ytdl-sub/src/ytdl_sub/downloaders/downloader.py

598 lines
22 KiB
Python

import abc
import contextlib
import os
from abc import ABC
from pathlib import Path
from typing import Dict
from typing import Generic
from typing import Iterable
from typing import Iterator
from typing import List
from typing import Optional
from typing import Set
from typing import Tuple
from typing import Type
from typing import TypeVar
from ytdl_sub.config.preset_options import AddsVariablesMixin
from ytdl_sub.config.preset_options import Overrides
from ytdl_sub.downloaders.generic.validators import MultiUrlValidator
from ytdl_sub.downloaders.generic.validators import UrlThumbnailListValidator
from ytdl_sub.downloaders.generic.validators import UrlValidator
from ytdl_sub.downloaders.ytdl_options_builder import YTDLOptionsBuilder
from ytdl_sub.downloaders.ytdlp import YTDLP
from ytdl_sub.entries.entry import Entry
from ytdl_sub.entries.entry_parent import EntryParent
from ytdl_sub.entries.variables.kwargs import COLLECTION_URL
from ytdl_sub.entries.variables.kwargs import COMMENTS
from ytdl_sub.entries.variables.kwargs import DOWNLOAD_INDEX
from ytdl_sub.entries.variables.kwargs import PLAYLIST_ENTRY
from ytdl_sub.entries.variables.kwargs import REQUESTED_SUBTITLES
from ytdl_sub.entries.variables.kwargs import SOURCE_ENTRY
from ytdl_sub.entries.variables.kwargs import SPONSORBLOCK_CHAPTERS
from ytdl_sub.entries.variables.kwargs import UPLOAD_DATE_INDEX
from ytdl_sub.plugins.plugin import Plugin
from ytdl_sub.plugins.plugin import PluginOptions
from ytdl_sub.utils.file_handler import FileHandler
from ytdl_sub.utils.logger import Logger
from ytdl_sub.utils.thumbnail import ThumbnailTypes
from ytdl_sub.utils.thumbnail import convert_download_thumbnail
from ytdl_sub.utils.thumbnail import download_and_convert_url_thumbnail
from ytdl_sub.validators.strict_dict_validator import StrictDictValidator
from ytdl_sub.ytdl_additions.enhanced_download_archive import DownloadArchiver
from ytdl_sub.ytdl_additions.enhanced_download_archive import EnhancedDownloadArchive
# pylint: disable=too-many-instance-attributes
download_logger = Logger.get(name="downloader")
class DownloaderValidator(StrictDictValidator, AddsVariablesMixin, ABC):
"""
Placeholder class to define downloader options
"""
@property
@abc.abstractmethod
def collection_validator(self) -> MultiUrlValidator:
"""
Returns
-------
MultiUrlValidator
To determine how the entries are downloaded
"""
def added_source_variables(self) -> List[str]:
"""
Returns
-------
Added source variables on the collection
"""
return self.collection_validator.added_source_variables()
def validate_with_variables(
self, source_variables: List[str], override_variables: Dict[str, str]
) -> None:
"""
Validates any source variables added by the collection
"""
self.collection_validator.validate_with_variables(
source_variables=source_variables, override_variables=override_variables
)
DownloaderOptionsT = TypeVar("DownloaderOptionsT", bound=DownloaderValidator)
class URLDownloadState:
def __init__(self, entries_total: int):
self.entries_total = entries_total
self.entries_downloaded = 0
class EmptyPluginOptions(PluginOptions):
_optional_keys = {"no-op"}
class BaseDownloaderPlugin(Plugin[EmptyPluginOptions], ABC):
def __init__(
self,
overrides: Overrides,
enhanced_download_archive: EnhancedDownloadArchive,
):
super().__init__(
# Downloader plugins do not have exposed YAML options, so keep it blank.
# Use init instead.
plugin_options=EmptyPluginOptions(name=self.__class__.__name__, value={}),
overrides=overrides,
enhanced_download_archive=enhanced_download_archive,
)
class BaseDownloader(DownloadArchiver, Generic[DownloaderOptionsT], ABC):
downloader_options_type: Type[DownloaderValidator] = DownloaderValidator
def __init__(
self,
download_options: DownloaderOptionsT,
enhanced_download_archive: EnhancedDownloadArchive,
download_ytdl_options: YTDLOptionsBuilder,
metadata_ytdl_options: YTDLOptionsBuilder,
overrides: Overrides,
):
super().__init__(enhanced_download_archive=enhanced_download_archive)
self.download_options = download_options
self.overrides = overrides
self._download_ytdl_options_builder = download_ytdl_options
self._metadata_ytdl_options_builder = metadata_ytdl_options
@abc.abstractmethod
def download_metadata(self) -> Iterable[Entry]:
"""Gathers metadata of all entries to download"""
@abc.abstractmethod
def download(self, entry: Entry) -> Entry:
"""The function to perform the download of all media entries"""
# pylint: disable=no-self-use
def added_plugins(self) -> List[BaseDownloaderPlugin]:
"""Add these plugins from the Downloader to the subscription"""
return []
# pylint: enable=no-self-use
class YtDlpThumbnailPlugin(BaseDownloaderPlugin):
def __init__(
self,
overrides: Overrides,
enhanced_download_archive: EnhancedDownloadArchive,
collection_urls: List[UrlValidator],
):
super().__init__(
overrides=overrides,
enhanced_download_archive=enhanced_download_archive,
)
self._thumbnails_downloaded: Set[str] = set()
self._collection_url_mapping: Dict[str, UrlValidator] = {
self.overrides.apply_formatter(collection_url.url): collection_url
for collection_url in collection_urls
}
def _download_parent_thumbnails(
self,
thumbnail_list_info: UrlThumbnailListValidator,
entry: Entry,
parent: EntryParent,
) -> None:
"""
Downloads and moves channel avatar and banner images to the output directory.
"""
for thumbnail_info in thumbnail_list_info.list:
thumbnail_name = self.overrides.apply_formatter(thumbnail_info.name, entry=entry)
thumbnail_id = self.overrides.apply_formatter(thumbnail_info.uid)
# If latest entry, always update the thumbnail on each entry
if thumbnail_id == ThumbnailTypes.LATEST_ENTRY:
# Make sure the entry's thumbnail is converted to jpg
convert_download_thumbnail(entry, error_if_not_found=False)
# always save in dry-run even if it doesn't exist...
if self.is_dry_run or os.path.isfile(entry.get_download_thumbnail_path()):
self.save_file(
file_name=entry.get_download_thumbnail_name(),
output_file_name=thumbnail_name,
copy_file=True,
)
self._thumbnails_downloaded.add(thumbnail_name)
continue
# If not latest entry and the thumbnail has already been downloaded, then skip
if thumbnail_name in self._thumbnails_downloaded:
continue
if (thumbnail_url := parent.get_thumbnail_url(thumbnail_id=thumbnail_id)) is None:
download_logger.debug("Failed to find thumbnail id '%s'", thumbnail_id)
continue
if download_and_convert_url_thumbnail(
thumbnail_url=thumbnail_url,
output_thumbnail_path=str(Path(self.working_directory) / thumbnail_name),
):
self.save_file(file_name=thumbnail_name)
self._thumbnails_downloaded.add(thumbnail_name)
else:
download_logger.debug("Failed to download thumbnail id '%s'", thumbnail_id)
def _download_url_thumbnails(self, collection_url: UrlValidator, entry: Entry):
"""
After all media entries have been downloaded, post processed, and moved to the output
directory, run this function. This lets the downloader add any extra files directly to the
output directory, for things like YT channel image, banner.
"""
if entry.kwargs_contains(PLAYLIST_ENTRY):
self._download_parent_thumbnails(
thumbnail_list_info=collection_url.playlist_thumbnails,
entry=entry,
parent=EntryParent(
entry.kwargs(PLAYLIST_ENTRY), working_directory=self.working_directory
),
)
if entry.kwargs_contains(SOURCE_ENTRY):
self._download_parent_thumbnails(
thumbnail_list_info=collection_url.source_thumbnails,
entry=entry,
parent=EntryParent(
entry.kwargs(SOURCE_ENTRY), working_directory=self.working_directory
),
)
def modify_entry(self, entry: Entry) -> Optional[Entry]:
"""
Use the entry to download thumbnails (or move if LATEST_ENTRY)
"""
if entry.kwargs(COLLECTION_URL) in self._collection_url_mapping:
self._download_url_thumbnails(
collection_url=self._collection_url_mapping[entry.kwargs(COLLECTION_URL)],
entry=entry,
)
return entry
class YtDlpCollectionVariablePlugin(BaseDownloaderPlugin):
def __init__(
self,
overrides: Overrides,
enhanced_download_archive: EnhancedDownloadArchive,
collection_urls: List[UrlValidator],
):
super().__init__(
overrides=overrides,
enhanced_download_archive=enhanced_download_archive,
)
self._thumbnails_downloaded: Set[str] = set()
self._collection_url_mapping: Dict[str, UrlValidator] = {
self.overrides.apply_formatter(collection_url.url): collection_url
for collection_url in collection_urls
}
def modify_entry_metadata(self, entry: Entry) -> Optional[Entry]:
"""
Add collection variables to the entry
"""
collection_url: Optional[UrlValidator] = self._collection_url_mapping.get(
entry.kwargs(COLLECTION_URL)
)
if collection_url:
entry.add_variables(variables_to_add=collection_url.variables.dict_with_format_strings)
return entry
class YtDlpDownloader(BaseDownloader[DownloaderOptionsT], ABC):
"""
Class that interacts with ytdl to perform the download of metadata and content,
and should translate that to list of Entry objects.
"""
def added_plugins(self) -> List[Plugin]:
"""
Adds
1. URL thumbnail download plugin
2. Collection variable plugin to add to each entry
"""
return [
YtDlpThumbnailPlugin(
overrides=self.overrides,
enhanced_download_archive=self._enhanced_download_archive,
collection_urls=self.collection.urls.list,
),
YtDlpCollectionVariablePlugin(
overrides=self.overrides,
enhanced_download_archive=self._enhanced_download_archive,
collection_urls=self.collection.urls.list,
),
]
@classmethod
def ytdl_option_defaults(cls) -> Dict:
"""
.. code-block:: yaml
ytdl_options:
ignoreerrors: True # ignore errors like hidden videos, age restriction, etc
"""
return {"ignoreerrors": True}
def __init__(
self,
download_options: DownloaderOptionsT,
enhanced_download_archive: EnhancedDownloadArchive,
download_ytdl_options: YTDLOptionsBuilder,
metadata_ytdl_options: YTDLOptionsBuilder,
overrides: Overrides,
):
"""
Parameters
----------
download_options
Options validator for this downloader
enhanced_download_archive
Download archive
download_ytdl_options
YTDL options builder for downloading media
metadata_ytdl_options
YTDL options builder for downloading metadata
overrides
Override variables
"""
super().__init__(
download_options=download_options,
enhanced_download_archive=enhanced_download_archive,
download_ytdl_options=download_ytdl_options,
metadata_ytdl_options=metadata_ytdl_options,
overrides=overrides,
)
self._downloaded_entries: Set[str] = set()
self._url_state: Optional[URLDownloadState] = None
@property
def download_ytdl_options(self) -> Dict:
"""
Returns
-------
YTLD options dict for downloading
"""
return (
self._download_ytdl_options_builder.clone()
.add(self.ytdl_option_defaults(), before=True)
.to_dict()
)
@property
def metadata_ytdl_options(self) -> Dict:
"""
Returns
-------
YTDL options dict for fetching metadata
"""
return (
self._metadata_ytdl_options_builder.clone()
.add(self.ytdl_option_defaults(), before=True)
.to_dict()
)
@property
def is_dry_run(self) -> bool:
"""
Returns
-------
True if dry-run is enabled. False otherwise.
"""
return self.download_ytdl_options.get("skip_download", False)
@property
def is_entry_thumbnails_enabled(self) -> bool:
"""
Returns
-------
True if entry thumbnails should be downloaded. False otherwise.
"""
return self.download_ytdl_options.get("writethumbnail", False)
###############################################################################################
# DOWNLOAD FUNCTIONS
def _is_downloaded(self, entry: Entry) -> bool:
return entry.ytdl_uid() in self._downloaded_entries
def _mark_downloaded(self, entry: Entry) -> None:
self._downloaded_entries.add(entry.ytdl_uid())
@property
def collection(self) -> MultiUrlValidator:
"""Return the download options collection"""
return self.download_options.collection_validator
@contextlib.contextmanager
def _separate_download_archives(self, clear_info_json_files: bool = False):
"""
Separate download archive writing between collection urls. This is so break_on_existing
does not break when downloading from subset urls.
Parameters
----------
clear_info_json_files
Whether to delete info.json files after yield
"""
archive_path = self.download_ytdl_options.get("download_archive", "")
backup_archive_path = f"{archive_path}.backup"
# If archive path exists, maintain download archive is enable
if archive_file_exists := archive_path and os.path.isfile(archive_path):
archive_file_exists = True
# If a backup exists, it's the one prior to any downloading, use that.
if os.path.isfile(backup_archive_path):
FileHandler.copy(src_file_path=backup_archive_path, dst_file_path=archive_path)
# If not, create the backup
else:
FileHandler.copy(src_file_path=archive_path, dst_file_path=backup_archive_path)
yield
# If an archive path did not exist at first, but now exists, delete it
if not archive_file_exists and os.path.isfile(archive_path):
FileHandler.delete(file_path=archive_path)
# If the archive file did exist, restore the backup
elif archive_file_exists:
FileHandler.move(src_file_path=backup_archive_path, dst_file_path=archive_path)
# Clear info json files if true
if clear_info_json_files:
info_json_files = [
Path(self.working_directory) / path
for path in os.listdir(self.working_directory)
if path.endswith(".info.json")
]
for info_json_file in info_json_files:
FileHandler.delete(info_json_file)
def _extract_entry_info_with_retry(self, entry: Entry) -> Entry:
download_entry_dict = YTDLP.extract_info_with_retry(
ytdl_options_overrides=self.download_ytdl_options,
is_downloaded_fn=None if self.is_dry_run else entry.is_downloaded,
is_thumbnail_downloaded_fn=None
if (self.is_dry_run or not self.is_entry_thumbnails_enabled)
else entry.is_thumbnail_downloaded,
url=entry.webpage_url,
)
return Entry(download_entry_dict, working_directory=self.working_directory)
def _iterate_child_entries(
self, url_validator: UrlValidator, entries: List[Entry]
) -> Iterator[Entry]:
entries_to_iterate = entries
if url_validator.download_reverse:
entries_to_iterate = reversed(entries)
for entry in entries_to_iterate:
self._url_state.entries_downloaded += 1
if self._is_downloaded(entry):
download_logger.info(
"Already downloaded entry %d/%d: %s",
self._url_state.entries_downloaded,
self._url_state.entries_total,
entry.title,
)
continue
yield entry
self._mark_downloaded(entry)
def _iterate_parent_entry(
self, url_validator: UrlValidator, parent: EntryParent
) -> Iterator[Entry]:
for entry_child in self._iterate_child_entries(
url_validator=url_validator, entries=parent.entry_children()
):
yield entry_child
# Recursion the parent's parent entries
for parent_child in reversed(parent.parent_children()):
for entry_child in self._iterate_parent_entry(
url_validator=url_validator, parent=parent_child
):
yield entry_child
def _download_url_metadata(
self, collection_url: UrlValidator
) -> Tuple[List[EntryParent], List[Entry]]:
"""
Downloads only info.json files and forms EntryParent trees
"""
url = self.overrides.apply_formatter(collection_url.url)
with self._separate_download_archives():
entry_dicts = YTDLP.extract_info_via_info_json(
working_directory=self.working_directory,
ytdl_options_overrides=self.metadata_ytdl_options,
log_prefix_on_info_json_dl="Downloading metadata for",
url=url,
)
parents = EntryParent.from_entry_dicts(
url=url,
entry_dicts=entry_dicts,
working_directory=self.working_directory,
)
orphans = EntryParent.from_entry_dicts_with_no_parents(
parents=parents, entry_dicts=entry_dicts, working_directory=self.working_directory
)
return parents, orphans
def _iterate_entries(
self,
url_validator: UrlValidator,
parents: List[EntryParent],
orphans: List[Entry],
) -> Iterator[Entry]:
"""
Downloads the leaf entries from EntryParent trees
"""
# Delete info json files afterwards so other collection URLs do not use them
with self._separate_download_archives(clear_info_json_files=True):
for parent in parents:
for entry_child in self._iterate_parent_entry(
url_validator=url_validator, parent=parent
):
yield entry_child
for orphan in self._iterate_child_entries(url_validator=url_validator, entries=orphans):
yield orphan
def download_metadata(self) -> Iterable[Entry]:
"""The function to perform the download of all media entries"""
# download the bottom-most urls first since they are top-priority
for collection_url in reversed(self.collection.urls.list):
parents, orphan_entries = self._download_url_metadata(collection_url=collection_url)
# TODO: Encapsulate this logic into its own class
self._url_state = URLDownloadState(
entries_total=sum(parent.num_children() for parent in parents) + len(orphan_entries)
)
download_logger.info(
"Beginning downloads for %s", self.overrides.apply_formatter(collection_url.url)
)
for entry in self._iterate_entries(
url_validator=collection_url, parents=parents, orphans=orphan_entries
):
# Add the collection URL to the info_dict to trace where it came from
entry.add_kwargs(
{COLLECTION_URL: self.overrides.apply_formatter(collection_url.url)}
)
yield entry
def download(self, entry: Entry) -> Entry:
"""
Parameters
----------
entry
Entry to download
Returns
-------
The entry that was downloaded successfully
"""
download_logger.info(
"Downloading entry %d/%d: %s",
self._url_state.entries_downloaded,
self._url_state.entries_total,
entry.title,
)
download_entry = self._extract_entry_info_with_retry(entry=entry)
upload_date_idx = self._enhanced_download_archive.mapping.get_num_entries_with_upload_date(
upload_date_standardized=entry.upload_date_standardized
)
download_idx = self._enhanced_download_archive.num_entries
entry.add_kwargs(
{
# Subtitles are not downloaded in metadata run, only here, so move over
REQUESTED_SUBTITLES: download_entry.kwargs_get(REQUESTED_SUBTITLES),
# Same with sponsorblock chapters
SPONSORBLOCK_CHAPTERS: download_entry.kwargs_get(SPONSORBLOCK_CHAPTERS),
COMMENTS: download_entry.kwargs_get(COMMENTS),
# Tracks number of entries downloaded
DOWNLOAD_INDEX: download_idx,
# Tracks number of entries with the same upload date to make them unique
UPLOAD_DATE_INDEX: upload_date_idx,
}
)
return entry