chapters and split by chapters plugins WIP

src/ytdl_sub/plugins/chapters.py

@@ -0,0 +1,107 @@
+import re
+from typing import Dict
+from typing import List
+from typing import Optional
+from typing import Set
+
+from ytdl_sub.entries.entry import Entry
+from ytdl_sub.plugins.plugin import Plugin
+from ytdl_sub.plugins.plugin import PluginOptions
+from ytdl_sub.validators.regex_validator import RegexListValidator
+from ytdl_sub.validators.string_select_validator import StringSelectValidator
+from ytdl_sub.validators.validators import ListValidator
+
+SPONSORBLOCK_HIGHLIGHT_CATEGORIES: Set[str] = {"poi_highlight"}
+SPONSORBLOCK_CATEGORIES: Set[str] = SPONSORBLOCK_HIGHLIGHT_CATEGORIES | {
+    "sponsor",
+    "intro",
+    "outro",
+    "selfpromo",
+    "preview",
+    "filler",
+    "interaction",
+    "music_offtopic",
+}
+
+
+class SponsorBlockCategoriesValidator(StringSelectValidator):
+    _expected_value_type = "sponsorblock category"
+    _select_values = {"all"} | SPONSORBLOCK_CATEGORIES
+
+
+class SponsorBlockCategoryListValidator(ListValidator[SponsorBlockCategoriesValidator]):
+    _expected_value_type = "sponsorblock category"
+    _inner_list_type = SponsorBlockCategoriesValidator
+
+
+class ChaptersOptions(PluginOptions):
+    """
+    Add chapters to video files if they are present. Options to add SponsorBlock chapters and
+    remove them or existing chapters.
+
+    Usage:
+
+    .. code-block:: yaml
+
+       presets:
+         my_example_preset:
+           chapters:
+             sponsorblock_categories:
+               - "outro"
+               - "selfpromo"
+               - "preview"
+               - "interaction"
+               - "sponsor"
+               - "music_offtopic"
+               - "intro"
+             remove_sponsorblock_categories: "all"
+             remove_chapters_regex:
+               - "sponsor"
+
+    """
+
+    _optional_keys = {
+        "sponsorblock_categories",
+        "remove_sponsorblock_categories",
+        "remove_chapters_regex",
+    }
+
+    def __init__(self, name, value):
+        super().__init__(name, value)
+        self._sponsorblock_categories = self._validate_key_if_present(
+            key="sponsorblock_categories", validator=SponsorBlockCategoryListValidator
+        )
+        self._remove_sponsorblock_categories = self._validate_key_if_present(
+            key="remove_sponsorblock_categories", validator=SponsorBlockCategoryListValidator
+        )
+        self._remove_chapters_regex = self._validate_key_if_present(
+            key="remove_chapters_regex", validator=RegexListValidator
+        )
+
+    @property
+    def sponsorblock_categories(self) -> Optional[List[str]]:
+        if self._sponsorblock_categories:
+            return [validator.value for validator in self._sponsorblock_categories.list]
+        return None
+
+    @property
+    def remove_sponsorblock_categories(self) -> Optional[List[str]]:
+        if self._remove_sponsorblock_categories:
+            return [validator.value for validator in self._remove_sponsorblock_categories.list]
+        return None
+
+    @property
+    def remove_chapters_regex(self) -> Optional[List[re.Pattern]]:
+        if self._remove_chapters_regex:
+            return [validator.compiled_regex for validator in self._remove_chapters_regex.list]
+        return None
+
+
+class ChaptersPlugin(Plugin[ChaptersOptions]):
+    plugin_options_type = ChaptersOptions
+
+    def modify_entry(self, entry: Entry) -> Optional[Entry]:
+        """
+        Tags the entry's audio file using values defined in the metadata options
+        """
+        return entry

src/ytdl_sub/plugins/split_by_chapters.py

@@ -0,0 +1,54 @@
+from typing import Dict
+from typing import Optional
+
+from ytdl_sub.entries.entry import Entry
+from ytdl_sub.plugins.plugin import Plugin
+from ytdl_sub.plugins.plugin import PluginOptions
+from ytdl_sub.validators.string_select_validator import StringSelectValidator
+
+
+class WhenNoChaptersValidator(StringSelectValidator):
+    _expected_value_type = "when no chapters option"
+    _select_values = {"pass", "drop", "error"}
+
+
+class SplitByChaptersOptions(PluginOptions):
+    """
+    Splits a file by chapters into multiple files. Each file becomes its own entry with ``title``
+    set to its chapter name, and is processed separately by other plugins.
+
+    Usage:
+
+    .. code-block:: yaml
+
+       presets:
+         my_example_preset:
+           split_by_chapters:
+             when_no_chapters: "pass"  # "drop"/"error"
+    """
+
+    _required_keys = {"when_no_chapters"}
+
+    def __init__(self, name, value):
+        super().__init__(name, value)
+        self._when_no_chapters = self._validate_key(
+            key="when_no_chapters", validator=WhenNoChaptersValidator
+        ).value
+
+    @property
+    def when_no_chapters(self) -> str:
+        """
+        Behavior to perform when no chapters are present. Supports "pass" (continue processing),
+        "drop" (exclude it from output), and "error" (stop processing for everything).
+        """
+        return self._when_no_chapters
+
+
+class SplitByChaptersPlugin(Plugin[SplitByChaptersOptions]):
+    plugin_options_type = SplitByChaptersOptions
+
+    def modify_entry(self, entry: Entry) -> Optional[Entry]:
+        """
+        Tags the entry's audio file using values defined in the metadata options
+        """
+        return entry

src/ytdl_sub/validators/regex_validator.py

@@ -28,6 +28,15 @@ class RegexValidator(StringValidator):
         """
         return self._compiled_regex.groups
 
+    @property
+    def compiled_regex(self) -> re.Pattern:
+        """
+        Returns
+        -------
+        The regex compiled
+        """
+        return self._compiled_regex
+
     def match(self, input_str: str) -> Optional[List[str]]:
         """
         Parameters

tests/e2e/plugins/test_audio_extract.py

@@ -7,10 +7,12 @@ from ytdl_sub.subscriptions.subscription import Subscription
 
 
 @pytest.fixture
-def single_song_preset_dict(output_directory):
+def single_song_preset_dict(output_directory, timestamps_file_path):
     return {
         "preset": "yt_song",
-        "youtube": {"video_url": "https://www.youtube.com/watch?v=2lAe1cqCOXo"},
+        "youtube": {
+            "video_url": "https://www.youtube.com/watch?v=gr0XWmEbiMQ",
+        },
         "output_options": {"output_directory": output_directory},
         # download the worst format so it is fast
         "ytdl_options": {

tests/e2e/resources/transaction_log_summaries/plugins/test_audio_extract_single.txt

@@ -1,11 +1,11 @@
 Files created in '{output_directory}'
 ----------------------------------------
-YouTube Rewind 2019： For the Record ｜ #YouTubeRewind.mp3
+Gabor Szabo - Dreams (1968) [full album].mp3
   Music Tags:
     album: Singles
-    albumartist: YouTube
-    artist: YouTube
+    albumartist: Cubensis Records
+    artist: Cubensis Records
     genre: Unset
-    title: YouTube Rewind 2019: For the Record | #YouTubeRewind
+    title: Gabor Szabo - Dreams (1968) [full album]
     track: 1
-    year: 2019
+    year: 2015