[FEATURE] Match filters plugin (#365)

docs/config.rst

@@ -152,6 +152,14 @@ file_convert
 
 -------------------------------------------------------------------------------
 
+match_filters
+'''''''''''''
+.. autoclass:: ytdl_sub.plugins.match_filters.MatchFiltersOptions()
+  :members:
+  :exclude-members: partial_validate
+
+-------------------------------------------------------------------------------
+
 music_tags
 ''''''''''
 .. autoclass:: ytdl_sub.plugins.music_tags.MusicTagsOptions()

src/ytdl_sub/config/preset_class_mappings.py

@@ -15,6 +15,7 @@ from ytdl_sub.plugins.chapters import ChaptersPlugin
 from ytdl_sub.plugins.date_range import DateRangePlugin
 from ytdl_sub.plugins.file_convert import FileConvertPlugin
 from ytdl_sub.plugins.internal.view import ViewPlugin
+from ytdl_sub.plugins.match_filters import MatchFiltersPlugin
 from ytdl_sub.plugins.music_tags import MusicTagsPlugin
 from ytdl_sub.plugins.nfo_tags import NfoTagsPlugin
 from ytdl_sub.plugins.output_directory_nfo_tags import OutputDirectoryNfoTagsPlugin
@@ -121,6 +122,7 @@ class PluginMapping:
         "audio_extract": AudioExtractPlugin,
         "date_range": DateRangePlugin,
         "file_convert": FileConvertPlugin,
+        "match_filters": MatchFiltersPlugin,
         "music_tags": MusicTagsPlugin,
         "video_tags": VideoTagsPlugin,
         "nfo_tags": NfoTagsPlugin,

src/ytdl_sub/plugins/match_filters.py

@@ -0,0 +1,78 @@
+from typing import Any
+from typing import Dict
+from typing import List
+from typing import Optional
+
+from yt_dlp import match_filter_func
+
+from ytdl_sub.plugins.plugin import Plugin
+from ytdl_sub.plugins.plugin import PluginOptions
+from ytdl_sub.utils.logger import Logger
+from ytdl_sub.validators.validators import StringListValidator
+
+logger = Logger.get("match_filters")
+
+
+class MatchFiltersOptions(PluginOptions):
+    """
+    Set ``--match-filters``` to pass into yt-dlp to filter entries from being downloaded.
+    Uses the same syntax as yt-dlp.
+
+    Usage:
+
+    .. code-block:: yaml
+
+       presets:
+         my_example_preset:
+           match_filters:
+             filters: "original_url!*=/shorts/"
+
+    Supports one or multiple filters:
+
+    .. code-block:: yaml
+
+       presets:
+         my_example_preset:
+           match_filters:
+             filters:
+               - "original_url!*=/shorts/"
+               - "!is_live"
+    """
+
+    _required_keys = {"filters"}
+
+    @classmethod
+    def partial_validate(cls, name: str, value: Any) -> None:
+        """Ensure filters looks right"""
+        if isinstance(value, dict):
+            value["filters"] = value.get("filters", [""])
+        _ = cls(name, value)
+
+    def __init__(self, name, value):
+        super().__init__(name, value)
+        self._filters = self._validate_key(key="filters", validator=StringListValidator).list
+
+    @property
+    def filters(self) -> List[str]:
+        """
+        The filters themselves. If used multiple times, the filter matches if at least one of the
+        conditions are met.
+        """
+        return [validator.value for validator in self._filters]
+
+
+class MatchFiltersPlugin(Plugin[MatchFiltersOptions]):
+    plugin_options_type = MatchFiltersOptions
+
+    def ytdl_options(self) -> Optional[Dict]:
+        """
+        Returns
+        -------
+        match_filter after calling the utility function for it
+        """
+        match_filters: List[str] = []
+        for filter_str in self.plugin_options.filters:
+            logger.debug("Adding match-filter %s", filter_str)
+            match_filters.append(filter_str)
+
+        return {"match_filter": match_filter_func(match_filters)}

src/ytdl_sub/subscriptions/subscription_ytdl_options.py

@@ -12,6 +12,7 @@ from ytdl_sub.plugins.audio_extract import AudioExtractPlugin
 from ytdl_sub.plugins.chapters import ChaptersPlugin
 from ytdl_sub.plugins.date_range import DateRangePlugin
 from ytdl_sub.plugins.file_convert import FileConvertPlugin
+from ytdl_sub.plugins.match_filters import MatchFiltersPlugin
 from ytdl_sub.plugins.plugin import Plugin
 from ytdl_sub.plugins.subtitles import SubtitlesPlugin
 from ytdl_sub.ytdl_additions.enhanced_download_archive import EnhancedDownloadArchive
@@ -110,6 +111,7 @@ class SubscriptionYTDLOptions:
             self._global_options,
             self._output_options,
             self._plugin_ytdl_options(DateRangePlugin),
+            self._plugin_ytdl_options(MatchFiltersPlugin),
             self._user_ytdl_options,  # user ytdl options...
             self._info_json_only_options,  # then info_json_only options
         )

tests/e2e/plugins/test_match_filters.py

@@ -0,0 +1,36 @@
+import pytest
+
+from ytdl_sub.subscriptions.subscription import Subscription
+
+
+@pytest.fixture
+def preset_dict(output_directory):
+    return {
+        "preset": "music_video",
+        "download": {"url": "https://www.youtube.com/watch?v=2zYF9JLHDmA"},
+        "output_options": {"output_directory": output_directory},
+        # download the worst format so it is fast
+        "ytdl_options": {
+            "postprocessor_args": {"ffmpeg": ["-bitexact"]},  # Must add this for reproducibility
+        },
+        "match_filters": {"filters": ["is_live", "view_count <=? 10"]},
+    }
+
+
+class TestFileConvert:
+    @pytest.mark.parametrize("dry_run", [True, False])
+    def test_match_filters(
+        self,
+        music_video_config,
+        preset_dict,
+        output_directory,
+        dry_run,
+    ):
+        subscription = Subscription.from_dict(
+            config=music_video_config,
+            preset_name="match_filter_test",
+            preset_dict=preset_dict,
+        )
+
+        transaction_log = subscription.download(dry_run=dry_run)
+        assert transaction_log.is_empty