github/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-21 20:46:36 -05:00
Code Issues Projects Releases Packages Wiki Activity

refactor YouTube subtitles code, it was ugly (my bad)

Browse source
This commit is contained in:
Filippo Valsorda 2012-12-20 11:26:38 +01:00
parent 69a3883199
commit 056d857571
2 changed files with 31 additions and 37 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

60
youtube_dl/InfoExtractors.py
View file

@ -219,6 +219,34 @@ def _closed_captions_xml_to_srt(self, xml_string):
srt += caption + '\n\n' srt += caption + '\n\n'
return srt return srt
def _extract_subtitles(self, video_id):
self.report_video_subtitles_download(video_id)
request = compat_urllib_request.Request('http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id)
try:
srt_list = compat_urllib_request.urlopen(request).read().decode('utf-8')
except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
return (u'WARNING: unable to download video subtitles: %s' % compat_str(err), None)
srt_lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', srt_list)
srt_lang_list = dict((l[1], l[0]) for l in srt_lang_list)
if not srt_lang_list:
return (u'WARNING: video has no closed captions', None)
if self._downloader.params.get('subtitleslang', False):
srt_lang = self._downloader.params.get('subtitleslang')
elif 'en' in srt_lang_list:
srt_lang = 'en'
else:
srt_lang = srt_lang_list.keys()[0]
if not srt_lang in srt_lang_list:
return (u'WARNING: no closed captions found in the specified language', None)
request = compat_urllib_request.Request('http://www.youtube.com/api/timedtext?lang=%s&name=%s&v=%s' % (srt_lang, srt_lang_list[srt_lang], video_id))
try:
srt_xml = compat_urllib_request.urlopen(request).read().decode('utf-8')
except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
return (u'WARNING: unable to download video subtitles: %s' % compat_str(err), None)
if not srt_xml:
return (u'WARNING: unable to download video subtitles', None)
return (None, self._closed_captions_xml_to_srt(srt_xml))
def _print_formats(self, formats): def _print_formats(self, formats):
print('Available formats:') print('Available formats:')
for x in formats: for x in formats:
@ -395,35 +423,9 @@ def _real_extract(self, url):
# closed captions # closed captions
video_subtitles = None video_subtitles = None
if self._downloader.params.get('writesubtitles', False): if self._downloader.params.get('writesubtitles', False):
try: (srt_error, video_subtitles) = self._extract_subtitles(video_id)
self.report_video_subtitles_download(video_id) if srt_error:
request = compat_urllib_request.Request('http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id) self._downloader.trouble(srt_error)
try:
srt_list = compat_urllib_request.urlopen(request).read().decode('utf-8')
except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
raise Trouble(u'WARNING: unable to download video subtitles: %s' % compat_str(err))
srt_lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', srt_list)
srt_lang_list = dict((l[1], l[0]) for l in srt_lang_list)
if not srt_lang_list:
raise Trouble(u'WARNING: video has no closed captions')
if self._downloader.params.get('subtitleslang', False):
srt_lang = self._downloader.params.get('subtitleslang')
elif 'en' in srt_lang_list:
srt_lang = 'en'
else:
srt_lang = srt_lang_list.keys()[0]
if not srt_lang in srt_lang_list:
raise Trouble(u'WARNING: no closed captions found in the specified language')
request = compat_urllib_request.Request('http://www.youtube.com/api/timedtext?lang=%s&name=%s&v=%s' % (srt_lang, srt_lang_list[srt_lang], video_id))
try:
srt_xml = compat_urllib_request.urlopen(request).read().decode('utf-8')
except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
raise Trouble(u'WARNING: unable to download video subtitles: %s' % compat_str(err))
if not srt_xml:
raise Trouble(u'WARNING: unable to download video subtitles')
video_subtitles = self._closed_captions_xml_to_srt(srt_xml)
except Trouble as trouble:
self._downloader.trouble(str(trouble))
if 'length_seconds' not in video_info: if 'length_seconds' not in video_info:
self._downloader.trouble(u'WARNING: unable to extract video duration') self._downloader.trouble(u'WARNING: unable to extract video duration')

8
youtube_dl/utils.py
View file

@ -465,14 +465,6 @@ def __init__(self, downloaded, expected):
self.downloaded = downloaded self.downloaded = downloaded
self.expected = expected self.expected = expected
class Trouble(Exception):
"""Trouble helper exception
This is an exception to be handled with
FileDownloader.trouble
"""
class YoutubeDLHandler(compat_urllib_request.HTTPHandler): class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
"""Handler for HTTP requests and responses. """Handler for HTTP requests and responses.