Archives/tubearchivist
2
0
Fork 0
mirror of https://github.com/tubearchivist/tubearchivist synced 2024-11-11 19:10:48 +00:00
Code Issues Projects Releases Wiki Activity

fix empty subtitle download and index, #288

Browse Source
This commit is contained in:
simon 2022-08-12 15:53:20 +07:00
parent cc40f46325
commit 6afe25a4af
No known key found for this signature in database
GPG Key ID: 2C15AA5E89985DD4
1 changed files with 15 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
tubearchivist/home/src/index/video.py
View File

@ -119,6 +119,7 @@ class YoutubeSubtitle:
def download_subtitles(self, relevant_subtitles):
"""download subtitle files to archive"""
videos_base = self.video.config["application"]["videos"]
indexed = []
for subtitle in relevant_subtitles:
dest_path = os.path.join(videos_base, subtitle["media_url"])
source = subtitle["source"]
@ -133,12 +134,19 @@ class YoutubeSubtitle:
parser = SubtitleParser(response.text, lang, source)
parser.process()
if not parser.all_cues:
continue
subtitle_str = parser.get_subtitle_str()
self._write_subtitle_file(dest_path, subtitle_str)
if self.video.config["downloads"]["subtitle_index"]:
query_str = parser.create_bulk_import(self.video, source)
self._index_subtitle(query_str)
indexed.append(subtitle)
return indexed
@staticmethod
def _write_subtitle_file(dest_path, subtitle_str):
"""write subtitle file to disk"""
@ -188,11 +196,15 @@ class SubtitleParser:
def process(self):
"""extract relevant que data"""
self.all_cues = []
all_events = self.subtitle_raw.get("events")
if not all_events:
return
if self.source == "auto":
all_events = self._flat_auto_caption(all_events)
self.all_cues = []
for idx, event in enumerate(all_events):
if "dDurationMs" not in event or "segs" not in event:
# some events won't have a duration or segs
@ -649,8 +661,8 @@ class YoutubeVideo(YouTubeItem, YoutubeSubtitle):
handler = YoutubeSubtitle(self)
subtitles = handler.get_subtitles()
if subtitles:
self.json_data["subtitles"] = subtitles
handler.download_subtitles(relevant_subtitles=subtitles)
indexed = handler.download_subtitles(relevant_subtitles=subtitles)
self.json_data["subtitles"] = indexed
def update_media_url(self):
"""update only media_url in es for reindex channel rename"""