WIP

src/yt2doc/formatting/interfaces.py

@@ -3,6 +3,13 @@
 from pydantic import BaseModel
 
 from yt2doc.extraction import interfaces as extraction_interfaces
+from yt2doc.transcription import interfaces as transcription_interfaces
+
+
+class Sentence(BaseModel):
+    start_second: float
+    end_second: float
+    text: str
 
 
 class Chapter(BaseModel):
@@ -20,10 +27,16 @@ class FormattedPlaylist(BaseModel):
     transcripts: typing.Sequence[FormattedTranscript]
 
 
+class IParagraphsSegmenter(typing.Protocol):
+    def segment(
+        self, transcription_segments: typing.Sequence[transcription_interfaces.Segment]
+    ) -> typing.Sequence[typing.Sequence[Sentence]]: ...
+
+
 class ILLMAdapter(typing.Protocol):
     def get_topic_changing_paragraph_indexes(
         self, paragraphs: typing.List[typing.List[str]]
-    ) -> typing.List[int]: ...
+    ) -> typing.Sequence[int]: ...
 
     def generate_title_for_paragraphs(
         self, paragraphs: typing.List[typing.List[str]]

src/yt2doc/formatting/llm_adapter.py

@@ -11,7 +11,7 @@ def __init__(self, llm_client: Instructor, llm_model: str) -> None:
 
     def get_topic_changing_paragraph_indexes(
         self, paragraphs: typing.List[typing.List[str]]
-    ) -> typing.List[int]:
+    ) -> typing.Sequence[int]:
         def validate_paragraph_indexes(v: typing.List[int]) -> typing.List[int]:
             n = len(paragraphs)
             unique_values = set(v)

src/yt2doc/formatting/paragraphs_segmenter.py

@@ -0,0 +1,23 @@
+import typing
+
+from wtpsplit import SaT
+
+from src.yt2doc.formatting import interfaces
+from yt2doc.transcription import interfaces as transcription_interfaces
+
+
+class ParagraphsSegmenter:
+    def __init__(self, sat: SaT) -> None:
+        self.sat = sat
+
+    def segment(
+        self, transcription_segments: typing.Sequence[transcription_interfaces.Segment]
+    ) -> typing.Sequence[typing.Sequence[interfaces.Sentence]]:
+        transcription_segment_texts = [s.text for s in transcription_segments]
+        full_text = "".join(transcription_segment_texts)
+        paragraphed_texts: typing.List[typing.List[str]] = self.sat.split(
+            full_text, do_paragraph_segmentation=True, verbose=True
+        )
+
+        # Paragraph-Transcription Segment timestamp alignment
+

src/yt2doc/transcription/faster_whisper_adapter.py

@@ -22,6 +22,8 @@ def transcribe(
             audio=audio_path, initial_prompt=initial_prompt, vad_filter=True
         )
         return (
-            interfaces.Segment(start=segment.start, end=segment.end, text=segment.text)
+            interfaces.Segment(
+                start_second=segment.start, end_second=segment.end, text=segment.text
+            )
             for segment in segments
         )

src/yt2doc/transcription/interfaces.py

@@ -8,8 +8,8 @@
 
 
 class Segment(BaseModel):
-    start: float
-    end: float
+    start_second: float
+    end_second: float
     text: str
 
 

src/yt2doc/transcription/transcriber.py

@@ -161,15 +161,15 @@ def transcribe(
                 )
                 for segment in segments:
                     aligned_segment = interfaces.Segment(
-                        start=chapter.start_time + segment.start,
-                        end=chapter.start_time + segment.end,
+                        start_second=chapter.start_time + segment.start_second,
+                        end_second=chapter.start_time + segment.end_second,
                         text=self._fix_comma(
                             segment_text=segment.text, language_code=language_code
                         ),
                     )
                     chapter_segments.append(aligned_segment)
-                    progress_bar.update(aligned_segment.end - current_timestamp)
-                    current_timestamp = aligned_segment.end
+                    progress_bar.update(aligned_segment.end_second - current_timestamp)
+                    current_timestamp = aligned_segment.end_second
 
                 chaptered_transcriptions.append(
                     interfaces.ChapterTranscription(

src/yt2doc/transcription/whisper_cpp_adapter.py

@@ -46,8 +46,8 @@ def _parse_whisper_line(self, line: str) -> interfaces.Segment:
         if match:
             start_time, end_time, text = match.groups()
             return interfaces.Segment(
-                start=self._time_to_seconds(start_time),
-                end=self._time_to_seconds(end_time),
+                start_second=self._time_to_seconds(start_time),
+                end_second=self._time_to_seconds(end_time),
                 text=text,
             )
         raise CannotParseWhisperCppLineException(

tests/integ/formatting/test_formatter.py

@@ -26,8 +26,8 @@ def parse_whisper_line(line: str) -> Segment:
     if match:
         start_time, end_time, text = match.groups()
         return Segment(
-            start=time_to_seconds(start_time),
-            end=time_to_seconds(end_time),
+            start_second=time_to_seconds(start_time),
+            end_second=time_to_seconds(end_time),
             text=text,
         )
     assert False, "Line does not match expected whisper segment pattern."
@@ -96,7 +96,7 @@ def test_format_chaptered_transcript_basic(
     formatter = MarkdownFormatter(sat=sat)
 
     segments_dicts = [
-        {"start": segment.start, "end": segment.end, "text": segment.text}
+        {"start": segment.start_second, "end": segment.end_second, "text": segment.text}
         for segment in mock_transcript_segments
     ]
 
@@ -158,7 +158,7 @@ def mock_generate_title_for_paragraphs(
     formatter = MarkdownFormatter(sat=sat, topic_segmenter=segmenter)
 
     segments_dicts = [
-        {"start": segment.start, "end": segment.end, "text": segment.text}
+        {"start": segment.start_second, "end": segment.end_second, "text": segment.text}
         for segment in mock_transcript_segments
     ]
 
@@ -183,4 +183,3 @@ def mock_generate_title_for_paragraphs(
     # Assert
     assert "# Test Video Title" in formatted_output.transcript
     assert "## Chapter Title" in formatted_output.transcript
-    # mock_llm_client.chat.completions.create.assert_called_once()