CDRM-Team/devine
1
0
Fork 0
mirror of https://github.com/devine-dl/devine.git synced 2025-04-29 17:49:44 +00:00
Code Issues Packages Projects Releases Wiki Activity

Ensure input is UTF-8 when parsing TTML and WebVTT Subtitles

Browse Source 
This fixes some conversion errors when working with non-latin languages like Russian (crylic) and Arabic.
This commit is contained in:
rlaphoenix 2024-01-12 00:36:43 +00:00
parent 96f1cbb260
commit 14ebe4ee1b
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
devine/core/tracks/subtitle.py
View File

@ -15,7 +15,7 @@ from pymp4.parser import MP4
from subtitle_filter import Subtitles from subtitle_filter import Subtitles
from devine.core.tracks.track import Track from devine.core.tracks.track import Track
from devine.core.utilities import get_binary_path from devine.core.utilities import get_binary_path, try_ensure_utf8
class Subtitle(Track): class Subtitle(Track):
@ -160,7 +160,7 @@ class Subtitle(Track):
caption_lists[lang].extend(segment.get_captions(lang)) caption_lists[lang].extend(segment.get_captions(lang))
caption_set: pycaption.CaptionSet = pycaption.CaptionSet(caption_lists) caption_set: pycaption.CaptionSet = pycaption.CaptionSet(caption_lists)
elif codec == Subtitle.Codec.TimedTextMarkupLang: elif codec == Subtitle.Codec.TimedTextMarkupLang:
text = data.decode("utf8") text = try_ensure_utf8(data).decode("utf8")
text = text.replace("tt:", "") text = text.replace("tt:", "")
# negative size values aren't allowed in TTML/DFXP spec, replace with 0 # negative size values aren't allowed in TTML/DFXP spec, replace with 0
text = re.sub(r'"(-\d+(\.\d+)?(px|em|%|c|pt))"', '"0"', text) text = re.sub(r'"(-\d+(\.\d+)?(px|em|%|c|pt))"', '"0"', text)
@ -171,7 +171,7 @@ class Subtitle(Track):
caption_lists[language] = caption_list caption_lists[language] = caption_list
caption_set: pycaption.CaptionSet = pycaption.CaptionSet(caption_lists) caption_set: pycaption.CaptionSet = pycaption.CaptionSet(caption_lists)
elif codec == Subtitle.Codec.WebVTT: elif codec == Subtitle.Codec.WebVTT:
text = data.decode("utf8") text = try_ensure_utf8(data).decode("utf8")
# Segmented VTT when merged may have the WEBVTT headers part of the next caption # Segmented VTT when merged may have the WEBVTT headers part of the next caption
# if they are not separated far enough from the previous caption, hence the \n\n # if they are not separated far enough from the previous caption, hence the \n\n
text = text. \ text = text. \