[WIP] break out a few contexts, start refactoring fast index modules

This commit is contained in:
Kieran Eglin 2024-03-12 09:04:16 -07:00
parent c0885cfaf2
commit f549b43793
No known key found for this signature in database
GPG key ID: 193984967FCF432D
20 changed files with 160 additions and 119 deletions

View file

@ -5,7 +5,7 @@ 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
@ -18,7 +18,7 @@ alias Pinchflat.MediaClient.MediaDownloader
alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
alias Pinchflat.YtDlp.Backend.MediaCollection, as: YtDlpCollection
alias Pinchflat.Api.YoutubeRss
alias Pinchflat.FastIndexing.YoutubeRss
alias Pinchflat.Metadata.MetadataFileHelpers
alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer

View file

@ -1,4 +1,4 @@
defmodule Pinchflat.YtDlp.DownloadOptionBuilder do
defmodule Pinchflat.Downloading.DownloadOptionBuilder do
@moduledoc """
Builds the options for yt-dlp to download media based on the given media profile.
"""

View file

@ -0,0 +1,59 @@
defmodule Pinchflat.FastIndexing.FastIndexingHelpers do
alias Pinchflat.Media
alias Pinchflat.Tasks
alias Pinchflat.Sources
alias Pinchflat.Sources.Source
alias Pinchflat.FastIndexing.YoutubeRss
alias Pinchflat.Media.MediaItem
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaIndexingWorker
alias Pinchflat.YtDlp.Backend.MediaCollection
alias Pinchflat.Workers.MediaCollectionIndexingWorker
alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer
@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.
Despite the similar name to `kickoff_fast_indexing_task`, this does work differently.
`kickoff_fast_indexing_task` starts a task that _calls_ this function whereas this
function starts individual indexing tasks for each new media item. I think it does
make sense grammatically, but I could see how that's confusing.
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
end

View file

@ -1,4 +1,4 @@
defmodule Pinchflat.Workers.FastIndexingWorker do
defmodule Pinchflat.FastIndexing.FastIndexingWorker do
@moduledoc false
use Oban.Worker,
@ -10,7 +10,7 @@ defmodule Pinchflat.Workers.FastIndexingWorker do
alias Pinchflat.Tasks
alias Pinchflat.Sources
alias Pinchflat.Sources.Source
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.FastIndexing.FastIndexingHelpers
@impl Oban.Worker
@doc """
@ -24,7 +24,7 @@ defmodule Pinchflat.Workers.FastIndexingWorker do
source = Sources.get_source!(source_id)
if source.fast_index do
SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
FastIndexingHelpers.kickoff_indexing_tasks_from_youtube_rss_feed(source)
reschedule_indexing(source)
else

View file

@ -1,4 +1,4 @@
defmodule Pinchflat.Api.YoutubeRss do
defmodule Pinchflat.FastIndexing.YoutubeRss do
@moduledoc """
Methods for interacting with YouTube RSS feeds
"""

View file

@ -9,7 +9,7 @@ defmodule Pinchflat.Media do
alias Pinchflat.Tasks
alias Pinchflat.Media.MediaItem
alias Pinchflat.Sources.Source
alias Pinchflat.Media.MediaMetadata
alias Pinchflat.Metadata.MediaMetadata
@doc """
Returns the list of media_items.

View file

@ -8,7 +8,7 @@ defmodule Pinchflat.Media.MediaItem do
alias Pinchflat.Tasks.Task
alias Pinchflat.Sources.Source
alias Pinchflat.Media.MediaMetadata
alias Pinchflat.Metadata.MediaMetadata
alias Pinchflat.Media.MediaItemSearchIndex
@allowed_fields [

View file

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

View file

@ -12,6 +12,7 @@ defmodule Pinchflat.Sources do
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.Profiles.MediaProfile
alias Pinchflat.YtDlp.Backend.MediaCollection
alias Pinchflat.FastIndexing.FastIndexingHelpers
@doc """
Returns the list of sources. Returns [%Source{}, ...]
@ -251,7 +252,7 @@ defmodule Pinchflat.Sources do
defp maybe_update_fast_indexing_task(changeset, source) do
case changeset.changes do
%{fast_index: true} ->
SourceTasks.kickoff_fast_indexing_task(source)
FastIndexingHelpers.kickoff_fast_indexing_task(source)
%{fast_index: false} ->
Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")

View file

@ -12,9 +12,9 @@ defmodule Pinchflat.Tasks.SourceTasks do
alias Pinchflat.Tasks
alias Pinchflat.Sources
alias Pinchflat.Sources.Source
alias Pinchflat.Api.YoutubeRss
alias Pinchflat.FastIndexing.YoutubeRss
alias Pinchflat.Media.MediaItem
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaIndexingWorker
alias Pinchflat.YtDlp.Backend.MediaCollection
@ -40,46 +40,6 @@ defmodule Pinchflat.Tasks.SourceTasks do
|> 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
@ -134,7 +94,7 @@ defmodule Pinchflat.Tasks.SourceTasks do
end)
end
def enqueue_pending_media_tasks(%Source{download_media: false} = _source) do
def enqueue_pending_media_tasks(%Source{download_media: false}) do
:ok
end

View file

@ -11,7 +11,7 @@ defmodule Pinchflat.Workers.MediaCollectionIndexingWorker do
alias Pinchflat.Sources
alias Pinchflat.Sources.Source
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
@impl Oban.Worker
@doc """

View file

@ -23,7 +23,7 @@ defmodule Pinchflat.Workers.MediaIndexingWorker do
and the media matches the profile's format preferences)
Order of operations:
1. SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed/1 (which is running
1. FastIndexingHelpers.kickoff_indexing_tasks_from_youtube_rss_feed/1 (which is running
in its own worker) periodically checks the YouTube RSS feed for new media
2. If new media is found, it enqueues a MediaIndexingWorker (this module) for each new media
item

View file

@ -10,7 +10,7 @@ defmodule Pinchflat.MediaClient.MediaDownloader do
alias Pinchflat.Media.MediaItem
alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
alias Pinchflat.YtDlp.DownloadOptionBuilder, as: YtDlpDownloadOptionBuilder
alias Pinchflat.Downloading.DownloadOptionBuilder, as: YtDlpDownloadOptionBuilder
alias Pinchflat.Metadata.MetadataParser, as: YtDlpMetadataParser
alias Pinchflat.Metadata.MetadataFileHelpers, as: YtDlpMetadataHelpers

View file

@ -1,11 +1,11 @@
defmodule Pinchflat.YtDlp.DownloadOptionBuilderTest do
defmodule Pinchflat.Downloading.DownloadOptionBuilderTest do
use Pinchflat.DataCase
import Pinchflat.MediaFixtures
import Pinchflat.ProfilesFixtures
import Pinchflat.SourcesFixtures
alias Pinchflat.Profiles
alias Pinchflat.YtDlp.DownloadOptionBuilder
alias Pinchflat.Downloading.DownloadOptionBuilder
setup do
media_profile = media_profile_fixture(%{output_path_template: "{{ title }}.%(ext)s"})

View file

@ -0,0 +1,74 @@
defmodule Pinchflat.FastIndexing.FastIndexingHelpersTest do
use Pinchflat.DataCase
import Mox
import Pinchflat.TasksFixtures
import Pinchflat.MediaFixtures
import Pinchflat.SourcesFixtures
import Pinchflat.ProfilesFixtures
alias Pinchflat.Tasks
alias Pinchflat.Tasks.Task
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.Media.MediaItem
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingHelpers
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaCollectionIndexingWorker
setup :verify_on_exit!
describe "kickoff_fast_indexing_task/1" do
test "it schedules a job" do
source = source_fixture()
assert {:ok, _} = FastIndexingHelpers.kickoff_fast_indexing_task(source)
assert_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
end
test "it creates and attaches a task" do
source = source_fixture()
assert {:ok, %Task{} = task} = FastIndexingHelpers.kickoff_fast_indexing_task(source)
assert task.source_id == source.id
end
test "it deletes any fast indexing tasks for the source" do
source = source_fixture()
{:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
task = task_fixture(source_id: source.id, job_id: job.id)
assert {:ok, _} = FastIndexingHelpers.kickoff_fast_indexing_task(source)
assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
end
end
describe "kickoff_indexing_tasks_from_youtube_rss_feed/1" do
setup do
{:ok, [source: source_fixture()]}
end
test "enqueues a new worker for each new media_id in the source's RSS feed", %{source: source} do
expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
assert :ok = FastIndexingHelpers.kickoff_indexing_tasks_from_youtube_rss_feed(source)
assert [worker] = all_enqueued(worker: MediaIndexingWorker)
assert worker.args["id"] == source.id
assert worker.args["media_url"] == "https://www.youtube.com/watch?v=test_1"
end
test "does not enqueue a new worker for the source's media IDs we already know about", %{source: source} do
expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
media_item_fixture(source_id: source.id, media_id: "test_1")
assert :ok = FastIndexingHelpers.kickoff_indexing_tasks_from_youtube_rss_feed(source)
refute_enqueued(worker: MediaIndexingWorker)
end
end
end

View file

@ -1,11 +1,11 @@
defmodule Pinchflat.Workers.FastIndexingWorkerTest do
defmodule Pinchflat.FastIndexing.FastIndexingWorkerTest do
use Pinchflat.DataCase
import Mox
import Pinchflat.SourcesFixtures
alias Pinchflat.Sources.Source
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
setup :verify_on_exit!

View file

@ -1,9 +1,9 @@
defmodule Pinchflat.Api.YoutubeRssTest do
defmodule Pinchflat.FastIndexing.YoutubeRssTest do
use Pinchflat.DataCase
import Mox
import Pinchflat.SourcesFixtures
alias Pinchflat.Api.YoutubeRss
alias Pinchflat.FastIndexing.YoutubeRss
setup :verify_on_exit!

View file

@ -9,7 +9,7 @@ defmodule Pinchflat.SourcesTest do
alias Pinchflat.Sources
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.Sources.Source
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaIndexingWorker
alias Pinchflat.Workers.MediaCollectionIndexingWorker

View file

@ -11,7 +11,7 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
alias Pinchflat.Tasks.Task
alias Pinchflat.Tasks.SourceTasks
alias Pinchflat.Media.MediaItem
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaIndexingWorker
alias Pinchflat.Workers.MediaCollectionIndexingWorker
@ -66,59 +66,6 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
end
end
describe "kickoff_fast_indexing_task/1" do
test "it schedules a job" do
source = source_fixture()
assert {:ok, _} = SourceTasks.kickoff_fast_indexing_task(source)
assert_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
end
test "it creates and attaches a task" do
source = source_fixture()
assert {:ok, %Task{} = task} = SourceTasks.kickoff_fast_indexing_task(source)
assert task.source_id == source.id
end
test "it deletes any fast indexing tasks for the source" do
source = source_fixture()
{:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
task = task_fixture(source_id: source.id, job_id: job.id)
assert {:ok, _} = SourceTasks.kickoff_fast_indexing_task(source)
assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
end
end
describe "kickoff_indexing_tasks_from_youtube_rss_feed/1" do
setup do
{:ok, [source: source_fixture()]}
end
test "enqueues a new worker for each new media_id in the source's RSS feed", %{source: source} do
expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
assert :ok = SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
assert [worker] = all_enqueued(worker: MediaIndexingWorker)
assert worker.args["id"] == source.id
assert worker.args["media_url"] == "https://www.youtube.com/watch?v=test_1"
end
test "does not enqueue a new worker for the source's media IDs we already know about", %{source: source} do
expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
media_item_fixture(source_id: source.id, media_id: "test_1")
assert :ok = SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
refute_enqueued(worker: MediaIndexingWorker)
end
end
describe "index_and_enqueue_download_for_media_items/1" do
setup do
stub(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->

View file

@ -8,7 +8,7 @@ defmodule Pinchflat.Workers.MediaCollectionIndexingWorkerTest do
alias Pinchflat.Tasks
alias Pinchflat.Sources.Source
alias Pinchflat.Workers.FastIndexingWorker
alias Pinchflat.FastIndexing.FastIndexingWorker
alias Pinchflat.Workers.MediaDownloadWorker
alias Pinchflat.Workers.MediaCollectionIndexingWorker