fix: whisper repetition loops, meeting transcript punctuation
- transcription: add temperature_inc=0 to whispercpp to disable fallback (prevents loops) - pipeline: punctuate meeting transcript in one pass (parallel with summarize) - output: write_meeting_docs accepts pre-built transcript_text - llm: punctuate prompt preserves speaker labels Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
@@ -152,6 +152,7 @@ def write_meeting_docs(
|
||||
dt: "datetime | None" = None,
|
||||
title: str = "",
|
||||
tldr: str = "",
|
||||
transcript_text: str = "",
|
||||
) -> dict[str, str]:
|
||||
"""Write index (in output_dir), transkript + zusammenfassung (in subdir)."""
|
||||
if dt is None:
|
||||
@@ -172,10 +173,13 @@ def write_meeting_docs(
|
||||
os.makedirs(subdir, exist_ok=True)
|
||||
|
||||
# --- transkript (in subdir) ---
|
||||
transcript_lines = []
|
||||
for speaker, text in aligned_segments:
|
||||
transcript_lines.append(f"**{speaker}:** {text}\n")
|
||||
transcript_content = "\n".join(transcript_lines)
|
||||
if transcript_text:
|
||||
transcript_content = transcript_text
|
||||
else:
|
||||
transcript_lines = []
|
||||
for speaker, text in aligned_segments:
|
||||
transcript_lines.append(f"**{speaker}:** {text}\n")
|
||||
transcript_content = "\n".join(transcript_lines)
|
||||
transkript_filename = f"{base}-transkript.md"
|
||||
transkript_path = os.path.join(subdir, transkript_filename)
|
||||
with open(transkript_path, "w", encoding="utf-8") as f:
|
||||
|
||||
Reference in New Issue
Block a user