dl-proxy / clean_sub.py
jumannendange's picture
Rename clean_sub (1).py to clean_sub.py
4c5584b verified
Raw History Blame Contribute Delete
1.64 kB
import re
_TS = re.compile(r'(\d{1,2}):(\d\d):(\d\d)[,.](\d{1,3})\s*-->\s*(\d{1,2}):(\d\d):(\d\d)[,.](\d{1,3})')
def decode_sub(raw: bytes) -> str:
"""UTF-16 (BOM) -> UTF-8 -> Windows-1252, kwa mpangilio huo."""
if raw[:2] in (b'\xff\xfe', b'\xfe\xff'):
return raw.decode('utf-16', errors='replace')
try:
return raw.decode('utf-8')
except UnicodeDecodeError:
return raw.decode('cp1252', errors='replace')
def _fmt(h, m, s, ms):
return f'{int(h):02d}:{m}:{s},{ms.ljust(3, "0")[:3]}'
def clean_srt(raw: bytes) -> str:
"""Geuza manukuu yoyote (SRT, WEBVTT, mchanganyiko wenye BOM/koma/nukta)
kuwa SRT safi yenye namba. Rudisha '' kama hakuna cue hata moja."""
text = decode_sub(raw).replace('\ufeff', '').replace('\r\n', '\n').replace('\r', '\n')
lines = text.split('\n')
cues, i = [], 0
while i < len(lines):
m = _TS.search(lines[i])
if not m:
i += 1
continue
i += 1
body = []
while i < len(lines) and lines[i].strip():
if _TS.search(lines[i]):
break
if lines[i].strip().isdigit() and i + 1 < len(lines) and _TS.search(lines[i + 1]):
break
body.append(lines[i])
i += 1
txt = re.sub(r'\{\\[^}]*\}', '', '\n'.join(body))
txt = re.sub(r'</?font[^>]*>', '', txt, flags=re.I).strip()
if txt:
g = m.groups()
cues.append(f'{_fmt(*g[0:4])} --> {_fmt(*g[4:8])}\n{txt}')
if not cues:
return ''
return '\n\n'.join(f'{n}\n{c}' for n, c in enumerate(cues, 1)) + '\n'