dschlueter/youtube_abschrift_korrigieren
Watch
1
0
Fork
You've already forked youtube_abschrift_korrigieren
 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

fix: glossar_vorschlag.py schickt keine Zeit-Überschriften ans Modell

Browse source 
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Dieter Schlüter 2026-10-11 20:16:21 +02:00
commit 5e9a11d0f4
1 changed files with 35 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

10
glossar_vorschlag.py
View file

@ -15,6 +15,7 @@ from rede_korrigieren import (
MODEL, apply_replacements, chat_request, check_ollama, ollama_open, read_glossary,
split_chunks,
)
from vtt2txt import split_segments
SYSTEM = """Du suchst Hörfehler in der automatischen Abschrift einer deutschen Rede.
@ -116,6 +117,11 @@ def model_suggestions(text, model, cache_dir, known):
return found
def plain_text(text):
"""Running text without time headers ("[00:04:57 - 00:06:04]") of excerpts."""
return " ".join(" ".join(body for _, body in split_segments(text)).split())
def contains(words, part):
"""Is `part` a contiguous run of `words`?"""
return any(words[i:i + len(part)] == part for i in range(len(words) - len(part) + 1))
@ -213,7 +219,7 @@ def parse_args():
def main():
args = parse_args()
text = " ".join(read_raw(args.input).split())
text = plain_text(read_raw(args.input))
if not text:
raise SystemExit("Die Eingabedatei ist leer.")
@ -234,7 +240,7 @@ def main():
changes = None
if args.korrigiert:
corrected = args.korrigiert.read_text(encoding="utf-8")
corrected = plain_text(args.korrigiert.read_text(encoding="utf-8"))
changes = correction_changes(raw, corrected, replacements, set(suggestions))
args.output.parent.mkdir(parents=True, exist_ok=True)