Dymas/abogen
1
0
Fork 0
mirror of https://github.com/denizsafak/abogen.git synced 2026-07-18 13:40:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

Refactor segment processing with overflow error handling

Browse Source 
fix: catch OverflowError in emit_text for very large numbers

Fixes #145

When text contains a very large number (e.g. a long decimal from a binary
hash), misaki's pipeline calls num2words() which raises OverflowError and
crashes the entire job. Wrap the segment iterator in try/except so the
chunk is skipped gracefully with a warning log instead of terminating.
This commit is contained in:
Yash
2026-05-22 10:07:38 +05:30
committed by GitHub
parent 9fa81fbe1e
commit da68f38b9b
1 changed files with 45 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files
abogen/webui/conversion_runner.py
+45 -39
View File
@@ -1868,45 +1868,51 @@ def run_conversion_job(job: Job) -> None:
split_pattern=split_pattern,
)
for segment in segment_iter:
canceller()
graphemes_raw = getattr(segment, "graphemes", "") or ""
graphemes = graphemes_raw.strip()
audio = _to_float32(getattr(segment, "audio", None))
if audio.size == 0:
continue
local_segments += 1
if chapter_sink:
chapter_sink.write(audio)
if audio_sink:
audio_sink.write(audio)
duration = len(audio) / SAMPLE_RATE
processed_chars += len(graphemes)
job.processed_characters = processed_chars
if job.total_characters:
job.progress = min(processed_chars / job.total_characters, 0.999)
else:
job.progress = 0.0 if processed_chars == 0 else 0.999
preview_text = graphemes or (graphemes_raw[:80] if graphemes_raw else "[silence]")
prefix = f"{preview_prefix} · " if preview_prefix else ""
job.add_log(f"{prefix}{processed_chars:,}/{job.total_characters or '—'}: {preview_text[:80]}")
if subtitle_writer and audio_sink and graphemes:
subtitle_writer.write_segment(
index=subtitle_index,
text=graphemes,
start=current_time,
end=current_time + duration,
)
subtitle_index += 1
if audio_sink:
current_time += duration
try:
for segment in segment_iter:
canceller()
graphemes_raw = getattr(segment, "graphemes", "") or ""
graphemes = graphemes_raw.strip()
audio = _to_float32(getattr(segment, "audio", None))
if audio.size == 0:
continue
local_segments += 1
if chapter_sink:
chapter_sink.write(audio)
if audio_sink:
audio_sink.write(audio)
duration = len(audio) / SAMPLE_RATE
processed_chars += len(graphemes)
job.processed_characters = processed_chars
if job.total_characters:
job.progress = min(processed_chars / job.total_characters, 0.999)
else:
job.progress = 0.0 if processed_chars == 0 else 0.999
preview_text = graphemes or (graphemes_raw[:80] if graphemes_raw else "[silence]")
prefix = f"{preview_prefix} · " if preview_prefix else ""
job.add_log(f"{prefix}{processed_chars:,}/{job.total_characters or '—'}: {preview_text[:80]}")
if subtitle_writer and audio_sink and graphemes:
subtitle_writer.write_segment(
index=subtitle_index,
text=graphemes,
start=current_time,
end=current_time + duration,
)
subtitle_index += 1
if audio_sink:
current_time += duration
except OverflowError as exc:
job.add_log(
f"Skipped chunk — number too large for TTS conversion: {exc}",
level="warning",
)
return local_segments
def append_silence(