WIP

src/yt2doc/extraction/extractor.py

@@ -11,12 +11,12 @@
 class Extractor:
     def __init__(
         self,
-        video_info_extractor: youtube_interfaces.IYtVideoInfoExtractor,
+        media_info_extractor: youtube_interfaces.IYtMediaInfoExtractor,
         transcriber: transcription_interfaces.ITranscriber,
         file_cache: interfaces.IFileCache,
         ignore_source_chapters: bool,
     ) -> None:
-        self.yt_dlp_adapter = video_info_extractor
+        self.yt_dlp_adapter = media_info_extractor
         self.transcriber = transcriber
         self.file_cache = file_cache
         self.ignore_source_chapters = ignore_source_chapters
@@ -28,16 +28,16 @@ def extract_by_chapter(
     ) -> interfaces.ChapteredTranscript:
         logger.info(f"Extracting video {video_url} by chapter.")
 
-        video_info = self.yt_dlp_adapter.extract_video_info(video_url=video_url)
+        media_info = self.yt_dlp_adapter.extract_media_info(video_url=video_url)
 
         if self.ignore_source_chapters:
-            video_info.chapters = []
+            media_info.chapters = []
 
         if (
             not skip_cache
             and (
                 cached_chaptered_transcript := self.file_cache.get_chaptered_transcript(
-                    video_id=video_info.video_id
+                    video_id=media_info.video_id
                 )
             )
             is not None
@@ -52,7 +52,7 @@ def extract_by_chapter(
         with Timer() as transcribe_timer:
             transcript = self.transcriber.transcribe(
                 audio_path=audio_path,
-                video_info=video_info,
+                media_info=media_info,
             )
             transcripts_by_chapter = [
                 interfaces.TranscriptChapter(
@@ -65,14 +65,16 @@ def extract_by_chapter(
 
         chaptered_transcript = interfaces.ChapteredTranscript(
             url=video_url,
-            title=video_info.title,
+            title=media_info.title,
+            webpage_url=media_info.webpage_url,
+            webpage_url_domain=media_info.webpage_url_domain,
             chapters=transcripts_by_chapter,
-            chaptered_at_source=len(video_info.chapters) > 0,
+            chaptered_at_source=len(media_info.chapters) > 0,
             language=transcript.language,
         )
 
         self.file_cache.cache_chaptered_transcript(
-            video_id=video_info.video_id,
+            video_id=media_info.video_id,
             transcript=chaptered_transcript,
         )
 

src/yt2doc/extraction/interfaces.py

@@ -13,6 +13,8 @@ class TranscriptChapter(BaseModel):
 class ChapteredTranscript(BaseModel):
     url: str
     title: str
+    webpage_url: str
+    webpage_url_domain: str
     language: str
     chapters: typing.Sequence[TranscriptChapter]
     chaptered_at_source: bool

src/yt2doc/factories.py

@@ -66,13 +66,13 @@ def get_yt2doc(
     else:
         formatter = MarkdownFormatter(paragraphs_segmenter=paragraphs_segmenter)
 
-    video_info_extractor = MediaInfoExtractor(temp_dir=temp_dir)
+    media_info_extractor = MediaInfoExtractor(temp_dir=temp_dir)
     transcriber = Transcriber(
         temp_dir=temp_dir,
         whisper_adapter=whisper_adapter,
     )
     extractor = Extractor(
-        video_info_extractor=video_info_extractor,
+        media_info_extractor=media_info_extractor,
         transcriber=transcriber,
         file_cache=file_cache,
         ignore_source_chapters=ignore_source_chapters,

src/yt2doc/formatting/formatter.py

@@ -11,20 +11,29 @@ class MarkdownFormatter:
     def __init__(
         self,
         paragraphs_segmenter: interfaces.IParagraphsSegmenter,
+        # timestamp_paragraphs: bool,
         topic_segmenter: typing.Optional[interfaces.ITopicSegmenter] = None,
     ) -> None:
         self.paragraphs_segmenter = paragraphs_segmenter
         self.topic_segmenter = topic_segmenter
         self.video_title_template = "# {name}"
         self.chapter_title_template = "## {name}"
+        # self.timestamp_paragraphs = timestamp_paragraphs
 
     @staticmethod
     def _paragraphs_to_text(
         paragraphs: typing.Sequence[typing.Sequence[interfaces.Sentence]],
+        # timestamp_paragraphs: bool,
+        # webpage_url: str,
+        # webpage_url_domain: str,
     ) -> str:
         paragraph_texts = []
         for paragraph in paragraphs:
+            first_sentence = paragraph[0]
             paragraph_text = "".join(sentence.text for sentence in paragraph)
+            # if timestamp_paragraphs:
+            #     if webpage_url_domain == "youtube.com":
+            #         timestamp_prefix = "[\({}\)]()"
             paragraph_texts.append(paragraph_text)
         return "\n\n".join(paragraph_texts)
 

src/yt2doc/media/interfaces.py

@@ -14,6 +14,8 @@ class MediaChapter(BaseModel):
 class MediaInfo(BaseModel):
     video_id: str
     title: str
+    webpage_url: str
+    webpage_url_domain: str
     chapters: typing.Sequence[MediaChapter]
     description: str
 
@@ -23,7 +25,7 @@ class YtPlaylistInfo(BaseModel):
     video_urls: typing.Sequence[str]
 
 
-class IYtVideoInfoExtractor(typing.Protocol):
-    def extract_video_info(self, video_url: str) -> MediaInfo: ...
+class IYtMediaInfoExtractor(typing.Protocol):
+    def extract_media_info(self, video_url: str) -> MediaInfo: ...
     def extract_audio(self, video_url: str) -> Path: ...
     def extract_playlist_info(self, playlist_url: str) -> YtPlaylistInfo: ...

src/yt2doc/media/media_info_extractor.py

@@ -16,6 +16,7 @@
 class YtDLPResponse(BaseModel):
     video_id: str = Field(alias="id")
     webpage_url: str
+    webpage_url_domain: str
     title: str
     description: str
     chapters: typing.Optional[typing.Sequence[interfaces.MediaChapter]] = None
@@ -67,7 +68,7 @@ class MediaInfoExtractor:
     def __init__(self, temp_dir: Path):
         self.temp_dir = temp_dir
 
-    def extract_video_info(self, video_url: str) -> interfaces.MediaInfo:
+    def extract_media_info(self, video_url: str) -> interfaces.MediaInfo:
         ydl_opts = {
             "quiet": True,
         }
@@ -80,6 +81,8 @@ def extract_video_info(self, video_url: str) -> interfaces.MediaInfo:
         return interfaces.MediaInfo(
             video_id=parsed_response.video_id,
             title=parsed_response.title,
+            webpage_url=parsed_response.webpage_url,
+            webpage_url_domain=parsed_response.webpage_url_domain,
             chapters=_merge_short_chapters(parsed_response.chapters or []),
             description=parsed_response.description,
         )

src/yt2doc/transcription/interfaces.py

@@ -33,5 +33,5 @@ def transcribe(
 
 class ITranscriber(typing.Protocol):
     def transcribe(
-        self, audio_path: Path, video_info: youtube_interfaces.MediaInfo
+        self, audio_path: Path, media_info: youtube_interfaces.MediaInfo
     ) -> Transcription: ...

src/yt2doc/transcription/transcriber.py

@@ -69,18 +69,18 @@ def _clean_title(title: str, punctuations: punctuations.Punctuations) -> str:
     def _get_initial_prompt(
         self,
         language_code: str,
-        video_info: youtube_interfaces.MediaInfo,
+        media_info: youtube_interfaces.MediaInfo,
     ) -> str:
         punctuations_ = punctuations.get_punctuations(language_code=language_code)
         cleaned_title = self._clean_title(
-            title=video_info.title,
+            title=media_info.title,
             punctuations=punctuations_,
         )
         cleaned_video_description = self._clean_video_description(
-            video_info.description, punctuations=punctuations_
+            media_info.description, punctuations=punctuations_
         )
         chapter_titles = f"{punctuations_.comma}".join(
-            c.title for c in video_info.chapters
+            c.title for c in media_info.chapters
         )
         return f"{cleaned_title}{punctuations_.full_stop} {cleaned_video_description} {chapter_titles}"
 
@@ -114,14 +114,14 @@ def _convert_audio_to_wav(audio_path: Path) -> Path:
         return wav_audio_path
 
     def transcribe(
-        self, audio_path: Path, video_info: youtube_interfaces.MediaInfo
+        self, audio_path: Path, media_info: youtube_interfaces.MediaInfo
     ) -> interfaces.Transcription:
         wav_audio_path = self._convert_audio_to_wav(audio_path=audio_path)
 
         language_code = self.whisper_adapter.detect_language(audio_path=wav_audio_path)
         initial_prompt = self._get_initial_prompt(
             language_code=language_code,
-            video_info=video_info,
+            media_info=media_info,
         )
         logger.info(f"Initial prompt: {initial_prompt}")
 
@@ -133,8 +133,8 @@ def transcribe(
 
         rounded_full_audio_duration = round(full_audio_duration, 2)
         current_timestamp = 0.0
-        if len(video_info.chapters) > 0:
-            chapters = video_info.chapters
+        if len(media_info.chapters) > 0:
+            chapters = media_info.chapters
         else:
             chapters = [
                 youtube_interfaces.MediaChapter(