max_duration_seconds) or fragment("duration_seconds <= ?", source.max_duration_seconds))
+    )
+  end
+
+  def past_retention_period do
+    dynamic(
+      [mi, source],
+      fragment("""
+        IFNULL(retention_period_days, 0) > 0 AND
+        DATETIME(media_downloaded_at, '+' || retention_period_days || ' day') < DATETIME('now')
+      """)
+    )
+  end
+
+  def past_redownload_delay do
+    dynamic(
+      [mi, source, media_profile],
+      # Returns media items where the uploaded_at is at least redownload_delay_days ago AND
+      # downloaded_at minus the redownload_delay_days is before the upload date
+      fragment("""
+        IFNULL(redownload_delay_days, 0) > 0 AND
+        DATE('now', '-' || redownload_delay_days || ' day') > DATE(uploaded_at) AND
+        DATE(media_downloaded_at, '-' || redownload_delay_days || ' day') < DATE(uploaded_at)
+      """)
+    )
+  end
+
+  def cullable do
+    dynamic(
+      [mi, source],
+      ^downloaded() and
+        ^past_retention_period() and
+        not (^culling_prevented())
+    )
+  end
+
+  def deletable_based_on_source_cutoff do
+    dynamic(
+      [mi, source],
+      ^downloaded() and
+        not (^upload_date_after_source_cutoff()) and
+        not (^culling_prevented())
+    )
+  end
+
+  def pending do
+    dynamic(
+      [mi],
+      not (^downloaded()) and
+        not (^download_prevented()) and
+        ^upload_date_after_source_cutoff() and
+        ^format_matching_profile_preference() and
+        ^matches_source_title_regex() and
+        ^meets_min_and_max_duration()
+    )
+  end
+
+  def upgradeable do
+    dynamic(
+      [mi, source],
+      ^downloaded() and
+        not (^download_prevented()) and
+        not (^redownloaded()) and
+        ^past_redownload_delay()
+    )
+  end
+
+  def matches_search_term(nil), do: dynamic([mi], true)
+
+  def matches_search_term(term) do
+    escaped_term = clean_search_term(term)
+
+    # Matching on `term` instead of `escaped_term` because the latter can mangle empty strings
+    case String.trim(term) do
+      "" -> dynamic([mi], true)
+      _ -> dynamic([mi], fragment("media_items_search_index MATCH ?", ^escaped_term))
+    end
+  end
+
+  def require_assoc(query, identifier) do
+    if has_named_binding?(query, identifier) do
+      query
+    else
+      do_require_assoc(query, identifier)
+    end
+  end
+
+  defp do_require_assoc(query, :media_items_search_index) do
+    from(mi in query, join: s in assoc(mi, :media_items_search_index), as: :media_items_search_index)
+  end
+
+  defp do_require_assoc(query, :source) do
+    from(mi in query, join: s in assoc(mi, :source), as: :source)
+  end
+
+  defp do_require_assoc(query, :media_profile) do
+    query
+    |> require_assoc(:source)
+    |> join(:inner, [mi, source], mp in assoc(source, :media_profile), as: :media_profile)
+  end
+
+  # This needs to be a non-dynamic query because it alone should control things like
+  # ordering and `snippets` for full-text search
+  def matching_search_term(query, nil), do: query
+
+  def matching_search_term(query, term) do
+    escaped_term = clean_search_term(term)
+
+    from(mi in query,
+      join: mi_search_index in assoc(mi, :media_items_search_index),
+      where: fragment("media_items_search_index MATCH ?", ^escaped_term),
+      select_merge: %{
+        matching_search_term:
+          fragment("""
+            coalesce(snippet(media_items_search_index, 0, '[PF_HIGHLIGHT]', '[/PF_HIGHLIGHT]', '...', 20), '') ||
+            ' ' ||
+            coalesce(snippet(media_items_search_index, 1, '[PF_HIGHLIGHT]', '[/PF_HIGHLIGHT]', '...', 20), '')
+          """)
+      },
+      order_by: [desc: fragment("rank")]
+    )
+  end
+
+  # SQLite's FTS5 is very picky about what it will accept as a search term.
+  # To that end, we need to clean up the search term before passing it to the
+  # MATCH clause.
+  # This method:
+  #   - Trims leading and trailing whitespace
+  #   - Collapses multiple spaces into a single space
+  #   - Removes quote characters
+  #   - Wraps any word in quotes (must happen after the double quote replacement)
+  #
+  # This allows for works with apostrophes and quotes to be searched for correctly
+  defp clean_search_term(nil), do: ""
+  defp clean_search_term(""), do: ""
+
+  defp clean_search_term(term) do
+    term
+    |> String.trim()
+    |> String.replace(~r/\s+/, " ")
+    |> String.split(~r/\s+/)
+    |> Enum.map(fn str -> String.replace(str, ~s("), "") end)
+    |> Enum.map_join(" ", fn str -> ~s("#{str}") end)
+  end
+end

lib/pinchflat/metadata/media_metadata.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.Media.MediaMetadata do
+defmodule Pinchflat.Metadata.MediaMetadata do
   @moduledoc """
   The MediaMetadata schema.
 

lib/pinchflat/metadata/metadata_file_helpers.ex

@@ -9,6 +9,27 @@ defmodule Pinchflat.Metadata.MetadataFileHelpers do
   needed
   """
 
+  alias Pinchflat.Sources
+  alias Pinchflat.Utils.FilesystemUtils
+
+  alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+
+  @doc """
+  Returns the directory where metadata for a database record should be stored.
+
+  Returns binary()
+  """
+  def metadata_directory_for(database_record) do
+    metadata_directory = Application.get_env(:pinchflat, :metadata_directory)
+    record_table_name = database_record.__meta__.source
+
+    Path.join([
+      metadata_directory,
+      record_table_name,
+      to_string(database_record.id)
+    ])
+  end
+
   @doc """
   Compresses and stores metadata for a media item, returning the filepath.
 
@@ -18,8 +39,7 @@ defmodule Pinchflat.Metadata.MetadataFileHelpers do
     filepath = generate_filepath_for(database_record, "metadata.json.gz")
     {:ok, json} = Phoenix.json_library().encode(metadata_map)
 
-    File.mkdir_p!(Path.dirname(filepath))
-    :ok = File.write(filepath, json, [:compressed])
+    :ok = FilesystemUtils.write_p!(filepath, json, [:compressed])
 
     filepath
   end
@@ -30,42 +50,101 @@ defmodule Pinchflat.Metadata.MetadataFileHelpers do
   Returns {:ok, map()} | {:error, any}
   """
   def read_compressed_metadata(filepath) do
-    {:ok, json} = File.open(filepath, [:read, :compressed], &IO.read(&1, :all))
+    {:ok, json} = File.open(filepath, [:read, :compressed], &IO.read(&1, :eof))
 
     Phoenix.json_library().decode(json)
   end
 
   @doc """
   Downloads and stores a thumbnail for a media item, returning the filepath.
+  Chooses the highest quality thumbnail available and converts it to a JPG
 
-  Returns binary()
+  Returns nil if no thumbnail is available or if yt-dlp encounters an error
+
+  Returns binary() | nil
   """
-  def download_and_store_thumbnail_for(database_record, metadata_map) do
-    thumbnail_url = metadata_map["thumbnail"]
-    filepath = generate_filepath_for(database_record, Path.basename(thumbnail_url))
-    thumbnail_blob = fetch_thumbnail_from_url(thumbnail_url)
+  def download_and_store_thumbnail_for(media_item_with_preloads) do
+    yt_dlp_filepath = generate_filepath_for(media_item_with_preloads, "thumbnail.%(ext)s")
+    real_filepath = generate_filepath_for(media_item_with_preloads, "thumbnail.jpg")
+    command_opts = [output: yt_dlp_filepath]
+    addl_opts = [use_cookies: Sources.use_cookies?(media_item_with_preloads.source, :metadata)]
 
-    File.mkdir_p!(Path.dirname(filepath))
-    :ok = File.write(filepath, thumbnail_blob)
-
-    filepath
+    case YtDlpMedia.download_thumbnail(media_item_with_preloads.original_url, command_opts, addl_opts) do
+      {:ok, _} -> real_filepath
+      _ -> nil
+    end
   end
 
-  defp fetch_thumbnail_from_url(url) do
-    http_client = Application.get_env(:pinchflat, :http_client, Pinchflat.HTTP.HTTPClient)
-    {:ok, body} = http_client.get(url, [], body_format: :binary)
+  @doc """
+  Parses an upload date from the YYYYMMDD string returned in yt-dlp metadata
+  and returns a DateTime struct, appending a time of 00:00:00Z.
 
-    body
+  Returns DateTime.t()
+  """
+  def parse_upload_date(upload_date) do
+    <<year::binary-size(4)>> <> <<month::binary-size(2)>> <> <<day::binary-size(2)>> = upload_date
+
+    case DateTime.from_iso8601("#{year}-#{month}-#{day}T00:00:00Z") do
+      {:ok, datetime, _} -> datetime
+      _ -> raise "Invalid upload date: #{upload_date}"
+    end
+  end
+
+  @doc """
+  Attempts to determine the series directory from a media filepath.
+  The series directory is the "root" directory for a given source
+  which should contain all the season-level folders of that source.
+
+  Used for determining where to store things like NFO data and banners
+  for media center apps. Not useful without a media center app.
+
+  Returns {:ok, binary()} | {:error, :indeterminable}
+  """
+  def series_directory_from_media_filepath(media_filepath) do
+    # Matches "s" or "season" (case-insensitive)
+    # followed by an optional non-word character (. or _ or <space>, etc)
+    # followed by at least one digit
+    # followed immediately by the end of the string
+    # Example matches: s1, s.1, s01 season 1, Season.01, Season_1, Season 1, Season1
+    # Example non-matches: s01e01, season, series 1,
+    season_regex = ~r/^s(eason)?(\W|_)?\d{1,}$/i
+
+    {series_directory, found_series_directory} =
+      media_filepath
+      |> Path.split()
+      |> Enum.reduce_while({[], false}, fn part, {directory_acc, _} ->
+        if String.match?(part, season_regex) do
+          {:halt, {directory_acc, true}}
+        else
+          {:cont, {directory_acc ++ [part], false}}
+        end
+      end)
+
+    if found_series_directory do
+      {:ok, Path.join(series_directory)}
+    else
+      {:error, :indeterminable}
+    end
+  end
+
+  @doc """
+  Attempts to determine the season and episode number from a media filepath.
+
+  Returns {:ok, {binary(), binary()}} | {:error, :indeterminable}
+  """
+  def season_and_episode_from_media_filepath(media_filepath) do
+    # matches s + 1 or more digits + e + 1 or more digits (case-insensitive)
+    season_episode_regex = ~r/s(\d+)e(\d+)/i
+
+    case Regex.scan(season_episode_regex, media_filepath) do
+      [[_, season, episode] | _] -> {:ok, {season, episode}}
+      _ -> {:error, :indeterminable}
+    end
   end
 
   defp generate_filepath_for(database_record, filename) do
-    metadata_directory = Application.get_env(:pinchflat, :metadata_directory)
-    record_table_name = database_record.__meta__.source
-
     Path.join([
-      metadata_directory,
-      record_table_name,
-      to_string(database_record.id),
+      metadata_directory_for(database_record),
       filename
     ])
   end

lib/pinchflat/metadata/metadata_parser.ex

@@ -8,6 +8,8 @@ defmodule Pinchflat.Metadata.MetadataParser do
   and not have it, ya know?
   """
 
+  alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+
   @doc """
   Parses the given JSON response from yt-dlp and returns a map of
   the needful media_item attributes, along with anything needed for
@@ -24,14 +26,12 @@ defmodule Pinchflat.Metadata.MetadataParser do
   end
 
   defp parse_media_metadata(metadata) do
-    %{
-      media_id: metadata["id"],
-      title: metadata["title"],
-      original_url: metadata["original_url"],
-      description: metadata["description"],
-      media_filepath: metadata["filepath"],
-      livestream: metadata["was_live"]
-    }
+    Map.merge(
+      Map.from_struct(YtDlpMedia.response_to_struct(metadata)),
+      %{
+        media_filepath: metadata["filepath"]
+      }
+    )
   end
 
   defp parse_subtitle_metadata(metadata) do
@@ -54,14 +54,42 @@ defmodule Pinchflat.Metadata.MetadataParser do
       |> Enum.reverse()
       |> Enum.find_value(fn attrs -> attrs["filepath"] end)
 
-    %{
-      thumbnail_filepath: thumbnail_filepath
-    }
+    if thumbnail_filepath do
+      # NOTE: whole ordeal needed due to a bug I found in yt-dlp
+      # https://github.com/yt-dlp/yt-dlp/issues/9445
+      # Can be reverted to remove this entire conditional once fixed
+      filepath =
+        thumbnail_filepath
+        |> String.split(~r{\.}, include_captures: true)
+        |> List.insert_at(-3, "-thumb")
+        |> Enum.join()
+
+      %{
+        thumbnail_filepath: filepath_if_exists(filepath)
+      }
+    else
+      %{
+        thumbnail_filepath: nil
+      }
+    end
   end
 
   defp parse_infojson_metadata(metadata) do
     %{
-      metadata_filepath: metadata["infojson_filename"]
+      metadata_filepath: filepath_if_exists(metadata["infojson_filename"])
     }
   end
+
+  # NOTE: this should not be needed, but it is due to a bug in yt-dlp.
+  # Can remove once this is resolved:
+  # https://github.com/yt-dlp/yt-dlp/issues/9445#issuecomment-2018724344
+  defp filepath_if_exists(nil), do: nil
+
+  defp filepath_if_exists(filepath) do
+    if File.exists?(filepath) do
+      filepath
+    else
+      nil
+    end
+  end
 end

lib/pinchflat/metadata/nfo_builder.ex

@@ -0,0 +1,82 @@
+defmodule Pinchflat.Metadata.NfoBuilder do
+  @moduledoc """
+  Provides methods for building and storing NFO files for
+  use by Kodi/Jellyfin and other media center software.
+  """
+
+  import Pinchflat.Utils.XmlUtils, only: [safe: 1]
+
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Metadata.MetadataFileHelpers
+
+  @doc """
+  Builds an NFO file for a media item (read: single "episode") and
+  stores it at the specified location.
+
+  Returns the filepath of the NFO file.
+  """
+  def build_and_store_for_media_item(nfo_filepath, metadata) do
+    nfo = build_for_media_item(nfo_filepath, metadata)
+
+    FilesystemUtils.write_p!(nfo_filepath, nfo)
+
+    nfo_filepath
+  end
+
+  @doc """
+  Builds an NFO file for a souce and stores it at the specified location.
+  Technically works for playlists, but it's really made for channels.
+
+  Returns the filepath of the NFO file.
+  """
+  def build_and_store_for_source(filepath, metadata) do
+    nfo = build_for_source(metadata)
+
+    FilesystemUtils.write_p!(filepath, nfo)
+
+    filepath
+  end
+
+  defp build_for_media_item(nfo_filepath, metadata) do
+    upload_date = MetadataFileHelpers.parse_upload_date(metadata["upload_date"])
+    # NOTE: the filepath here isn't the path of the media item, it's the path that
+    # the NFO should be saved to. This works because the NFO's path is the same as
+    # the media's path, just with a different extension. If this ever changes I'll
+    # need to pass in the media item's path as well.
+    {season, episode} = determine_season_and_episode_number(nfo_filepath, upload_date)
+
+    # Cribbed from a combination of the Kodi wiki, ytdl-nfo, and ytdl-sub.
+    """
+    <?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
+    <episodedetails>
+      <title>#{safe(metadata["title"])}</title>
+      <showtitle>#{safe(metadata["uploader"])}</showtitle>
+      <uniqueid type="youtube" default="true">#{safe(metadata["id"])}</uniqueid>
+      <plot>#{safe(metadata["description"])}</plot>
+      <aired>#{safe(upload_date)}</aired>
+      <season>#{safe(season)}</season>
+      <episode>#{episode}</episode>
+      <genre>YouTube</genre>
+    </episodedetails>
+    """
+  end
+
+  defp build_for_source(metadata) do
+    """
+    <?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
+    <tvshow>
+      <title>#{safe(metadata["title"])}</title>
+      <plot>#{safe(metadata["description"])}</plot>
+      <uniqueid type="youtube" default="true">#{safe(metadata["id"])}</uniqueid>
+      <genre>YouTube</genre>
+    </tvshow>
+    """
+  end
+
+  defp determine_season_and_episode_number(filepath, upload_date) do
+    case MetadataFileHelpers.season_and_episode_from_media_filepath(filepath) do
+      {:ok, {season, episode}} -> {season, episode}
+      {:error, _} -> {upload_date.year, Calendar.strftime(upload_date, "%m%d")}
+    end
+  end
+end

lib/pinchflat/metadata/source_image_parser.ex

@@ -0,0 +1,99 @@
+defmodule Pinchflat.Metadata.SourceImageParser do
+  @moduledoc """
+  Functions for parsing and storing source images.
+  """
+  alias Pinchflat.Utils.FilesystemUtils
+
+  @doc """
+  Given a base directory and source metadata, look for the appropriate images
+  and move them to the base directory. Returns a map of image types and their
+  filepaths (specifically for a source). If a given image type is not found,
+  the key will not be present in the map.
+
+  The metadata is expected to contain the `filepath` key for images, implying
+  that the images have already been downloaded by yt-dlp. This method will NOT
+  download images from the internet - it just copys them around.
+
+  Returns a map with the possible keys :poster_filepath, :fanart_filepath, and
+  :banner_filepath.
+  """
+  def store_source_images(base_directory, source_metadata) do
+    (source_metadata["thumbnails"] || [])
+    |> Enum.filter(&(&1["filepath"] != nil))
+    |> select_useful_images(source_metadata)
+    |> Enum.map(&move_image(&1, base_directory))
+    |> Enum.into(%{})
+  end
+
+  defp select_useful_images(images, source_metadata) do
+    labelled_images =
+      Enum.reduce(images, %{}, fn image_map, acc ->
+        case image_map do
+          %{"id" => "avatar_uncropped"} -> put_image_key(acc, :poster, image_map["filepath"])
+          %{"id" => "banner_uncropped"} -> put_image_key(acc, :fanart, image_map["filepath"])
+          _ -> acc
+        end
+      end)
+
+    labelled_images
+    |> add_fallback_poster(source_metadata)
+    |> put_image_key(:banner, determine_best_banner(images))
+    |> Enum.filter(fn {_key, attrs} -> attrs.current_filepath end)
+  end
+
+  # If a poster is set, short-circuit and return the images as-is
+  defp add_fallback_poster(%{poster: _} = images, _), do: images
+
+  # If a poster is NOT set, see if we can find a suitable image to use as a fallback
+  defp add_fallback_poster(images, source_metadata) do
+    case source_metadata["entries"] do
+      nil -> images
+      [] -> images
+      [first_entry | _] -> add_poster_from_entry_thumbnail(images, first_entry)
+    end
+  end
+
+  defp add_poster_from_entry_thumbnail(images, entry) do
+    thumbnail =
+      (entry["thumbnails"] || [])
+      |> Enum.reverse()
+      |> Enum.find(& &1["filepath"])
+
+    case thumbnail do
+      nil -> images
+      _ -> put_image_key(images, :poster, thumbnail["filepath"])
+    end
+  end
+
+  defp determine_best_banner(images) do
+    best_candidate =
+      images
+      # Filter out images that don't have a width and height attribute
+      |> Enum.filter(&(&1["width"] && &1["height"]))
+      # Sort images with the highest width first
+      |> Enum.sort(&(&1["width"] >= &2["width"]))
+      # Find the first image where the ratio of width to height is greater than 3
+      |> Enum.find(&(&1["width"] / &1["height"] > 3))
+
+    Map.get(best_candidate || %{}, "filepath")
+  end
+
+  defp move_image({_key, attrs}, base_directory) do
+    extension = Path.extname(attrs.current_filepath)
+    final_filepath = Path.join([base_directory, "#{attrs.final_filename}#{extension}"])
+
+    FilesystemUtils.cp_p!(attrs.current_filepath, final_filepath)
+
+    {attrs.attribute_name, final_filepath}
+  end
+
+  defp put_image_key(map, key, image) do
+    attribute_atom = String.to_existing_atom("#{key}_filepath")
+
+    Map.put(map, key, %{
+      attribute_name: attribute_atom,
+      final_filename: to_string(key),
+      current_filepath: image
+    })
+  end
+end

lib/pinchflat/metadata/source_metadata.ex

@@ -0,0 +1,50 @@
+defmodule Pinchflat.Metadata.SourceMetadata do
+  @moduledoc """
+  The SourceMetadata schema.
+
+  Look. Don't @ me about Metadata vs. Metadatum. I'm very sensitive.
+  """
+
+  use Ecto.Schema
+  import Ecto.Changeset
+
+  alias Pinchflat.Sources.Source
+
+  @allowed_fields ~w(
+    metadata_filepath
+    fanart_filepath
+    poster_filepath
+    banner_filepath
+  )a
+
+  @required_fields ~w(metadata_filepath)a
+
+  schema "source_metadata" do
+    field :metadata_filepath, :string
+    field :fanart_filepath, :string
+    field :poster_filepath, :string
+    field :banner_filepath, :string
+
+    belongs_to :source, Source
+
+    timestamps(type: :utc_datetime)
+  end
+
+  @doc false
+  def changeset(source_metadata, attrs) do
+    source_metadata
+    |> cast(attrs, @allowed_fields)
+    |> validate_required(@required_fields)
+    |> unique_constraint([:source_id])
+  end
+
+  @doc false
+  def filepath_attributes do
+    ~w(
+      metadata_filepath
+      fanart_filepath
+      poster_filepath
+      banner_filepath
+    )a
+  end
+end

lib/pinchflat/metadata/source_metadata_storage_worker.ex

@@ -0,0 +1,131 @@
+defmodule Pinchflat.Metadata.SourceMetadataStorageWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :remote_metadata,
+    tags: ["media_source", "source_metadata", "remote_metadata", "show_in_dashboard"],
+    max_attempts: 3
+
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Repo
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Utils.StringUtils
+  alias Pinchflat.Metadata.NfoBuilder
+  alias Pinchflat.YtDlp.MediaCollection
+  alias Pinchflat.Metadata.SourceImageParser
+  alias Pinchflat.Metadata.MetadataFileHelpers
+  alias Pinchflat.Downloading.DownloadOptionBuilder
+
+  @doc """
+  Starts the source metadata storage worker and creates a task for the source.
+
+  Returns {:ok, %Task{}} | {:error, :duplicate_job} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff_with_task(source, opts \\ []) do
+    %{id: source.id}
+    |> SourceMetadataStorageWorker.new(opts)
+    |> Tasks.create_job_with_task(source)
+  end
+
+  @doc """
+  Fetches and stores various forms of metadata for a source:
+    - Attributes like `description`
+    - JSON metadata for internal use
+    - The series directory for the source
+    - The NFO file for the source (if specified)
+    - Downloads and stores source images (if specified)
+
+  The worker is kicked off after a source is inserted or it's original_url
+  is updated - this can take an unknown amount of time so don't rely on this
+  data being here before, say, the first indexing or downloading task is complete.
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+    source = Repo.preload(Sources.get_source!(source_id), [:metadata, :media_profile])
+    series_directory = determine_series_directory(source)
+
+    {source_metadata, source_image_attrs, metadata_image_attrs} =
+      fetch_source_metadata_and_images(series_directory, source)
+
+    source_metadata_filepath = MetadataFileHelpers.compress_and_store_metadata_for(source, source_metadata)
+
+    Sources.update_source(
+      source,
+      Map.merge(
+        %{
+          series_directory: series_directory,
+          nfo_filepath: store_source_nfo(source, series_directory, source_metadata),
+          description: source_metadata["description"],
+          metadata: Map.merge(%{metadata_filepath: source_metadata_filepath}, metadata_image_attrs)
+        },
+        source_image_attrs
+      ),
+      # `run_post_commit_tasks: false` prevents this from running in an infinite loop
+      run_post_commit_tasks: false
+    )
+
+    :ok
+  rescue
+    Ecto.NoResultsError -> Logger.info("#{__MODULE__} discarded: source #{source_id} not found")
+    Ecto.StaleEntryError -> Logger.info("#{__MODULE__} discarded: source #{source_id} stale")
+  end
+
+  defp fetch_source_metadata_and_images(series_directory, source) do
+    metadata_directory = MetadataFileHelpers.metadata_directory_for(source)
+
+    {:ok, metadata} = fetch_metadata_for_source(source)
+    metadata_image_attrs = SourceImageParser.store_source_images(metadata_directory, metadata)
+
+    if source.media_profile.download_source_images && series_directory do
+      source_image_attrs = SourceImageParser.store_source_images(series_directory, metadata)
+
+      {metadata, source_image_attrs, metadata_image_attrs}
+    else
+      {metadata, %{}, metadata_image_attrs}
+    end
+  end
+
+  defp determine_series_directory(source) do
+    output_path = DownloadOptionBuilder.build_output_path_for(source)
+    runner_opts = [output: output_path]
+    addl_opts = [use_cookies: Sources.use_cookies?(source, :metadata)]
+    {:ok, %{filepath: filepath}} = MediaCollection.get_source_details(source.original_url, runner_opts, addl_opts)
+
+    case MetadataFileHelpers.series_directory_from_media_filepath(filepath) do
+      {:ok, series_directory} -> series_directory
+      {:error, _} -> nil
+    end
+  end
+
+  defp store_source_nfo(source, series_directory, metadata) do
+    if source.media_profile.download_nfo && series_directory do
+      nfo_filepath = Path.join(series_directory, "tvshow.nfo")
+
+      NfoBuilder.build_and_store_for_source(nfo_filepath, metadata)
+    end
+  end
+
+  defp fetch_metadata_for_source(source) do
+    tmp_output_path = "#{tmp_directory()}/#{StringUtils.random_string(16)}/source_image.%(ext)S"
+    base_opts = [convert_thumbnails: "jpg", output: tmp_output_path]
+    should_use_cookies = Sources.use_cookies?(source, :metadata)
+
+    opts =
+      if source.collection_type == :channel do
+        base_opts ++ [:write_all_thumbnails, playlist_items: 0]
+      else
+        base_opts ++ [:write_thumbnail, playlist_items: 1]
+      end
+
+    MediaCollection.get_source_metadata(source.original_url, opts, use_cookies: should_use_cookies)
+  end
+
+  defp tmp_directory do
+    Application.get_env(:pinchflat, :tmpfile_directory)
+  end
+end

lib/pinchflat/podcasts/opml_feed_builder.ex

@@ -0,0 +1,40 @@
+defmodule Pinchflat.Podcasts.OpmlFeedBuilder do
+  @moduledoc """
+  Methods for building an OPML feed for a list of sources.
+  """
+
+  import Pinchflat.Utils.XmlUtils, only: [safe: 1]
+
+  alias PinchflatWeb.Router.Helpers, as: Routes
+
+  @doc """
+  Builds an OPML feed for a given list of sources.
+
+  Returns an XML document as a string.
+  """
+  def build(url_base, sources) do
+    sources_xml =
+      Enum.map(
+        sources,
+        &"""
+        <outline type="rss" text="#{safe(&1.custom_name)}" xmlUrl="#{safe(source_route(url_base, &1))}" />
+        """
+      )
+
+    """
+    <?xml version="1.0" encoding="UTF-8"?>
+    <opml version="2.0">
+      <head>
+        <title>All Sources</title>
+      </head>
+      <body>
+        #{Enum.join(sources_xml, "\n")}
+      </body>
+    </opml>
+    """
+  end
+
+  defp source_route(url_base, source) do
+    Path.join(url_base, "#{Routes.podcast_path(PinchflatWeb.Endpoint, :rss_feed, source.uuid)}.xml")
+  end
+end

lib/pinchflat/podcasts/podcast_helpers.ex

@@ -0,0 +1,93 @@
+defmodule Pinchflat.Podcasts.PodcastHelpers do
+  @moduledoc """
+  Methods for fetching postcast-related data from a source
+  or its media items
+  """
+
+  use Pinchflat.Media.MediaQuery
+  use Pinchflat.Sources.SourcesQuery
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Metadata.MediaMetadata
+  alias Pinchflat.Metadata.SourceMetadata
+
+  @doc """
+  Returns a list of sources that are not marked for deletion.
+
+  Returns: [%Source{}]
+  """
+  def opml_sources() do
+    SourcesQuery.new()
+    |> select([s], %{custom_name: s.custom_name, uuid: s.uuid})
+    |> where([s], is_nil(s.marked_for_deletion_at))
+    |> order_by(asc: :custom_name)
+    |> Repo.all()
+  end
+
+  @doc """
+  Returns a list of media items that have been downloaded to disk
+  and have been proven to still exist there.
+
+  Useful for podcasts since we don't want to serve media that
+  has been deleted or moved, but it's also fairly generally useful
+  so I could see this being moved in the future.
+
+  Options:
+    - limit: integer - the maximum number of media items to return
+
+  Returns: [%MediaItem{}]
+  """
+  def persisted_media_items_for(source, opts \\ []) do
+    limit = Keyword.get(opts, :limit, 1_000)
+
+    MediaQuery.new()
+    |> where(^dynamic(^MediaQuery.for_source(source) and ^MediaQuery.downloaded()))
+    |> order_by(desc: :uploaded_at)
+    |> Repo.maybe_limit(limit)
+    |> Repo.all()
+    |> Enum.filter(fn media_item -> File.exists?(media_item.media_filepath) end)
+  end
+
+  @doc """
+  Selects a cover image for a source based on the source's metadata
+  and the metadata of the media items associated with the source. Also
+  ensures images exist on disk.
+
+  Only one media item should need to be returned since this is using the
+  internal metadata which, so long as the media_item was _downloaded_, should
+  be guaranteed to exist.
+
+  Prefers the source's poster, then fanart, then the media item's thumbnail.
+
+  Returns: {:ok, filepath} | {:error, :no_suitable_image}
+  """
+  def select_cover_image(source, media_items) do
+    source_with_preloads = Repo.preload(source, :metadata)
+
+    source_with_preloads
+    |> get_images_by_preference(media_items)
+    |> Enum.reject(&is_nil(&1))
+    |> Enum.find(&File.exists?/1)
+    |> case do
+      nil -> {:error, :no_suitable_image}
+      filepath -> {:ok, filepath}
+    end
+  end
+
+  defp get_images_by_preference(source_with_preloads, []) do
+    source_metadata = source_with_preloads.metadata || %SourceMetadata{}
+
+    [
+      source_metadata.poster_filepath,
+      source_metadata.fanart_filepath
+    ]
+  end
+
+  defp get_images_by_preference(source_with_preloads, [media_item | _]) do
+    media_item_with_preloads = Repo.preload(media_item, :metadata)
+    media_item_metadata = media_item_with_preloads.metadata || %MediaMetadata{}
+    source_images = get_images_by_preference(source_with_preloads, [])
+
+    source_images ++ [media_item_metadata.thumbnail_filepath]
+  end
+end

lib/pinchflat/podcasts/rss_feed_builder.ex

@@ -0,0 +1,142 @@
+defmodule Pinchflat.Podcasts.RssFeedBuilder do
+  @moduledoc """
+  Methods for building an RSS feed for a source and its media items.
+  """
+
+  @datetime_format "%a, %d %b %Y %H:%M:%S %z"
+
+  import Pinchflat.Utils.XmlUtils, only: [safe: 1]
+
+  alias Pinchflat.Podcasts.PodcastHelpers
+  alias PinchflatWeb.Router.Helpers, as: Routes
+
+  @doc """
+  Builds an RSS feed for a given source and its media items.
+  Only MediaItems that have been persisted will be included in the feed.
+
+  ## Options:
+    - `:limit` - The maximum number of media items to include in the feed. Defaults to 2,000.
+
+  Returns an XML document as a string.
+  """
+  def build(source, opts \\ []) do
+    limit = Keyword.get(opts, :limit, 2_000)
+    url_base = Keyword.get(opts, :url_base, PinchflatWeb.Endpoint.url())
+
+    media_items = PodcastHelpers.persisted_media_items_for(source, limit: limit)
+    build_source_xml(source, media_items, url_base)
+  end
+
+  defp build_source_xml(source, media_items, url_base) do
+    media_item_xml = Enum.map(media_items, &build_media_item_xml(source, &1, url_base))
+    # "caching" the image path since it requires some DB calls and is used twice
+    feed_image_path = feed_image_path(url_base, source, media_items)
+
+    # Useful: resources:
+    #   - https://validator.w3.org/feed/#validate_by_input
+    #   - https://github.com/Podcastindex-org/podcast-namespace/blob/main/docs/1.0.md
+    #   - https://podba.se/validate
+    """
+    <?xml version="1.0" encoding="UTF-8"?>
+    <rss version="2.0"
+      xmlns:itunes="http://www.itunes.com/dtds/podcast-1.0.dtd"
+      xmlns:podcast="https://podcastindex.org/namespace/1.0"
+      xmlns:atom="http://www.w3.org/2005/Atom">
+      <channel>
+        <title>#{safe(source.custom_name)}</title>
+        <link>#{safe(source.original_url)}</link>
+        <description>#{safe(source.description)}</description>
+        <category>TV &amp; Film</category>
+        <generator>Generated by Pinchflat</generator>
+        <language>en-us</language>
+        <lastBuildDate>#{Calendar.strftime(source.updated_at, @datetime_format)}</lastBuildDate>
+        <pubDate>#{Calendar.strftime(source.inserted_at, @datetime_format)}</pubDate>
+        <atom:link href="#{safe(generate_self_link(url_base, source))}" rel="self" type="application/rss+xml" />
+        <podcast:locked>yes</podcast:locked>
+        <podcast:guid>#{source.uuid}</podcast:guid>
+        <image>
+          <url>#{safe(feed_image_path)}</url>
+          <title>#{safe(source.custom_name)}</title>
+          <link>#{safe(source.original_url)}</link>
+        </image>
+        <itunes:author>#{safe(source.custom_name)}</itunes:author>
+        <itunes:subtitle>#{safe(source.custom_name)}</itunes:subtitle>
+        <itunes:block>yes</itunes:block>
+        <itunes:image href="#{safe(feed_image_path)}"></itunes:image>
+        <itunes:explicit>false</itunes:explicit>
+        <itunes:category text="TV &amp; Film"></itunes:category>
+
+        #{Enum.join(media_item_xml, "\n")}
+
+      </channel>
+    </rss>
+    """
+  end
+
+  defp build_media_item_xml(source, media_item, url_base) do
+    item_image_path = item_image_path(url_base, media_item)
+
+    """
+    <item>
+      <guid isPermaLink="false">#{media_item.uuid}</guid>
+      <title>#{safe(media_item.title)}</title>
+      <link>#{safe(media_item.original_url)}</link>
+      <description>#{safe(media_item.description)}</description>
+      <pubDate>#{Calendar.strftime(media_item.uploaded_at, @datetime_format)}</pubDate>
+      <itunes:duration>#{media_item.duration_seconds}</itunes:duration>
+      <enclosure
+        url="#{media_stream_path(url_base, media_item)}"
+        length="#{media_item.media_size_bytes}"
+        type="#{MIME.from_path(media_item.media_filepath)}"
+      />
+      <itunes:author>#{safe(source.custom_name)}</itunes:author>
+      <itunes:subtitle>#{safe(media_item.title)}</itunes:subtitle>
+      <itunes:summary><![CDATA[#{media_item.description}]]></itunes:summary>
+
+      #{item_image_path && ~s(<itunes:image href="#{safe(item_image_path)}"></itunes:image>)}
+      #{item_image_path && ~s(<podcast:images srcset="#{safe(item_image_path)}" />)}
+
+      <itunes:explicit>false</itunes:explicit>
+    </item>
+    """
+  end
+
+  defp generate_self_link(url_base, source) do
+    Path.join(url_base, "#{podcast_route(:rss_feed, source.uuid)}.xml")
+  end
+
+  defp media_stream_path(url_base, media_item) do
+    extension = Path.extname(media_item.media_filepath)
+
+    Path.join(url_base, "#{media_route(:stream, media_item.uuid)}#{extension}")
+  end
+
+  defp feed_image_path(url_base, source, media_items) do
+    case PodcastHelpers.select_cover_image(source, media_items) do
+      {:error, _} ->
+        ""
+
+      {:ok, filepath} ->
+        extension = Path.extname(filepath)
+        Path.join(url_base, "#{podcast_route(:feed_image, source.uuid)}#{extension}")
+    end
+  end
+
+  def item_image_path(url_base, media_item) do
+    if media_item.thumbnail_filepath && File.exists?(media_item.thumbnail_filepath) do
+      extension = Path.extname(media_item.thumbnail_filepath)
+
+      Path.join(url_base, "#{podcast_route(:episode_image, media_item.uuid)}#{extension}")
+    else
+      nil
+    end
+  end
+
+  defp podcast_route(action, params) do
+    Routes.podcast_path(PinchflatWeb.Endpoint, action, params)
+  end
+
+  defp media_route(action, params) do
+    Routes.media_item_path(PinchflatWeb.Endpoint, action, params)
+  end
+end

lib/pinchflat/profiles/media_profile.ex

@@ -6,6 +6,7 @@ defmodule Pinchflat.Profiles.MediaProfile do
   use Ecto.Schema
   import Ecto.Changeset
 
+  alias __MODULE__
   alias Pinchflat.Sources.Source
 
   @allowed_fields ~w(
@@ -17,32 +18,45 @@ defmodule Pinchflat.Profiles.MediaProfile do
     sub_langs
     download_thumbnail
     embed_thumbnail
+    download_source_images
     download_metadata
     embed_metadata
+    download_nfo
+    sponsorblock_behaviour
+    sponsorblock_categories
     shorts_behaviour
     livestream_behaviour
+    audio_track
     preferred_resolution
+    media_container
+    redownload_delay_days
+    marked_for_deletion_at
   )a
 
   @required_fields ~w(name output_path_template)a
 
   schema "media_profiles" do
     field :name, :string
+    field :redownload_delay_days, :integer
 
     field :output_path_template, :string,
-      default: "/{{ source_custom_name }}/{{ title }}/{{ title }} [{{ id }}].{{ ext }}"
+      default: "/{{ source_custom_name }}/{{ upload_yyyy_mm_dd }} {{ title }}/{{ title }} [{{ id }}].{{ ext }}"
 
-    field :download_subs, :boolean, default: true
-    field :download_auto_subs, :boolean, default: true
-    field :embed_subs, :boolean, default: true
+    field :download_subs, :boolean, default: false
+    field :download_auto_subs, :boolean, default: false
+    field :embed_subs, :boolean, default: false
     field :sub_langs, :string, default: "en"
 
-    field :download_thumbnail, :boolean, default: true
-    field :embed_thumbnail, :boolean, default: true
+    field :download_thumbnail, :boolean, default: false
+    field :embed_thumbnail, :boolean, default: false
+    field :download_source_images, :boolean, default: false
 
-    field :download_metadata, :boolean, default: true
-    field :embed_metadata, :boolean, default: true
+    field :download_metadata, :boolean, default: false
+    field :embed_metadata, :boolean, default: false
 
+    field :download_nfo, :boolean, default: false
+    field :sponsorblock_behaviour, Ecto.Enum, values: [:disabled, :mark, :remove], default: :disabled
+    field :sponsorblock_categories, {:array, :string}, default: []
     # NOTE: these do NOT speed up indexing - the indexer still has to go
     # through the entire collection to determine if a media is a short or
     # a livestream.
@@ -52,8 +66,11 @@ defmodule Pinchflat.Profiles.MediaProfile do
     # See `build_format_clauses` in the Media context for more.
     field :shorts_behaviour, Ecto.Enum, values: ~w(include exclude only)a, default: :include
     field :livestream_behaviour, Ecto.Enum, values: ~w(include exclude only)a, default: :include
+    field :audio_track, :string
+    field :preferred_resolution, Ecto.Enum, values: ~w(4320p 2160p 1440p 1080p 720p 480p 360p audio)a, default: :"1080p"
+    field :media_container, :string, default: nil
 
-    field :preferred_resolution, Ecto.Enum, values: ~w(2160p 1080p 720p 480p 360p audio)a, default: :"1080p"
+    field :marked_for_deletion_at, :utc_datetime
 
     has_many :sources, Source
 
@@ -65,6 +82,27 @@ defmodule Pinchflat.Profiles.MediaProfile do
     media_profile
     |> cast(attrs, @allowed_fields)
     |> validate_required(@required_fields)
+    # Ensures it ends with `.{{ ext }}` or `.%(ext)s` or similar (with a little wiggle room)
+    |> validate_format(:output_path_template, ext_regex(), message: "must end with .{{ ext }}")
+    |> validate_number(:redownload_delay_days, greater_than_or_equal_to: 0)
     |> unique_constraint(:name)
   end
+
+  @doc false
+  def ext_regex do
+    ~r/\.({{ ?ext ?}}|%\( ?ext ?\)[sS])$/
+  end
+
+  @doc false
+  def json_exluded_fields do
+    ~w(__meta__ __struct__ sources)a
+  end
+
+  defimpl Jason.Encoder, for: MediaProfile do
+    def encode(value, opts) do
+      value
+      |> Map.drop(MediaProfile.json_exluded_fields())
+      |> Jason.Encode.map(opts)
+    end
+  end
 end

lib/pinchflat/profiles/media_profile_deletion_worker.ex

@@ -0,0 +1,38 @@
+defmodule Pinchflat.Profiles.MediaProfileDeletionWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :local_data,
+    tags: ["media_profiles", "local_data"]
+
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Profiles
+
+  @doc """
+  Starts the profile deletion worker. Does not attach it to a task like `kickoff_with_task/2`
+  since deletion also cancels all tasks for the profile
+
+  Returns {:ok, %Oban.Job{}} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff(profile, job_args \\ %{}, job_opts \\ []) do
+    %{id: profile.id}
+    |> Map.merge(job_args)
+    |> MediaProfileDeletionWorker.new(job_opts)
+    |> Oban.insert()
+  end
+
+  @doc """
+  Deletes a profile and optionally deletes its files
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => profile_id} = args}) do
+    delete_files = Map.get(args, "delete_files", false)
+    profile = Profiles.get_media_profile!(profile_id)
+
+    Profiles.delete_media_profile(profile, delete_files: delete_files)
+  end
+end

lib/pinchflat/profiles/profiles_query.ex

@@ -0,0 +1,29 @@
+defmodule Pinchflat.Profiles.ProfilesQuery do
+  @moduledoc """
+  Query helpers for the Profiles context.
+
+  These methods are made to be one-ish liners used
+  to compose queries. Each method should strive to do
+  _one_ thing. These don't need to be tested as
+  they are just building blocks for other functionality
+  which, itself, will be tested.
+  """
+  import Ecto.Query, warn: false
+
+  alias Pinchflat.Profiles.MediaProfile
+
+  # This allows the module to be aliased and query methods to be used
+  # all in one go
+  # usage: use Pinchflat.Profiles.ProfilesQuery
+  defmacro __using__(_opts) do
+    quote do
+      import Ecto.Query, warn: false
+
+      alias unquote(__MODULE__)
+    end
+  end
+
+  def new do
+    MediaProfile
+  end
+end

lib/pinchflat/prom_ex.ex

@@ -0,0 +1,40 @@
+defmodule Pinchflat.PromEx do
+  @moduledoc """
+  Configuration for the PromEx library which provides Prometheus metrics
+  """
+
+  use PromEx, otp_app: :pinchflat
+
+  alias PromEx.Plugins
+
+  @impl true
+  def plugins do
+    [
+      Plugins.Application,
+      Plugins.Beam,
+      {Plugins.Phoenix, router: PinchflatWeb.Router, endpoint: PinchflatWeb.Endpoint},
+      Plugins.Ecto,
+      Plugins.Oban,
+      Plugins.PhoenixLiveView
+    ]
+  end
+
+  @impl true
+  def dashboard_assigns do
+    [
+      default_selected_interval: "30s"
+    ]
+  end
+
+  @impl true
+  def dashboards do
+    [
+      {:prom_ex, "application.json"},
+      {:prom_ex, "beam.json"},
+      {:prom_ex, "phoenix.json"},
+      {:prom_ex, "ecto.json"},
+      {:prom_ex, "oban.json"},
+      {:prom_ex, "phoenix_live_view.json"}
+    ]
+  end
+end

lib/pinchflat/release.ex

@@ -5,6 +5,10 @@ defmodule Pinchflat.Release do
   """
   @app :pinchflat
 
+  require Logger
+
+  alias Pinchflat.Utils.FilesystemUtils
+
   def migrate do
     load_app()
 
@@ -18,6 +22,43 @@ defmodule Pinchflat.Release do
     {:ok, _, _} = Ecto.Migrator.with_repo(repo, &Ecto.Migrator.run(&1, :down, to: version))
   end
 
+  def check_file_permissions do
+    load_app()
+
+    directories =
+      [
+        "/config",
+        "/downloads",
+        "/etc/yt-dlp",
+        "/etc/yt-dlp/plugins",
+        Application.get_env(:pinchflat, :media_directory),
+        Application.get_env(:pinchflat, :tmpfile_directory),
+        Application.get_env(:pinchflat, :extras_directory),
+        Application.get_env(:pinchflat, :metadata_directory),
+        Application.get_env(:tzdata, :data_dir)
+      ]
+      |> Enum.uniq()
+      |> Enum.filter(&(&1 != nil))
+
+    Enum.each(directories, fn dir ->
+      Logger.info("Checking permissions for #{dir}")
+      filepath = Path.join([dir, ".keep"])
+
+      case FilesystemUtils.write_p(filepath, "") do
+        :ok ->
+          Logger.info("Permissions OK")
+
+        {:error, :eacces} ->
+          Logger.error(permission_denied_screed(dir))
+          raise "Permission denied"
+
+        err ->
+          Logger.error("Permissions check failed: #{inspect(err)}")
+          raise "Unknown error"
+      end
+    end)
+  end
+
   defp repos do
     Application.fetch_env!(@app, :ecto_repos)
   end
@@ -25,4 +66,31 @@ defmodule Pinchflat.Release do
   defp load_app do
     Application.load(@app)
   end
+
+  defp permission_denied_screed(dir) do
+    """
+    The directory "#{dir}" is not writeable by the Docker container.
+
+    Please ensure that the directory exists and is writeable by the Docker
+    container. All setups are different, but you may be able to run something
+    like this on the *host*:
+
+      chown nobody -R <host path that maps to #{dir}>
+      chmod 755 -R <host path that maps to #{dir}>
+
+    Swapping in your real host path. Then, you should set the user running
+    this container by editing your `docker run` command like so:
+
+        docker run --user 99:100 <rest of the command>
+
+    Or adding `user: '99:100'` to the Pinchflat service of your Docker Compose
+    file. Again, there are many ways to do this depending on your setup and
+    this is just one example. See issue #106 in the Pinchflat Github for more.
+
+    No matter the case, this _is_ a permissions error and allowing the container
+    to write to the directory is the only way to fix it. It is not recommended
+    to run the container as `root` because files created by Pinchflat may not
+    be accessible to other apps that want to modify them.
+    """
+  end
 end

lib/pinchflat/settings.ex

@@ -1,95 +0,0 @@
-defmodule Pinchflat.Settings do
-  @moduledoc """
-  The Settings context.
-  """
-
-  import Ecto.Query, warn: false
-  alias Pinchflat.Repo
-
-  alias Pinchflat.Settings.Setting
-
-  @doc """
-  Returns the list of settings.
-
-  Returns [%Setting{}, ...]
-  """
-  def list_settings do
-    Repo.all(Setting)
-  end
-
-  @doc """
-  Creates or updates a setting, returning the parsed value.
-  Raises if an unsupported datatype is used. Optionally allows
-  specifying the datatype.
-
-  Returns value in type of `Ecto.Enum.mappings(Setting, :datatype)`
-  """
-  def set!(name, value) do
-    set!(name, value, infer_datatype(value))
-  end
-
-  def set!(name, value, datatype) do
-    # Only create if doesn't exist
-    case Repo.get_by(Setting, name: to_string(name)) do
-      nil -> create_setting!(name, value, datatype)
-      setting -> update_setting!(setting, value, datatype)
-    end
-  end
-
-  @doc """
-  Gets the parsed value of a setting. Raises if the setting does not exist.
-
-  Returns value in type of `Ecto.Enum.mappings(Setting, :datatype)`
-  """
-  def get!(name) do
-    Setting
-    |> Repo.get_by!(name: to_string(name))
-    |> read_setting()
-  end
-
-  @doc """
-  Attempts to find a setting by name or creates a setting with value
-  if one doesn't exist, returning the parsed value. Optionally allows
-  specifying the datatype.
-
-  Returns value in type of `Ecto.Enum.mappings(Setting, :datatype)`
-  """
-  def fetch!(name, value) do
-    fetch!(name, value, infer_datatype(value))
-  end
-
-  def fetch!(name, value, datatype) do
-    case Repo.get_by(Setting, name: to_string(name)) do
-      nil -> create_setting!(name, value, datatype)
-      setting -> read_setting(setting)
-    end
-  end
-
-  defp change_setting(setting, attrs) do
-    Setting.changeset(setting, attrs)
-  end
-
-  defp create_setting!(name, value, datatype) do
-    %Setting{}
-    |> change_setting(%{name: to_string(name), value: to_string(value), datatype: datatype})
-    |> Repo.insert!()
-    |> read_setting()
-  end
-
-  defp update_setting!(setting, value, datatype) do
-    setting
-    |> change_setting(%{value: to_string(value), datatype: datatype})
-    |> Repo.update!()
-    |> read_setting()
-  end
-
-  defp read_setting(%{value: value, datatype: :string}), do: value
-  defp read_setting(%{value: value, datatype: :boolean}), do: value in ["true", "t", "1"]
-  defp read_setting(%{value: value, datatype: :integer}), do: String.to_integer(value)
-  defp read_setting(%{value: value, datatype: :float}), do: String.to_float(value)
-
-  defp infer_datatype(value) when is_boolean(value), do: :boolean
-  defp infer_datatype(value) when is_integer(value), do: :integer
-  defp infer_datatype(value) when is_float(value), do: :float
-  defp infer_datatype(value) when is_binary(value), do: :string
-end

lib/pinchflat/settings/setting.ex

@@ -1,24 +1,55 @@
 defmodule Pinchflat.Settings.Setting do
   @moduledoc """
-  A Setting is a key-value pair with a datatype used to track user-level settings.
+  The Setting schema.
   """
 
   use Ecto.Schema
   import Ecto.Changeset
 
-  schema "settings" do
-    field :name, :string
-    field :value, :string
-    field :datatype, Ecto.Enum, values: ~w(boolean string integer float)a
+  @allowed_fields [
+    :onboarding,
+    :pro_enabled,
+    :yt_dlp_version,
+    :apprise_version,
+    :apprise_server,
+    :video_codec_preference,
+    :audio_codec_preference,
+    :youtube_api_key,
+    :extractor_sleep_interval_seconds,
+    :download_throughput_limit,
+    :restrict_filenames
+  ]
 
-    timestamps(type: :utc_datetime)
+  @required_fields [
+    :onboarding,
+    :pro_enabled,
+    :video_codec_preference,
+    :audio_codec_preference,
+    :extractor_sleep_interval_seconds
+  ]
+
+  schema "settings" do
+    field :onboarding, :boolean, default: true
+    field :pro_enabled, :boolean, default: false
+    field :yt_dlp_version, :string
+    field :apprise_version, :string
+    field :apprise_server, :string
+    field :youtube_api_key, :string
+    field :route_token, :string
+    field :extractor_sleep_interval_seconds, :integer, default: 0
+    # This is a string because it accepts values like "100K" or "4.2M"
+    field :download_throughput_limit, :string
+    field :restrict_filenames, :boolean, default: false
+
+    field :video_codec_preference, :string
+    field :audio_codec_preference, :string
   end
 
   @doc false
   def changeset(setting, attrs) do
     setting
-    |> cast(attrs, [:name, :value, :datatype])
-    |> validate_required([:name, :value, :datatype])
-    |> unique_constraint([:name])
+    |> cast(attrs, @allowed_fields)
+    |> validate_required(@required_fields)
+    |> validate_number(:extractor_sleep_interval_seconds, greater_than_or_equal_to: 0)
   end
 end

lib/pinchflat/settings/settings.ex

@@ -0,0 +1,81 @@
+defmodule Pinchflat.Settings do
+  @moduledoc """
+  The Settings context.
+  """
+  import Ecto.Query, warn: false
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Settings.Setting
+
+  @doc """
+  Returns the only setting record. It _should_ be impossible
+  to create or delete this record, so it's assertive about
+  assuming it's the only one.
+
+  Returns %Setting{}
+  """
+  def record do
+    Setting
+    |> limit(1)
+    |> Repo.one()
+  end
+
+  @doc """
+  Updates the setting record.
+
+  Returns {:ok, %Setting{}} | {:error, %Ecto.Changeset{}}
+  """
+  def update_setting(%Setting{} = setting, attrs) do
+    setting
+    |> Setting.changeset(attrs)
+    |> Repo.update()
+  end
+
+  @doc """
+  Updates a setting, returning the new value.
+  Is setup to take a keyword list argument so you
+  can call it like `Settings.set(onboarding: true)`
+
+  Returns {:ok, value} | {:error, :invalid_key} | {:error, %Ecto.Changeset{}}
+  """
+  def set([{attr, value}]) do
+    record()
+    |> update_setting(%{attr => value})
+    |> case do
+      {:ok, %{^attr => _}} -> {:ok, value}
+      {:ok, _} -> {:error, :invalid_key}
+      {:error, changeset} -> {:error, changeset}
+    end
+  end
+
+  @doc """
+  Gets the value of a setting.
+
+  Returns {:ok, value} | {:error, :invalid_key}
+  """
+  def get(name) do
+    case Map.fetch(record(), name) do
+      {:ok, value} -> {:ok, value}
+      :error -> {:error, :invalid_key}
+    end
+  end
+
+  @doc """
+  Gets the value of a setting, raising if it doesn't exist.
+
+  Returns value
+  """
+  def get!(name) do
+    case get(name) do
+      {:ok, value} -> value
+      {:error, _} -> raise "Setting `#{name}` not found"
+    end
+  end
+
+  @doc """
+  Returns `%Ecto.Changeset{}`
+  """
+  def change_setting(%Setting{} = setting, attrs \\ %{}) do
+    Setting.changeset(setting, attrs)
+  end
+end

lib/pinchflat/slow_indexing/file_follower_server.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.Utils.FilesystemUtils.FileFollowerServer do
+defmodule Pinchflat.SlowIndexing.FileFollowerServer do
   @moduledoc """
   A GenServer that watches a file for new lines and processes them as they come in.
   This is useful for tailing log files and other similar tasks. If there's no activity
@@ -9,7 +9,7 @@ defmodule Pinchflat.Utils.FilesystemUtils.FileFollowerServer do
   require Logger
 
   @poll_interval_ms Application.compile_env(:pinchflat, :file_watcher_poll_interval)
-  @activity_timeout_ms 60_000
+  @activity_timeout_ms 600_000
 
   # Client API
   @doc """
@@ -106,7 +106,7 @@ defmodule Pinchflat.Utils.FilesystemUtils.FileFollowerServer do
         {:noreply, %{state | last_activity: DateTime.utc_now()}}
 
       :eof ->
-        Logger.debug("EOF reached, waiting before trying to read new lines")
+        Logger.debug("Current batch of media processed. Will check again in #{@poll_interval_ms}ms")
         Process.send_after(self(), :read_new_lines, @poll_interval_ms)
 
         {:noreply, state}

lib/pinchflat/slow_indexing/media_collection_indexing_worker.ex

@@ -1,25 +1,40 @@
-defmodule Pinchflat.Workers.MediaCollectionIndexingWorker do
+defmodule Pinchflat.SlowIndexing.MediaCollectionIndexingWorker do
   @moduledoc false
 
   use Oban.Worker,
     queue: :media_collection_indexing,
     unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
-    tags: ["media_source", "media_collection_indexing"]
+    tags: ["media_source", "media_collection_indexing", "show_in_dashboard"]
+
+  require Logger
 
   alias __MODULE__
   alias Pinchflat.Tasks
   alias Pinchflat.Sources
+  alias Pinchflat.Settings
   alias Pinchflat.Sources.Source
-  alias Pinchflat.Tasks.SourceTasks
-  alias Pinchflat.Workers.FastIndexingWorker
+  alias Pinchflat.FastIndexing.FastIndexingWorker
+  alias Pinchflat.SlowIndexing.SlowIndexingHelpers
+  alias Pinchflat.Lifecycle.Notifications.SourceNotifications
+
+  @doc """
+  Starts the source slow indexing worker and creates a task for the source.
+
+  Returns {:ok, %Task{}} | {:error, :duplicate_job} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff_with_task(source, job_args \\ %{}, job_opts \\ []) do
+    %{id: source.id}
+    |> Map.merge(job_args)
+    |> MediaCollectionIndexingWorker.new(job_opts)
+    |> Tasks.create_job_with_task(source)
+  end
 
-  @impl Oban.Worker
   @doc """
   The ID is that of a source _record_, not a YouTube channel/playlist ID. Indexes
   the provided source, kicks off downloads for each new MediaItem, and
   reschedules the job to run again in the future. It will ALWAYS index a source
-  if it's never been indexed before, but rescheduling is determined by the
-  `index_frequency_minutes` field.
+  if it's never been indexed before or if `force` is set to `true`, but rescheduling
+  is determined by the `index_frequency_minutes` field.
 
   README: Re-scheduling here works a little different than you may expect.
   The reschedule time is relative to the time the job has actually _completed_.
@@ -40,15 +55,14 @@ defmodule Pinchflat.Workers.MediaCollectionIndexingWorker do
        by the `download_media` field on the source as well as the profile's
        shorts/livestream behaviour. At this step we also attach a file reader
        to the `yt-dlp` output file so we can create media items as they come in
-       for a little speedup (see SourceTasks comments for more)
+       for a little speedup (see {Fast,Slow}IndexingHelpers comments for more)
     4. If this job is meant to reschedule (ie: has an index frequency > 0),
        it reschedules itself. If not, it runs once and does not reschedule
     5. If the source uses fast indexing, that job is kicked off as well. It
        uses RSS to run a smaller, faster, and more frequent index. That job
        handles rescheduling itself but largely has a similar behaviour to this
-       job in that it kicks off index and maybe download jobs. The biggest difference
-       is that an index job is kicked off _for each new media item_ as opposed
-       to one larger index job. Check out `MediaIndexingWorker` comments for more.
+       job in that it runs and index and maybe kicks off media download jobs.
+       Check out `FastIndexingWorker` comments for more.
     6. If the job reschedules, the cycle from step 3 repeats until the heat death
        of the universe. The user changing things like the index frequency can
        dequeue or reschedule jobs as well
@@ -56,31 +70,45 @@ defmodule Pinchflat.Workers.MediaCollectionIndexingWorker do
   NOTE: Since indexing can take a LONG time, I should check what happens if an
   application restart occurs while a job is running. Will the job be lost?
 
-  IDEA: Should I use paging and do indexing in chunks? Is that even faster?
-
   Returns :ok | {:ok, %Task{}}
   """
-  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => source_id} = args}) do
     source = Sources.get_source!(source_id)
 
     case {source.index_frequency_minutes, source.last_indexed_at} do
       {index_freq, _} when index_freq > 0 ->
         # If the indexing is on a schedule simply run indexing and reschedule
-        SourceTasks.index_and_enqueue_download_for_media_items(source)
+        perform_indexing_and_notification(source, was_forced: args["force"])
         maybe_enqueue_fast_indexing_task(source)
         reschedule_indexing(source)
 
       {_, nil} ->
         # If the source has never been indexed, index it once
         # even if it's not meant to reschedule
-        SourceTasks.index_and_enqueue_download_for_media_items(source)
+        perform_indexing_and_notification(source, was_forced: args["force"])
         :ok
 
       _ ->
         # If the source HAS been indexed and is not meant to reschedule,
-        # perform a no-op
+        # perform a no-op (unless forced)
+        if args["force"] do
+          perform_indexing_and_notification(source, was_forced: true)
+        end
+
         :ok
     end
+  rescue
+    Ecto.NoResultsError -> Logger.info("#{__MODULE__} discarded: source #{source_id} not found")
+    Ecto.StaleEntryError -> Logger.info("#{__MODULE__} discarded: source #{source_id} stale")
+  end
+
+  defp perform_indexing_and_notification(source, indexing_opts) do
+    apprise_server = Settings.get!(:apprise_server)
+
+    SourceNotifications.wrap_new_media_notification(apprise_server, source, fn ->
+      SlowIndexingHelpers.index_and_enqueue_download_for_media_items(source, indexing_opts)
+    end)
   end
 
   defp reschedule_indexing(source) do

lib/pinchflat/slow_indexing/slow_indexing_helpers.ex

@@ -0,0 +1,244 @@
+defmodule Pinchflat.SlowIndexing.SlowIndexingHelpers do
+  @moduledoc """
+  Methods for performing slow indexing tasks and managing the indexing process.
+
+  Many of these methods are made to be kickoff or be consumed by workers.
+  """
+
+  use Pinchflat.Media.MediaQuery
+
+  require Logger
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Media.MediaItem
+  alias Pinchflat.YtDlp.MediaCollection
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Downloading.DownloadingHelpers
+  alias Pinchflat.SlowIndexing.FileFollowerServer
+  alias Pinchflat.Downloading.DownloadOptionBuilder
+  alias Pinchflat.SlowIndexing.MediaCollectionIndexingWorker
+
+  alias Pinchflat.YtDlp.Media, as: YtDlpMedia
+
+  @doc """
+  Kills old indexing tasks and starts a new task to index the media collection.
+
+  The job is delayed based on the source's `index_frequency_minutes` setting unless
+  one of the following is true:
+    - The `force` option is set to true
+    - The source has never been indexed before
+    - The source has been indexed before, but the last indexing job was more than
+      `index_frequency_minutes` ago
+
+  Returns {:ok, %Task{}}
+  """
+  def kickoff_indexing_task(%Source{} = source, job_args \\ %{}, job_opts \\ []) do
+    job_offset_seconds = if job_args[:force], do: 0, else: calculate_job_offset_seconds(source)
+
+    Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker", include_executing: true)
+
+    MediaCollectionIndexingWorker.kickoff_with_task(source, job_args, job_opts ++ [schedule_in: job_offset_seconds])
+  end
+
+  @doc """
+  A helper method to delete all indexing-related tasks for a source.
+  Optionally, you can include executing tasks in the deletion process.
+
+  Returns :ok
+  """
+  def delete_indexing_tasks(%Source{} = source, opts \\ []) do
+    include_executing = Keyword.get(opts, :include_executing, false)
+
+    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker", include_executing: include_executing)
+    Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker", include_executing: include_executing)
+  end
+
+  @doc """
+  Given a media source, creates (indexes) the media by creating media_items for each
+  media ID in the source. Afterward, kicks off a download task for each pending media
+  item belonging to the source. Returns a list of media items or changesets
+  (if the media item couldn't be created).
+
+  Indexing is slow and usually returns a list of all media data at once for record creation.
+  To help with this, we use a file follower to watch the file that yt-dlp writes to
+  so we can create media items as they come in. This parallelizes the process and adds
+  clarity to the user experience. This has a few things to be aware of which are documented
+  below in the file watcher setup method.
+
+  Additionally, in the case of a repeat index we create a download archive file that
+  contains some media IDs that we've indexed in the past. Note that this archive doesn't
+  contain the most recent IDs but rather a subset of IDs that are offset by some amount.
+  Practically, this means that we'll re-index a small handful of media that we've recently
+  indexed, but this is a good thing since it'll let us pick up on any recent changes to the
+  most recent media items.
+
+  We don't create a download archive for playlists (only channels), nor do we create one if
+  the indexing was forced by the user.
+
+  NOTE: downloads are only enqueued if the source is set to download media. Downloads are
+  also enqueued for ALL pending media items, not just the ones that were indexed in this
+  job run. This should ensure that any stragglers are caught if, for some reason, they
+  weren't enqueued or somehow got de-queued.
+
+  Available options:
+    - `was_forced`: Whether the indexing was forced by the user
+
+  Returns [%MediaItem{} | %Ecto.Changeset{}]
+  """
+  def index_and_enqueue_download_for_media_items(%Source{} = source, opts \\ []) do
+    # The media_profile is needed to determine the quality options to _then_ determine a more
+    # accurate predicted filepath
+    source = Repo.preload(source, [:media_profile])
+    # See the method definition below for more info on how file watchers work
+    # (important reading if you're not familiar with it)
+    {:ok, media_attributes} = setup_file_watcher_and_kickoff_indexing(source, opts)
+    # Reload because the source may have been updated during the (long-running) indexing process
+    # and important settings like `download_media` may have changed.
+    source = Repo.reload!(source)
+
+    result =
+      Enum.map(media_attributes, fn media_attrs ->
+        case Media.create_media_item_from_backend_attrs(source, media_attrs) do
+          {:ok, media_item} -> media_item
+          {:error, changeset} -> changeset
+        end
+      end)
+
+    Sources.update_source(source, %{last_indexed_at: DateTime.utc_now()})
+    DownloadingHelpers.enqueue_pending_download_tasks(source)
+
+    result
+  end
+
+  # The file follower is a GenServer that watches a file for new lines and
+  # processes them. This works well, but we have to be resilliant to partially-written
+  # lines (ie: you should gracefully fail if you can't parse a line).
+  #
+  # This works in-tandem with the normal (blocking) media indexing behaviour. When
+  # the `setup_file_watcher_and_kickoff_indexing` method completes it'll return the
+  # FULL result to the caller for parsing. Ideally, every item in the list will have already
+  # been processed by the file follower, but if not, the caller handles creation
+  # of any media items that were missed/initially failed.
+  #
+  # It attempts a graceful shutdown of the file follower after the indexing is done,
+  # but the FileFollowerServer will also stop itself if it doesn't see any activity
+  # for a sufficiently long time.
+  defp setup_file_watcher_and_kickoff_indexing(source, opts) do
+    was_forced = Keyword.get(opts, :was_forced, false)
+    {:ok, pid} = FileFollowerServer.start_link()
+
+    handler = fn filepath -> setup_file_follower_watcher(pid, filepath, source) end
+    should_use_cookies = Sources.use_cookies?(source, :indexing)
+
+    command_opts =
+      [output: DownloadOptionBuilder.build_output_path_for(source)] ++
+        DownloadOptionBuilder.build_quality_options_for(source) ++
+        build_download_archive_options(source, was_forced)
+
+    runner_opts = [file_listener_handler: handler, use_cookies: should_use_cookies]
+    result = MediaCollection.get_media_attributes_for_collection(source.original_url, command_opts, runner_opts)
+
+    FileFollowerServer.stop(pid)
+
+    result
+  end
+
+  defp setup_file_follower_watcher(pid, filepath, source) do
+    FileFollowerServer.watch_file(pid, filepath, fn line ->
+      case Phoenix.json_library().decode(line) do
+        {:ok, media_attrs} ->
+          Logger.debug("FileFollowerServer Handler: Got media attributes: #{inspect(media_attrs)}")
+
+          media_struct = YtDlpMedia.response_to_struct(media_attrs)
+          create_media_item_and_enqueue_download(source, media_struct)
+
+        err ->
+          Logger.debug("FileFollowerServer Handler: Error decoding JSON: #{inspect(err)}")
+
+          err
+      end
+    end)
+  end
+
+  defp create_media_item_and_enqueue_download(source, media_attrs) do
+    # Reload because the source may have been updated during the (long-running) indexing process
+    # and important settings like `download_media` may have changed.
+    source = Repo.reload!(source)
+
+    case Media.create_media_item_from_backend_attrs(source, media_attrs) do
+      {:ok, %MediaItem{} = media_item} ->
+        DownloadingHelpers.kickoff_download_if_pending(media_item)
+
+      {:error, changeset} ->
+        changeset
+    end
+  end
+
+  # Find the difference between the current time and the last time the source was indexed
+  defp calculate_job_offset_seconds(%Source{last_indexed_at: nil}), do: 0
+
+  defp calculate_job_offset_seconds(source) do
+    offset_seconds = DateTime.diff(DateTime.utc_now(), source.last_indexed_at, :second)
+    index_frequency_seconds = source.index_frequency_minutes * 60
+
+    max(0, index_frequency_seconds - offset_seconds)
+  end
+
+  # The download archive file works in tandem with --break-on-existing to stop
+  # yt-dlp once we've hit media items we've already indexed. But we generate
+  # this list with a bit of an offset so we do intentionally re-scan some media
+  # items to pick up any recent changes (see `get_media_items_for_download_archive`).
+  #
+  # From there, we format the media IDs in the way that yt-dlp expects (ie: "<extractor> <media_id>")
+  # and return the filepath to the caller.
+  defp create_download_archive_file(source) do
+    tmpfile = FilesystemUtils.generate_metadata_tmpfile(:txt)
+
+    archive_contents =
+      source
+      |> get_media_items_for_download_archive()
+      |> Enum.map_join("\n", fn media_item -> "youtube #{media_item.media_id}" end)
+
+    case File.write(tmpfile, archive_contents) do
+      :ok -> tmpfile
+      err -> err
+    end
+  end
+
+  # Sorting by `uploaded_at` is important because we want to re-index the most recent
+  # media items first but there is no guarantee of any correlation between ID and uploaded_at.
+  #
+  # The offset is important because we want to re-index some media items that we've
+  # recently indexed to pick up on any changes. The limit is because we want this mechanism
+  # to work even if, for example, the video we were using as a stopping point was deleted.
+  # It's not a perfect system, but it should do well enough.
+  #
+  # The chosen limit and offset are arbitary, independent, and vibes-based. Feel free to
+  # tweak as-needed
+  defp get_media_items_for_download_archive(source) do
+    MediaQuery.new()
+    |> where(^MediaQuery.for_source(source))
+    |> order_by(desc: :uploaded_at)
+    |> limit(50)
+    |> offset(20)
+    |> Repo.all()
+  end
+
+  # The download archive isn't useful for playlists (since those are ordered arbitrarily)
+  # and we don't want to use it if the indexing was forced by the user. In other words,
+  # only create an archive for channels that are being indexed as part of their regular
+  # indexing schedule. The first indexing pass should also not create an archive.
+  defp build_download_archive_options(%Source{collection_type: :playlist}, _was_forced), do: []
+  defp build_download_archive_options(%Source{last_indexed_at: nil}, _was_forced), do: []
+  defp build_download_archive_options(_source, true), do: []
+
+  defp build_download_archive_options(source, _was_forced) do
+    archive_file = create_download_archive_file(source)
+
+    [:break_on_existing, download_archive: archive_file]
+  end
+end

lib/pinchflat/sources.ex

@@ -1,263 +0,0 @@
-defmodule Pinchflat.Sources do
-  @moduledoc """
-  The Sources context.
-  """
-
-  import Ecto.Query, warn: false
-  alias Pinchflat.Repo
-
-  alias Pinchflat.Media
-  alias Pinchflat.Tasks
-  alias Pinchflat.Sources.Source
-  alias Pinchflat.Tasks.SourceTasks
-  alias Pinchflat.Profiles.MediaProfile
-  alias Pinchflat.YtDlp.Backend.MediaCollection
-
-  @doc """
-  Returns the list of sources. Returns [%Source{}, ...]
-  """
-  def list_sources do
-    Repo.all(Source)
-  end
-
-  @doc """
-  Returns the list of sources for a media_profile.
-
-  Returns [%Source{}, ...]
-  """
-  def list_sources_for(%MediaProfile{} = media_profile) do
-    Repo.all(from s in Source, where: s.media_profile_id == ^media_profile.id)
-  end
-
-  @doc """
-  Gets a single source.
-
-  Returns %Source{}. Raises `Ecto.NoResultsError` if the Source does not exist.
-  """
-  def get_source!(id), do: Repo.get!(Source, id)
-
-  @doc """
-  Creates a source. May attempt to pull additional source details from the
-  original_url (if provided). Will attempt to start indexing the source's
-  media if successfully inserted.
-
-  Runs an initial `change_source` check to ensure most of the source is valid
-  before making an expensive API call. Runs it through `Repo.insert` even
-  though we know it's going to fail so it picks up any addl. database errors
-  and fulfills our return contract.
-
-  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
-  """
-  def create_source(attrs) do
-    case change_source(%Source{}, attrs, :initial) do
-      %Ecto.Changeset{valid?: true} ->
-        %Source{}
-        |> change_source_from_url(attrs)
-        |> maybe_change_indexing_frequency()
-        |> commit_and_handle_tasks()
-
-      changeset ->
-        Repo.insert(changeset)
-    end
-  end
-
-  @doc """
-  Updates a source. May attempt to pull additional source details from the
-  original_url (if changed). May attempt to start indexing the source's
-  media if the indexing frequency has been changed.
-
-  Existing indexing tasks will be cancelled if the indexing frequency has been
-  changed (logic in `SourceTasks.kickoff_indexing_task`)
-
-  Runs an initial `change_source` check to ensure most of the source is valid
-  before making an expensive API call. Runs it through `Repo.update` even
-  though we know it's going to fail so it picks up any addl. database errors
-  and fulfills our return contract.
-
-  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
-  """
-  def update_source(%Source{} = source, attrs) do
-    case change_source(source, attrs, :initial) do
-      %Ecto.Changeset{valid?: true} ->
-        source
-        |> change_source_from_url(attrs)
-        |> maybe_change_indexing_frequency()
-        |> commit_and_handle_tasks()
-
-      changeset ->
-        Repo.update(changeset)
-    end
-  end
-
-  @doc """
-  Deletes a source, its media items, and its associated tasks (of any state).
-  Can optionally delete the source's media files.
-
-  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
-  """
-  def delete_source(%Source{} = source, opts \\ []) do
-    delete_files = Keyword.get(opts, :delete_files, false)
-
-    source
-    |> Media.list_media_items_for()
-    |> Enum.each(fn media_item ->
-      Media.delete_media_item(media_item, delete_files: delete_files)
-    end)
-
-    Tasks.delete_tasks_for(source)
-    Repo.delete(source)
-  end
-
-  @doc """
-  Returns an `%Ecto.Changeset{}` for tracking source changes.
-  """
-  def change_source(%Source{} = source, attrs \\ %{}, validation_stage \\ :pre_insert) do
-    Source.changeset(source, attrs, validation_stage)
-  end
-
-  @doc """
-  Returns an `%Ecto.Changeset{}` for tracking source changes and additionally
-  fetches source details from the original_url (if provided). If the source
-  details cannot be fetched, an error is added to the changeset.
-
-  Note that this fetches source details as long as the `original_url` is present.
-  This means that it'll go for it even if a changeset is otherwise invalid. This
-  is pretty easy to change, but for MVP I'm not concerned.
-
-  NOTE: When operating in the ideal path, this effectively adds an API call
-  to the source creation/update process. Should be used only when needed.
-  """
-  def change_source_from_url(%Source{} = source, attrs) do
-    case change_source(source, attrs) do
-      %Ecto.Changeset{changes: %{original_url: _}} = changeset ->
-        add_source_details_to_changeset(source, changeset)
-
-      changeset ->
-        changeset
-    end
-  end
-
-  defp add_source_details_to_changeset(source, changeset) do
-    %Ecto.Changeset{changes: changes} = changeset
-
-    case MediaCollection.get_source_details(changes.original_url) do
-      {:ok, source_details} ->
-        add_source_details_by_collection_type(source, changeset, source_details)
-
-      {:error, runner_error, _status_code} ->
-        Ecto.Changeset.add_error(
-          changeset,
-          :original_url,
-          "could not fetch source details from URL",
-          error: runner_error
-        )
-    end
-  end
-
-  defp add_source_details_by_collection_type(source, changeset, source_details) do
-    %Ecto.Changeset{changes: changes} = changeset
-
-    collection_changes =
-      if source_details.playlist_id == source_details.channel_id do
-        %{
-          collection_type: :channel,
-          collection_id: source_details.channel_id,
-          collection_name: source_details.channel_name
-        }
-      else
-        %{
-          collection_type: :playlist,
-          collection_id: source_details.playlist_id,
-          collection_name: source_details.playlist_name
-        }
-      end
-
-    change_source(source, Map.merge(changes, collection_changes))
-  end
-
-  defp maybe_change_indexing_frequency(changeset) do
-    fast_index = Ecto.Changeset.get_field(changeset, :fast_index)
-
-    if fast_index do
-      Ecto.Changeset.put_change(
-        changeset,
-        :index_frequency_minutes,
-        Source.index_frequency_when_fast_indexing()
-      )
-    else
-      changeset
-    end
-  end
-
-  defp commit_and_handle_tasks(changeset) do
-    case Repo.insert_or_update(changeset) do
-      {:ok, %Source{} = source} ->
-        maybe_handle_media_tasks(changeset, source)
-        maybe_run_indexing_task(changeset, source)
-
-      err ->
-        err
-    end
-  end
-
-  # If the source is NOT new (ie: updated) and the download_media flag has changed,
-  # enqueue or dequeue media download tasks as necessary.
-  defp maybe_handle_media_tasks(changeset, source) do
-    case {changeset.data, changeset.changes} do
-      {%{__meta__: %{state: :loaded}}, %{download_media: true}} ->
-        SourceTasks.enqueue_pending_media_tasks(source)
-
-      {%{__meta__: %{state: :loaded}}, %{download_media: false}} ->
-        SourceTasks.dequeue_pending_media_tasks(source)
-
-      _ ->
-        :ok
-    end
-
-    {:ok, source}
-  end
-
-  defp maybe_run_indexing_task(changeset, source) do
-    case changeset.data do
-      # If the changeset is new (not persisted), attempt indexing no matter what
-      %{__meta__: %{state: :built}} ->
-        SourceTasks.kickoff_indexing_task(source)
-
-      # If the record has been persisted, only run indexing if the
-      # indexing frequency has been changed and is now greater than 0
-      %{__meta__: %{state: :loaded}} ->
-        maybe_update_slow_indexing_task(changeset, source)
-        maybe_update_fast_indexing_task(changeset, source)
-    end
-
-    {:ok, source}
-  end
-
-  defp maybe_update_slow_indexing_task(changeset, source) do
-    case changeset.changes do
-      %{index_frequency_minutes: mins} when mins > 0 ->
-        SourceTasks.kickoff_indexing_task(source)
-
-      %{index_frequency_minutes: _} ->
-        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
-        Tasks.delete_pending_tasks_for(source, "MediaIndexingWorker")
-        Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker")
-
-      _ ->
-        :ok
-    end
-  end
-
-  defp maybe_update_fast_indexing_task(changeset, source) do
-    case changeset.changes do
-      %{fast_index: true} ->
-        SourceTasks.kickoff_fast_indexing_task(source)
-
-      %{fast_index: false} ->
-        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
-
-      _ ->
-        :ok
-    end
-  end
-end

lib/pinchflat/sources/source.ex

@@ -7,28 +7,46 @@ defmodule Pinchflat.Sources.Source do
   import Ecto.Changeset
   import Pinchflat.Utils.ChangesetUtils
 
+  alias __MODULE__
+  alias Pinchflat.Repo
   alias Pinchflat.Tasks.Task
   alias Pinchflat.Media.MediaItem
   alias Pinchflat.Profiles.MediaProfile
+  alias Pinchflat.Metadata.SourceMetadata
 
   @allowed_fields ~w(
+    enabled
     collection_name
     collection_id
     collection_type
     custom_name
+    description
+    nfo_filepath
+    poster_filepath
+    fanart_filepath
+    banner_filepath
+    series_directory
     index_frequency_minutes
     fast_index
+    cookie_behaviour
     download_media
     last_indexed_at
     original_url
     download_cutoff_date
+    retention_period_days
+    title_filter_regex
     media_profile_id
+    output_path_template_override
+    marked_for_deletion_at
+    min_duration_seconds
+    max_duration_seconds
   )a
 
   # Expensive API calls are made when a source is inserted/updated so
   # we want to ensure that the source is valid before making the call.
   # This way, we check that the other attributes are valid before ensuring
-  # that all fields are valid.
+  # that all fields are valid. This is still only one DB insert but it's
+  # a two-stage validation process to fail fast before the API call.
   @initially_required_fields ~w(
     index_frequency_minutes
     fast_index
@@ -39,6 +57,7 @@ defmodule Pinchflat.Sources.Source do
 
   @pre_insert_required_fields @initially_required_fields ++
                                 ~w(
+                                  uuid
                                   custom_name
                                   collection_name
                                   collection_id
@@ -46,20 +65,44 @@ defmodule Pinchflat.Sources.Source do
                                 )a
 
   schema "sources" do
+    field :enabled, :boolean, default: true
+    # This is _not_ used as the primary key or internally in the database
+    # relations. This is only used to prevent an enumeration attack on the streaming
+    # and RSS feed endpoints since those _must_ be public (ie: no basic auth)
+    field :uuid, Ecto.UUID
+
     field :custom_name, :string
+    field :description, :string
     field :collection_name, :string
     field :collection_id, :string
     field :collection_type, Ecto.Enum, values: [:channel, :playlist]
     field :index_frequency_minutes, :integer, default: 60 * 24
     field :fast_index, :boolean, default: false
+    field :cookie_behaviour, Ecto.Enum, values: [:disabled, :when_needed, :all_operations], default: :disabled
     field :download_media, :boolean, default: true
     field :last_indexed_at, :utc_datetime
     # Only download media items that were published after this date
     field :download_cutoff_date, :date
+    field :retention_period_days, :integer
     field :original_url, :string
+    field :title_filter_regex, :string
+    field :output_path_template_override, :string
+
+    field :min_duration_seconds, :integer
+    field :max_duration_seconds, :integer
+
+    field :series_directory, :string
+    field :nfo_filepath, :string
+    field :poster_filepath, :string
+    field :fanart_filepath, :string
+    field :banner_filepath, :string
+
+    field :marked_for_deletion_at, :utc_datetime
 
     belongs_to :media_profile, MediaProfile
 
+    has_one :metadata, SourceMetadata, on_replace: :update
+
     has_many :tasks, Task
     has_many :media_items, MediaItem, foreign_key: :source_id
 
@@ -79,8 +122,16 @@ defmodule Pinchflat.Sources.Source do
     source
     |> cast(attrs, @allowed_fields)
     |> dynamic_default(:custom_name, fn cs -> get_field(cs, :collection_name) end)
+    |> dynamic_default(:uuid, fn _ -> Ecto.UUID.generate() end)
     |> validate_required(required_fields)
-    |> unique_constraint([:collection_id, :media_profile_id])
+    |> validate_title_regex()
+    |> validate_min_and_max_durations()
+    |> validate_number(:retention_period_days, greater_than_or_equal_to: 0)
+    # Ensures it ends with `.{{ ext }}` or `.%(ext)s` or similar (with a little wiggle room)
+    |> validate_format(:output_path_template_override, MediaProfile.ext_regex(), message: "must end with .{{ ext }}")
+    |> validate_format(:original_url, youtube_channel_or_playlist_regex(), message: "must be a channel or playlist URL")
+    |> cast_assoc(:metadata, with: &SourceMetadata.changeset/2, required: false)
+    |> unique_constraint([:collection_id, :media_profile_id, :title_filter_regex], error_key: :original_url)
   end
 
   @doc false
@@ -92,6 +143,52 @@ defmodule Pinchflat.Sources.Source do
   @doc false
   def fast_index_frequency do
     # minutes
-    15
+    10
+  end
+
+  @doc false
+  def filepath_attributes do
+    ~w(nfo_filepath fanart_filepath poster_filepath banner_filepath)a
+  end
+
+  @doc false
+  def json_exluded_fields do
+    ~w(__meta__ __struct__ metadata tasks media_items)a
+  end
+
+  def youtube_channel_or_playlist_regex do
+    # Validate that the original URL is not a video URL
+    # Also matches if the string does NOT contain youtube.com or youtu.be. This preserves my tenuous support
+    # for non-youtube sources.
+    ~r<^(?:(?!youtube\.com/(watch|shorts|embed)|youtu\.be).)*$>
+  end
+
+  defp validate_title_regex(%{changes: %{title_filter_regex: regex}} = changeset) when is_binary(regex) do
+    case Ecto.Adapters.SQL.query(Repo, "SELECT regexp_like('', ?)", [regex]) do
+      {:ok, _} -> changeset
+      _ -> add_error(changeset, :title_filter_regex, "is invalid")
+    end
+  end
+
+  defp validate_title_regex(changeset), do: changeset
+
+  defp validate_min_and_max_durations(changeset) do
+    min_duration = get_change(changeset, :min_duration_seconds)
+    max_duration = get_change(changeset, :max_duration_seconds)
+
+    case {min_duration, max_duration} do
+      {min, max} when is_nil(min) or is_nil(max) -> changeset
+      {min, max} when min >= max -> add_error(changeset, :max_duration_seconds, "must be greater than minumum duration")
+      _ -> changeset
+    end
+  end
+
+  defimpl Jason.Encoder, for: Source do
+    def encode(value, opts) do
+      value
+      |> Repo.preload(:media_profile)
+      |> Map.drop(Source.json_exluded_fields())
+      |> Jason.Encode.map(opts)
+    end
   end
 end

lib/pinchflat/sources/source_deletion_worker.ex

@@ -0,0 +1,38 @@
+defmodule Pinchflat.Sources.SourceDeletionWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :local_data,
+    tags: ["sources", "local_data"]
+
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Sources
+
+  @doc """
+  Starts the source deletion worker. Does not attach it to a task like `kickoff_with_task/2`
+  since deletion also cancels all tasks for the source
+
+  Returns {:ok, %Task{}} | {:error, %Ecto.Changeset{}}
+  """
+  def kickoff(source, job_args \\ %{}, job_opts \\ []) do
+    %{id: source.id}
+    |> Map.merge(job_args)
+    |> SourceDeletionWorker.new(job_opts)
+    |> Oban.insert()
+  end
+
+  @doc """
+  Deletes a source and optionally deletes its files
+
+  Returns :ok
+  """
+  @impl Oban.Worker
+  def perform(%Oban.Job{args: %{"id" => source_id} = args}) do
+    delete_files = Map.get(args, "delete_files", false)
+    source = Sources.get_source!(source_id)
+
+    Sources.delete_source(source, delete_files: delete_files)
+  end
+end

lib/pinchflat/sources/sources.ex

@@ -0,0 +1,392 @@
+defmodule Pinchflat.Sources do
+  @moduledoc """
+  The Sources context.
+  """
+
+  import Ecto.Query, warn: false
+  use Pinchflat.Media.MediaQuery
+
+  alias Pinchflat.Repo
+  alias Pinchflat.Media
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Profiles.MediaProfile
+  alias Pinchflat.YtDlp.MediaCollection
+  alias Pinchflat.Metadata.SourceMetadata
+  alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.Downloading.DownloadingHelpers
+  alias Pinchflat.SlowIndexing.SlowIndexingHelpers
+  alias Pinchflat.FastIndexing.FastIndexingHelpers
+  alias Pinchflat.Metadata.SourceMetadataStorageWorker
+
+  @doc """
+  Returns the relevant output path template for a source.
+  Pulls from the source's override if present, otherwise uses the media profile's.
+
+  Returns binary()
+  """
+  def output_path_template(source) do
+    source = Repo.preload(source, :media_profile)
+    media_profile = source.media_profile
+
+    source.output_path_template_override || media_profile.output_path_template
+  end
+
+  @doc """
+  Returns a boolean indicating whether or not cookies should be used for a given operation.
+
+  Returns boolean()
+  """
+  def use_cookies?(source, operation) when operation in [:indexing, :downloading, :metadata, :error_recovery] do
+    case source.cookie_behaviour do
+      :disabled -> false
+      :all_operations -> true
+      :when_needed -> operation in [:indexing, :error_recovery]
+    end
+  end
+
+  @doc """
+  Returns the list of sources. Returns [%Source{}, ...]
+  """
+  def list_sources do
+    Repo.all(Source)
+  end
+
+  @doc """
+  Returns the list of sources for a media_profile.
+
+  Returns [%Source{}, ...]
+  """
+  def list_sources_for(%MediaProfile{} = media_profile) do
+    Repo.all(from s in Source, where: s.media_profile_id == ^media_profile.id)
+  end
+
+  @doc """
+  Gets a single source.
+
+  Returns %Source{}. Raises `Ecto.NoResultsError` if the Source does not exist.
+  """
+  def get_source!(id), do: Repo.get!(Source, id)
+
+  @doc """
+  Creates a source. May attempt to pull additional source details from the
+  original_url (if provided). Will attempt to start indexing the source's
+  media if successfully inserted.
+
+  Runs an initial `change_source` check to ensure most of the source is valid
+  before making an expensive API call. Runs it through `Repo.insert` even
+  though we know it's going to fail so it picks up any addl. database errors
+  and fulfills our return contract.
+
+  You can pass options to control the behavior of the function:
+    - `run_post_commit_tasks` (default: true) - If false, the function will not
+      enqueue any tasks in `commit_and_handle_tasks`.
+
+  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
+  """
+  def create_source(attrs, opts \\ []) do
+    case change_source(%Source{}, attrs, :initial) do
+      %Ecto.Changeset{valid?: true} ->
+        %Source{}
+        |> maybe_change_source_from_url(attrs)
+        |> maybe_change_indexing_frequency()
+        |> commit_and_handle_tasks(opts)
+
+      changeset ->
+        Repo.insert(changeset)
+    end
+  end
+
+  @doc """
+  Updates a source. May attempt to pull additional source details from the
+  original_url (if changed). May attempt to start indexing the source's
+  media if the indexing frequency has been changed.
+
+  Existing indexing tasks will be cancelled if the indexing frequency has been
+  changed (logic in `SlowIndexingHelpers.kickoff_indexing_task`)
+
+  Runs an initial `change_source` check to ensure most of the source is valid
+  before making an expensive API call. Runs it through `Repo.update` even
+  though we know it's going to fail so it picks up any addl. database errors
+  and fulfills our return contract.
+
+  You can pass options to control the behavior of the function:
+    - `run_post_commit_tasks` (default: true) - If false, the function will not
+      enqueue any tasks in `commit_and_handle_tasks`.
+
+  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
+  """
+  def update_source(%Source{} = source, attrs, opts \\ []) do
+    case change_source(source, attrs, :initial) do
+      %Ecto.Changeset{valid?: true} ->
+        source
+        |> maybe_change_source_from_url(attrs)
+        |> maybe_change_indexing_frequency()
+        |> commit_and_handle_tasks(opts)
+
+      changeset ->
+        Repo.update(changeset)
+    end
+  end
+
+  @doc """
+  Deletes a source, its media items, and its associated tasks (of any state).
+  Can optionally delete the source's media files.
+
+  Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
+  """
+  def delete_source(%Source{} = source, opts \\ []) do
+    delete_files = Keyword.get(opts, :delete_files, false)
+    Tasks.delete_tasks_for(source)
+
+    MediaQuery.new()
+    |> where(^MediaQuery.for_source(source))
+    |> Repo.all()
+    |> Enum.each(fn media_item ->
+      Media.delete_media_item(media_item, delete_files: delete_files)
+    end)
+
+    if delete_files do
+      delete_source_files(source)
+    end
+
+    delete_internal_metadata_files(source)
+    Repo.delete(source)
+  end
+
+  @doc """
+  Returns an `%Ecto.Changeset{}` for tracking source changes.
+  """
+  def change_source(%Source{} = source, attrs \\ %{}, validation_stage \\ :pre_insert) do
+    Source.changeset(source, attrs, validation_stage)
+  end
+
+  # NOTE: When operating in the ideal path, this effectively adds an API call
+  # to the source creation/update process. Should be used only when needed.
+  defp maybe_change_source_from_url(%Source{} = source, attrs) do
+    case change_source(source, attrs) do
+      %Ecto.Changeset{changes: %{original_url: _}} = changeset ->
+        add_source_details_to_changeset(source, changeset)
+
+      changeset ->
+        changeset
+    end
+  end
+
+  defp delete_source_files(source) do
+    mapped_struct = Map.from_struct(source)
+
+    Source.filepath_attributes()
+    |> Enum.map(fn field -> mapped_struct[field] end)
+    |> Enum.filter(&is_binary/1)
+    |> Enum.each(&FilesystemUtils.delete_file_and_remove_empty_directories/1)
+  end
+
+  defp delete_internal_metadata_files(source) do
+    metadata = Repo.preload(source, :metadata).metadata || %SourceMetadata{}
+    mapped_struct = Map.from_struct(metadata)
+
+    SourceMetadata.filepath_attributes()
+    |> Enum.map(fn field -> mapped_struct[field] end)
+    |> Enum.filter(&is_binary/1)
+    |> Enum.each(&FilesystemUtils.delete_file_and_remove_empty_directories/1)
+  end
+
+  defp add_source_details_to_changeset(source, changeset) do
+    original_url = changeset.changes.original_url
+    should_use_cookies = Ecto.Changeset.get_field(changeset, :cookie_behaviour) == :all_operations
+    # Skipping sleep interval since this is UI blocking and we want to keep this as fast as possible
+    addl_opts = [use_cookies: should_use_cookies, skip_sleep_interval: true]
+
+    case MediaCollection.get_source_details(original_url, [], addl_opts) do
+      {:ok, source_details} ->
+        add_source_details_by_collection_type(source, changeset, source_details)
+
+      err ->
+        runner_error =
+          case err do
+            {:error, error_msg, _status_code} -> error_msg
+            {:error, error_msg} -> error_msg
+          end
+
+        Ecto.Changeset.add_error(
+          changeset,
+          :original_url,
+          "could not fetch source details from URL",
+          error: runner_error
+        )
+    end
+  end
+
+  defp add_source_details_by_collection_type(source, changeset, source_details) do
+    %Ecto.Changeset{changes: changes} = changeset
+
+    collection_changes =
+      if source_details.playlist_id == source_details.channel_id do
+        %{
+          collection_type: :channel,
+          collection_id: source_details.channel_id,
+          collection_name: source_details.channel_name
+        }
+      else
+        %{
+          collection_type: :playlist,
+          collection_id: source_details.playlist_id,
+          collection_name: source_details.playlist_name
+        }
+      end
+
+    change_source(source, Map.merge(changes, collection_changes))
+  end
+
+  defp maybe_change_indexing_frequency(changeset) do
+    fast_index = Ecto.Changeset.get_field(changeset, :fast_index)
+
+    if fast_index do
+      Ecto.Changeset.put_change(
+        changeset,
+        :index_frequency_minutes,
+        Source.index_frequency_when_fast_indexing()
+      )
+    else
+      changeset
+    end
+  end
+
+  defp commit_and_handle_tasks(changeset, opts) do
+    run_post_commit_tasks = Keyword.get(opts, :run_post_commit_tasks, true)
+
+    case Repo.insert_or_update(changeset) do
+      {:ok, %Source{} = source} ->
+        if run_post_commit_tasks do
+          maybe_handle_media_tasks(changeset, source)
+          maybe_run_indexing_task(changeset, source)
+          maybe_run_metadata_storage_task(changeset, source)
+        end
+
+        {:ok, source}
+
+      err ->
+        err
+    end
+  end
+
+  # If the source is new (ie: not persisted), do nothing
+  defp maybe_handle_media_tasks(%{data: %{__meta__: %{state: state}}}, _source) when state != :loaded do
+    :ok
+  end
+
+  # If the source is NOT new (ie: updated),
+  # enqueue or dequeue media download tasks as necessary.
+  defp maybe_handle_media_tasks(changeset, source) do
+    current_changes = changeset.changes
+    applied_changes = Ecto.Changeset.apply_changes(changeset)
+
+    # We need both current_changes and applied_changes to determine
+    # the course of action to take. For example, we only care if a source is supposed
+    # to be `enabled` or not - we don't care if that information comes from the
+    # current changes or if that's how it already was in the database.
+    # Rephrased, we're essentially using it in place of `get_field/2`
+    case {current_changes, applied_changes} do
+      {%{download_media: true}, %{enabled: true}} ->
+        DownloadingHelpers.enqueue_pending_download_tasks(source)
+
+      {%{enabled: true}, %{download_media: true}} ->
+        DownloadingHelpers.enqueue_pending_download_tasks(source)
+
+      {%{download_media: false}, _} ->
+        DownloadingHelpers.dequeue_pending_download_tasks(source)
+
+      {%{enabled: false}, _} ->
+        DownloadingHelpers.dequeue_pending_download_tasks(source)
+
+      _ ->
+        nil
+    end
+
+    :ok
+  end
+
+  defp maybe_run_indexing_task(changeset, source) do
+    case changeset.data do
+      # If the changeset is new (not persisted), attempt indexing no matter what
+      %{__meta__: %{state: :built}} ->
+        SlowIndexingHelpers.kickoff_indexing_task(source)
+
+        if Ecto.Changeset.get_field(changeset, :fast_index) do
+          FastIndexingHelpers.kickoff_indexing_task(source)
+        end
+
+      # If the record has been persisted, only run indexing if the
+      # indexing frequency has been changed and is now greater than 0
+      %{__meta__: %{state: :loaded}} ->
+        maybe_update_slow_indexing_task(changeset, source)
+        maybe_update_fast_indexing_task(changeset, source)
+    end
+  end
+
+  defp maybe_run_metadata_storage_task(changeset, source) do
+    case {changeset.data, changeset.changes} do
+      # If the changeset is new (not persisted), fetch metadata no matter what
+      {%{__meta__: %{state: :built}}, _} ->
+        SourceMetadataStorageWorker.kickoff_with_task(source)
+
+      # If the record has been persisted, only fetch metadata if the
+      # original_url has changed
+      {_, %{original_url: _}} ->
+        SourceMetadataStorageWorker.kickoff_with_task(source)
+
+      _ ->
+        :ok
+    end
+  end
+
+  defp maybe_update_slow_indexing_task(changeset, source) do
+    # See comment in `maybe_handle_media_tasks` as to why we need these
+    current_changes = changeset.changes
+    applied_changes = Ecto.Changeset.apply_changes(changeset)
+
+    case {current_changes, applied_changes} do
+      {%{index_frequency_minutes: mins}, %{enabled: true}} when mins > 0 ->
+        SlowIndexingHelpers.kickoff_indexing_task(source)
+
+      {%{enabled: true}, %{index_frequency_minutes: mins}} when mins > 0 ->
+        SlowIndexingHelpers.kickoff_indexing_task(source)
+
+      {%{index_frequency_minutes: _}, _} ->
+        SlowIndexingHelpers.delete_indexing_tasks(source, include_executing: true)
+
+      {%{enabled: false}, _} ->
+        SlowIndexingHelpers.delete_indexing_tasks(source, include_executing: true)
+
+      _ ->
+        :ok
+    end
+  end
+
+  defp maybe_update_fast_indexing_task(changeset, source) do
+    # See comment in `maybe_handle_media_tasks` as to why we need these
+    current_changes = changeset.changes
+    applied_changes = Ecto.Changeset.apply_changes(changeset)
+
+    # This technically could be simplified since `maybe_update_slow_indexing_task`
+    # has some overlap re: deleting pending tasks, but I'm keeping it separate
+    # for clarity and explicitness.
+    case {current_changes, applied_changes} do
+      {%{fast_index: true}, %{enabled: true}} ->
+        FastIndexingHelpers.kickoff_indexing_task(source)
+
+      {%{enabled: true}, %{fast_index: true}} ->
+        FastIndexingHelpers.kickoff_indexing_task(source)
+
+      {%{fast_index: false}, _} ->
+        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker", include_executing: true)
+
+      {%{enabled: false}, _} ->
+        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker", include_executing: true)
+
+      _ ->
+        :ok
+    end
+  end
+end

lib/pinchflat/sources/sources_query.ex

@@ -0,0 +1,37 @@
+defmodule Pinchflat.Sources.SourcesQuery do
+  @moduledoc """
+  Query helpers for the Sources context.
+
+  These methods are made to be one-ish liners used
+  to compose queries. Each method should strive to do
+  _one_ thing. These don't need to be tested as
+  they are just building blocks for other functionality
+  which, itself, will be tested.
+  """
+  import Ecto.Query, warn: false
+
+  alias Pinchflat.Sources.Source
+
+  # This allows the module to be aliased and query methods to be used
+  # all in one go
+  # usage: use Pinchflat.Sources.SourcesQuery
+  defmacro __using__(_opts) do
+    quote do
+      import Ecto.Query, warn: false
+
+      alias unquote(__MODULE__)
+    end
+  end
+
+  def new do
+    Source
+  end
+
+  def for_media_profile(media_profile_id) when is_integer(media_profile_id) do
+    dynamic([s], s.media_profile_id == ^media_profile_id)
+  end
+
+  def for_media_profile(media_profile) do
+    dynamic([s], s.media_profile_id == ^media_profile.id)
+  end
+end

lib/pinchflat/tasks/media_item_tasks.ex

@@ -1,61 +0,0 @@
-defmodule Pinchflat.Tasks.MediaItemTasks do
-  @moduledoc """
-  Contains methods used by OR used to create/manage tasks for media items.
-
-  Tasks/workers are meant to be thin wrappers so most of the actual work they
-  do is also defined here. Essentially, a one-stop-shop for media-related tasks/workers.
-  """
-  alias Pinchflat.Media
-  alias Pinchflat.Tasks
-  alias Pinchflat.Sources.Source
-  alias Pinchflat.Workers.MediaDownloadWorker
-
-  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
-
-  @doc """
-  Fetches the file size of a media item and saves it to the database.
-
-  Returns {:ok, media_item} | {:error, any()}
-  """
-  def compute_and_save_media_filesize(media_item) do
-    case File.stat(media_item.media_filepath) do
-      {:ok, %{size: size}} ->
-        Media.update_media_item(media_item, %{media_size_bytes: size})
-
-      err ->
-        err
-    end
-  end
-
-  @doc """
-  Indexes a single media item for a source and enqueues a download job if the
-  media should be downloaded. This method creates the media item record so it's
-  the one-stop-shop for adding a media item (and possibly downloading it) just
-  by a URL and source.
-
-  Returns {:ok, media_item} | {:error, any()}
-  """
-  def index_and_enqueue_download_for_media_item(%Source{} = source, url) do
-    maybe_media_item = create_media_item_from_url(source, url)
-
-    case maybe_media_item do
-      {:ok, media_item} ->
-        if source.download_media && Media.pending_download?(media_item) do
-          %{id: media_item.id}
-          |> MediaDownloadWorker.new()
-          |> Tasks.create_job_with_task(media_item)
-        end
-
-        {:ok, media_item}
-
-      err ->
-        err
-    end
-  end
-
-  defp create_media_item_from_url(source, url) do
-    {:ok, media_attrs} = YtDlpMedia.get_media_attributes(url)
-
-    Media.create_media_item_from_backend_attrs(source, media_attrs)
-  end
-end

lib/pinchflat/tasks/source_tasks.ex

@@ -1,217 +0,0 @@
-defmodule Pinchflat.Tasks.SourceTasks do
-  @moduledoc """
-  Contains methods used by OR used to create/manage tasks for sources.
-
-  Tasks/workers are meant to be thin wrappers so most of the actual work they
-  do is also defined here. Essentially, a one-stop-shop for source-related tasks/workers.
-  """
-
-  require Logger
-
-  alias Pinchflat.Media
-  alias Pinchflat.Tasks
-  alias Pinchflat.Sources
-  alias Pinchflat.Sources.Source
-  alias Pinchflat.Api.YoutubeRss
-  alias Pinchflat.Media.MediaItem
-  alias Pinchflat.Workers.FastIndexingWorker
-  alias Pinchflat.Workers.MediaDownloadWorker
-  alias Pinchflat.Workers.MediaIndexingWorker
-  alias Pinchflat.YtDlp.Backend.MediaCollection
-  alias Pinchflat.Workers.MediaCollectionIndexingWorker
-  alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer
-
-  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
-
-  @doc """
-  Starts tasks for indexing a source's media regardless of the source's indexing
-  frequency. It's assumed the caller will check for indexing frequency.
-
-  Returns {:ok, %Task{}}.
-  """
-  def kickoff_indexing_task(%Source{} = source) do
-    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
-    Tasks.delete_pending_tasks_for(source, "MediaIndexingWorker")
-    Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker")
-
-    %{id: source.id}
-    # Schedule this one immediately, but future ones will be on an interval
-    |> MediaCollectionIndexingWorker.new()
-    |> Tasks.create_job_with_task(source)
-  end
-
-  @doc """
-  Starts tasks for running a fast indexing task for a source's media
-  regardless of the source's fast_index state. It's assumed the
-  caller will check for fast_index.
-
-  This is used for running fast index tasks on update. On creation, the
-  fast index is enqueued after the slow index is complete.
-
-  Returns {:ok, %Task{}}.
-  """
-  def kickoff_fast_indexing_task(%Source{} = source) do
-    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
-
-    %{id: source.id}
-    # Schedule this one immediately, but future ones will be on an interval
-    |> FastIndexingWorker.new()
-    |> Tasks.create_job_with_task(source)
-  end
-
-  @doc """
-  Fetches new media IDs from a source's YouTube RSS feed and kicks off indexing tasks
-  for any new media items. See comments in `MediaIndexingWorker` for more info on the
-  order of operations and how this fits into the indexing process.
-
-  Returns :ok
-  """
-  def kickoff_indexing_tasks_from_youtube_rss_feed(%Source{} = source) do
-    {:ok, media_ids} = YoutubeRss.get_recent_media_ids_from_rss(source)
-    existing_media_items = Media.list_media_items_by_media_id_for(source, media_ids)
-    new_media_ids = media_ids -- Enum.map(existing_media_items, & &1.media_id)
-
-    Enum.each(new_media_ids, fn media_id ->
-      url = "https://www.youtube.com/watch?v=#{media_id}"
-
-      %{id: source.id, media_url: url}
-      |> MediaIndexingWorker.new()
-      |> Tasks.create_job_with_task(source)
-    end)
-  end
-
-  @doc """
-  Given a media source, creates (indexes) the media by creating media_items for each
-  media ID in the source. Afterward, kicks off a download task for each pending media
-  item belonging to the source. You can't tell me the method name isn't descriptive!
-
-  Indexing is slow and usually returns a list of all media data at once for record creation.
-  To help with this, we use a file follower to watch the file that yt-dlp writes to
-  so we can create media items as they come in. This parallelizes the process and adds
-  clarity to the user experience. This has a few things to be aware of which are documented
-  below in the file watcher setup method.
-
-  NOTE: downloads are only enqueued if the source is set to download media. Downloads are
-  also enqueued for ALL pending media items, not just the ones that were indexed in this
-  job run. This should ensure that any stragglers are caught if, for some reason, they
-  weren't enqueued or somehow got de-queued.
-
-  Since indexing returns all media data EVERY TIME, we that that opportunity to update
-  indexing metadata for media items that have already been created.
-
-  Returns [%MediaItem{}, ...]
-  """
-  def index_and_enqueue_download_for_media_items(%Source{} = source) do
-    # See the method definition below for more info on how file watchers work
-    # (important reading if you're not familiar with it)
-    {:ok, media_attributes} = get_media_attributes_for_collection_and_setup_file_watcher(source)
-    result = Enum.map(media_attributes, fn media_attrs -> create_media_item_from_attributes(source, media_attrs) end)
-
-    Sources.update_source(source, %{last_indexed_at: DateTime.utc_now()})
-    enqueue_pending_media_tasks(source)
-
-    result
-  end
-
-  @doc """
-  Starts tasks for downloading media for any of a sources _pending_ media items.
-  Jobs are not enqueued if the source is set to not download media. This will return :ok.
-
-  NOTE: this starts a download for each media item that is pending,
-  not just the ones that were indexed in this job run. This should ensure
-  that any stragglers are caught if, for some reason, they weren't enqueued
-  or somehow got de-queued.
-
-  Returns :ok
-  """
-  def enqueue_pending_media_tasks(%Source{download_media: true} = source) do
-    source
-    |> Media.list_pending_media_items_for()
-    |> Enum.each(fn media_item ->
-      %{id: media_item.id}
-      |> MediaDownloadWorker.new()
-      |> Tasks.create_job_with_task(media_item)
-    end)
-  end
-
-  def enqueue_pending_media_tasks(%Source{download_media: false} = _source) do
-    :ok
-  end
-
-  @doc """
-  Deletes ALL pending tasks for a source's media items.
-
-  Returns :ok
-  """
-  def dequeue_pending_media_tasks(%Source{} = source) do
-    source
-    |> Media.list_pending_media_items_for()
-    |> Enum.each(&Tasks.delete_pending_tasks_for/1)
-  end
-
-  # The file follower is a GenServer that watches a file for new lines and
-  # processes them. This works well, but we have to be resilliant to partially-written
-  # lines (ie: you should gracefully fail if you can't parse a line).
-  #
-  # This works in-tandem with the normal (blocking) media indexing behaviour. When
-  # the `get_media_attributes_for_collection` method completes it'll return the FULL result to
-  # the caller for parsing. Ideally, every item in the list will have already
-  # been processed by the file follower, but if not, the caller handles creation
-  # of any media items that were missed/initially failed.
-  #
-  # It attempts a graceful shutdown of the file follower after the indexing is done,
-  # but the FileFollowerServer will also stop itself if it doesn't see any activity
-  # for a sufficiently long time.
-  defp get_media_attributes_for_collection_and_setup_file_watcher(source) do
-    {:ok, pid} = FileFollowerServer.start_link()
-
-    handler = fn filepath -> setup_file_follower_watcher(pid, filepath, source) end
-    result = MediaCollection.get_media_attributes_for_collection(source.original_url, file_listener_handler: handler)
-
-    FileFollowerServer.stop(pid)
-
-    result
-  end
-
-  defp setup_file_follower_watcher(pid, filepath, source) do
-    FileFollowerServer.watch_file(pid, filepath, fn line ->
-      case Phoenix.json_library().decode(line) do
-        {:ok, media_attrs} ->
-          Logger.debug("FileFollowerServer Handler: Got media attributes: #{inspect(media_attrs)}")
-
-          media_struct = YtDlpMedia.response_to_struct(media_attrs)
-          create_media_item_and_enqueue_download(source, media_struct)
-
-        err ->
-          Logger.debug("FileFollowerServer Handler: Error decoding JSON: #{inspect(err)}")
-
-          err
-      end
-    end)
-  end
-
-  defp create_media_item_and_enqueue_download(source, media_attrs) do
-    maybe_media_item = create_media_item_from_attributes(source, media_attrs)
-
-    case maybe_media_item do
-      %MediaItem{} = media_item ->
-        if source.download_media && Media.pending_download?(media_item) do
-          Logger.debug("FileFollowerServer Handler: Enqueuing download task for #{inspect(media_attrs)}")
-
-          %{id: media_item.id}
-          |> MediaDownloadWorker.new()
-          |> Tasks.create_job_with_task(media_item)
-        end
-
-      changeset ->
-        changeset
-    end
-  end
-
-  defp create_media_item_from_attributes(source, media_attrs) do
-    case Media.create_media_item_from_backend_attrs(source, media_attrs) do
-      {:ok, media_item} -> media_item
-      {:error, changeset} -> changeset
-    end
-  end
-end

lib/pinchflat/tasks/tasks.ex

@@ -22,9 +22,15 @@ defmodule Pinchflat.Tasks do
 
   Returns [%Task{}, ...]
   """
-  def list_tasks_for(attached_record_type, attached_record_id, worker_name \\ nil, job_states \\ Oban.Job.states()) do
+  def list_tasks_for(record, worker_name \\ nil, job_states \\ Oban.Job.states()) do
     stringified_states = Enum.map(job_states, &to_string/1)
 
+    record_type =
+      case record do
+        %Source{} -> :source_id
+        %MediaItem{} -> :media_item_id
+      end
+
     worker_name_finder =
       if worker_name do
         # Workers are the full module name - we want to match on the string ENDING with
@@ -41,27 +47,12 @@ defmodule Pinchflat.Tasks do
     Repo.all(
       from t in Task,
         join: j in assoc(t, :job),
-        where: field(t, ^attached_record_type) == ^attached_record_id,
+        where: field(t, ^record_type) == ^record.id,
         where: ^worker_name_finder,
         where: j.state in ^stringified_states
     )
   end
 
-  @doc """
-  Returns the list of pending tasks for a given record type and ID. Optionally allows you to specify
-  which worker to include.
-
-  Returns [%Task{}, ...]
-  """
-  def list_pending_tasks_for(attached_record_type, attached_record_id, worker_name \\ nil) do
-    list_tasks_for(
-      attached_record_type,
-      attached_record_id,
-      worker_name,
-      [:available, :scheduled, :retryable]
-    )
-  end
-
   @doc """
   Gets a single task.
 
@@ -122,18 +113,14 @@ defmodule Pinchflat.Tasks do
 
   @doc """
   Deletes all tasks attached to a given record, cancelling any attached jobs.
-  Optionally allows you to specify which worker to include.
+  Optionally allows you to specify which worker and job states to include.
 
   Returns :ok
   """
-  def delete_tasks_for(attached_record, worker_name \\ nil) do
-    tasks =
-      case attached_record do
-        %Source{} = source -> list_tasks_for(:source_id, source.id, worker_name)
-        %MediaItem{} = media_item -> list_tasks_for(:media_item_id, media_item.id, worker_name)
-      end
-
-    Enum.each(tasks, &delete_task/1)
+  def delete_tasks_for(record, worker_name \\ nil, job_states \\ Oban.Job.states()) do
+    record
+    |> list_tasks_for(worker_name, job_states)
+    |> Enum.each(&delete_task/1)
   end
 
   @doc """
@@ -142,14 +129,12 @@ defmodule Pinchflat.Tasks do
 
   Returns :ok
   """
-  def delete_pending_tasks_for(attached_record, worker_name \\ nil) do
-    tasks =
-      case attached_record do
-        %Source{} = source -> list_pending_tasks_for(:source_id, source.id, worker_name)
-        %MediaItem{} = media_item -> list_pending_tasks_for(:media_item_id, media_item.id, worker_name)
-      end
+  def delete_pending_tasks_for(record, worker_name \\ nil, opts \\ []) do
+    include_executing = Keyword.get(opts, :include_executing, false)
+    base_job_states = [:available, :scheduled, :retryable]
+    job_states = if include_executing, do: base_job_states ++ [:executing], else: base_job_states
 
-    Enum.each(tasks, &delete_task/1)
+    delete_tasks_for(record, worker_name, job_states)
   end
 
   @doc """

lib/pinchflat/tasks/tasks_query.ex

@@ -0,0 +1,45 @@
+defmodule Pinchflat.Tasks.TasksQuery do
+  @moduledoc """
+  Query helpers for the Tasks context.
+
+  These methods are made to be one-ish liners used
+  to compose queries. Each method should strive to do
+  _one_ thing. These don't need to be tested as
+  they are just building blocks for other functionality
+  which, itself, will be tested.
+  """
+  import Ecto.Query, warn: false
+
+  alias Pinchflat.Tasks.Task
+
+  # This allows the module to be aliased and query methods to be used
+  # all in one go
+  # usage: use Pinchflat.Tasks.TasksQuery
+  defmacro __using__(_opts) do
+    quote do
+      import Ecto.Query, warn: false
+
+      alias unquote(__MODULE__)
+    end
+  end
+
+  def new do
+    Task
+  end
+
+  def join_job(query) do
+    query
+    |> join(:left, [t], j in assoc(t, :job))
+    |> preload([t, j], job: j)
+  end
+
+  def in_state(states) when is_list(states) do
+    dynamic([t, j], j.state in ^states)
+  end
+
+  def in_state(state), do: in_state([state])
+
+  def has_tag(tag) do
+    dynamic([t, j], ^tag in j.tags)
+  end
+end

lib/pinchflat/utils/cli_utils.ex

@@ -0,0 +1,94 @@
+defmodule Pinchflat.Utils.CliUtils do
+  @moduledoc """
+  Utility methods for working with CLI executables
+  """
+
+  require Logger
+
+  alias Pinchflat.Utils.StringUtils
+
+  @doc """
+  Wraps a command in a shell script that will terminate
+  the command if stdin is closed. Useful for stopping
+  commands if the job runner is cancelled.
+
+  Delegates to `System.cmd/3` and any options/output
+  are passed through. Custom options can be passed in.
+
+  Custom options:
+    - logging_arg_override: if set, the passed value will be logged in place of
+      the actual arguments passed to the command
+
+  Returns {binary(), integer()}
+  """
+  def wrap_cmd(command, args, passthrough_opts \\ [], opts \\ []) do
+    wrapper_command = Path.join(:code.priv_dir(:pinchflat), "cmd_wrapper.sh")
+    actual_command = [command] ++ args
+    command_opts = set_command_opts() ++ passthrough_opts
+    logging_arg_override = Keyword.get(opts, :logging_arg_override, Enum.join(args, " "))
+
+    Logger.info("[command_wrapper]: #{command} called with: #{logging_arg_override}")
+
+    {output, status} = System.cmd(wrapper_command, actual_command, command_opts)
+    log_cmd_result(command, logging_arg_override, status, output)
+
+    {output, status}
+  end
+
+  @doc """
+  Parses a list of command options into a list of strings suitable for passing to
+  `System.cmd/3`.
+
+  We want to satisfy the following behaviours:
+    1. If the key is an atom, convert it to a string and convert it to kebab case (for convenience)
+    2. If the key is a string, assume we want it as-is and don't convert it
+    3. If the key is accompanied by a value, append the value to the list
+    4. If the key is not accompanied by a value, assume it's a flag and PREpend it to the list
+
+  Returns [binary()]
+  """
+  def parse_options(command_opts) do
+    command_opts
+    |> List.wrap()
+    |> Enum.reduce([], &parse_option/2)
+  end
+
+  defp parse_option({k, v}, acc) when is_atom(k) do
+    stringified_key = StringUtils.to_kebab_case(Atom.to_string(k))
+
+    parse_option({"--#{stringified_key}", v}, acc)
+  end
+
+  defp parse_option({k, v}, acc) when is_binary(k) do
+    acc ++ [k, to_string(v)]
+  end
+
+  defp parse_option(arg, acc) when is_atom(arg) do
+    stringified_arg =
+      arg
+      |> Atom.to_string()
+      |> StringUtils.to_kebab_case()
+
+    parse_option("--#{stringified_arg}", acc)
+  end
+
+  defp parse_option(arg, acc) when is_binary(arg) do
+    acc ++ [arg]
+  end
+
+  defp log_cmd_result(command, logging_arg_override, status, output) do
+    log_message = "[command_wrapper]: #{command} called with: #{logging_arg_override} exited: #{status} with: #{output}"
+    log_level = if status == 0, do: :debug, else: :error
+
+    Logger.log(log_level, log_message)
+  end
+
+  defp set_command_opts do
+    # This resolves an issue where yt-dlp would attempt to write to a read-only directory
+    # if you scanned a new video with `--windows-filenames` enabled. Hopefully can be removed
+    # in the future.
+    [
+      cd: Application.get_env(:pinchflat, :tmpfile_directory)
+    ]
+  end
+end

lib/pinchflat/utils/filesystem_utils.ex

@@ -2,9 +2,42 @@ defmodule Pinchflat.Utils.FilesystemUtils do
   @moduledoc """
   Utility methods for working with the filesystem
   """
-
+  alias Pinchflat.Media
   alias Pinchflat.Utils.StringUtils
 
+  @doc """
+  Checks if a file exists and has non-whitespace contents.
+
+  Returns boolean()
+  """
+  def exists_and_nonempty?(filepath) do
+    case File.read(filepath) do
+      {:ok, contents} ->
+        String.trim(contents) != ""
+
+      _ ->
+        false
+    end
+  end
+
+  @doc """
+  Checks if two filepaths reference the same file.
+
+  Useful if you have a relative and absolute filepath and want to be sure they're the same file.
+  Also works with symlinks.
+
+  Returns boolean()
+  """
+  def filepaths_reference_same_file?(filepath_1, filepath_2) do
+    {:ok, stat_1} = File.stat(filepath_1)
+    {:ok, stat_2} = File.stat(filepath_2)
+
+    identifier_1 = "#{stat_1.major_device}:#{stat_1.minor_device}:#{stat_1.inode}"
+    identifier_2 = "#{stat_2.major_device}:#{stat_2.minor_device}:#{stat_2.inode}"
+
+    identifier_1 == identifier_2
+  end
+
   @doc """
   Generates a temporary file and returns its path. The file is empty and has the given type.
   Generates all the directories in the path if they don't exist.
@@ -12,12 +45,108 @@ defmodule Pinchflat.Utils.FilesystemUtils do
   Returns binary()
   """
   def generate_metadata_tmpfile(type) do
+    filename = StringUtils.random_string(64)
+    # This "namespacing" is more to help with development since things get
+    # weird in my editor when there are thousands of files in a single directory
+    first_two = String.slice(filename, 0..1)
+    second_two = String.slice(filename, 2..3)
     tmpfile_directory = Application.get_env(:pinchflat, :tmpfile_directory)
-    filepath = Path.join([tmpfile_directory, "#{StringUtils.random_string(64)}.#{type}"])
 
-    :ok = File.mkdir_p!(Path.dirname(filepath))
-    :ok = File.write(filepath, "")
+    filepath =
+      Path.join([
+        tmpfile_directory,
+        first_two,
+        second_two,
+        "#{filename}.#{type}"
+      ])
+
+    :ok = write_p!(filepath, "")
 
     filepath
   end
+
+  @doc """
+  Writes content to a file, creating directories as needed.
+  Takes the same args as File.write/3.
+
+  Returns :ok | {:error, any()}
+  """
+  def write_p(file, content, modes \\ []) do
+    dirname = Path.dirname(file)
+
+    case File.mkdir_p(dirname) do
+      :ok -> File.write(file, content, modes)
+      err -> err
+    end
+  end
+
+  @doc """
+  Writes content to a file, creating directories as needed.
+  Takes the same args as File.write!/3.
+
+  Returns :ok | raises on error
+  """
+  def write_p!(filepath, content, modes \\ []) do
+    :ok = write_p(filepath, content, modes)
+  end
+
+  @doc """
+  Copies a file from source to destination, creating directories as needed.
+
+  Returns :ok | raises on error
+  """
+  def cp_p!(source, destination) do
+    destination
+    |> Path.dirname()
+    |> File.mkdir_p!()
+
+    File.cp!(source, destination)
+  end
+
+  @doc """
+  Fetches the file size of a media item and saves it to the database.
+
+  Returns {:ok, media_item} | {:error, any()}
+  """
+  def compute_and_save_media_filesize(media_item) do
+    case File.stat(media_item.media_filepath) do
+      {:ok, %{size: size}} ->
+        Media.update_media_item(media_item, %{media_size_bytes: size})
+
+      err ->
+        err
+    end
+  end
+
+  @doc """
+  Deletes a file and removes any empty directories in the path.
+  Does NOT remove any directories that are not empty.
+
+  Returns :ok | {:error, any()}
+  """
+  def delete_file_and_remove_empty_directories(filepath) do
+    case File.rm(filepath) do
+      :ok ->
+        filepath
+        |> Path.dirname()
+        |> recursively_delete_empty_directories()
+
+      err ->
+        err
+    end
+  end
+
+  defp recursively_delete_empty_directories(directory) do
+    case File.rmdir(directory) do
+      :ok ->
+        directory
+        |> Path.dirname()
+        |> recursively_delete_empty_directories()
+
+      err ->
+        err
+    end
+
+    :ok
+  end
 end

lib/pinchflat/utils/map_utils.ex

@@ -0,0 +1,17 @@
+defmodule Pinchflat.Utils.MapUtils do
+  @moduledoc """
+  Utility methods for working with maps
+  """
+
+  @doc """
+  Converts a nested list of 2-element tuples or lists into a map.
+
+  Returns map()
+  """
+  def from_nested_list(list) do
+    Enum.reduce(list, %{}, fn
+      [key, value], acc -> Map.put(acc, key, value)
+      {key, value}, acc -> Map.put(acc, key, value)
+    end)
+  end
+end

lib/pinchflat/utils/number_utils.ex

@@ -0,0 +1,53 @@
+defmodule Pinchflat.Utils.NumberUtils do
+  @moduledoc """
+  Utility methods for working with numbers
+  """
+
+  @doc """
+  Clamps a number between a minimum and maximum value
+
+  Returns integer() | float()
+  """
+  def clamp(num, minimum, maximum) do
+    num
+    |> max(minimum)
+    |> min(maximum)
+  end
+
+  @doc """
+  Converts a number to a human readable byte size. Can take a precision
+  option to specify the number of decimal places to round to.
+
+  Returns {integer(), String.t()}
+  """
+  def human_byte_size(number, opts \\ [])
+  def human_byte_size(nil, opts), do: human_byte_size(0, opts)
+
+  def human_byte_size(number, opts) do
+    precision = Keyword.get(opts, :precision, 2)
+    suffixes = ["B", "KB", "MB", "GB", "TB", "PB", "EB", "ZB", "YB"]
+    base = 1024
+
+    Enum.reduce_while(suffixes, {number / 1.0, "B"}, fn suffix, {value, _} ->
+      if value < base do
+        {:halt, {Float.round(value, precision), suffix}}
+      else
+        {:cont, {value / base, suffix}}
+      end
+    end)
+  end
+
+  @doc """
+  Adds jitter to a number based on a percentage. Returns 0 if the number is less than or equal to 0.
+
+  Returns integer()
+  """
+  def add_jitter(num, jitter_percentage \\ 0.5)
+  def add_jitter(num, _jitter_percentage) when num <= 0, do: 0
+
+  def add_jitter(num, jitter_percentage) do
+    jitter = :rand.uniform(round(num * jitter_percentage))
+
+    round(num + jitter)
+  end
+end

lib/pinchflat/utils/string_utils.ex

@@ -26,19 +26,22 @@ defmodule Pinchflat.Utils.StringUtils do
   end
 
   @doc """
-  Truncates a string to the given length and adds `...` if the string is longer than the given length.
-  Will break on a word boundary. Nothing happens if the string is shorter than the given length.
+  Wraps a string in double braces. Useful as a UI helper now that
+  LiveView 1.0.0 allows `{}` for interpolation so now we can't use braces
+  directly in the view.
 
   Returns binary()
   """
-  def truncate(string, length) do
-    if String.length(string) > length do
-      string
-      |> String.slice(0..(length - 1))
-      |> String.replace(~r/\s+\S*$/, "")
-      |> Kernel.<>("...")
-    else
-      string
-    end
+  def double_brace(string) do
+    "{{ #{string} }}"
   end
+
+  @doc """
+  Wraps a string in quotes if it's not already a string. Useful for working with
+  error messages whose types can vary.
+
+  Returns binary()
+  """
+  def wrap_string(message) when is_binary(message), do: message
+  def wrap_string(message), do: "#{inspect(message)}"
 end

lib/pinchflat/utils/xml_utils.ex

@@ -0,0 +1,17 @@
+defmodule Pinchflat.Utils.XmlUtils do
+  @moduledoc """
+  Utility methods for working with XML documents
+  """
+
+  @doc """
+  Escapes invalid XML characters in a string
+
+  Returns binary()
+  """
+  def safe(value) do
+    value
+    |> to_string()
+    |> Phoenix.HTML.html_escape()
+    |> Phoenix.HTML.safe_to_string()
+  end
+end