Don't use bare 'except:'
[youtube-dl] / youtube_dl / YoutubeDL.py
index ca7c3f5c6cc1e2fdb0669f2bfb22146510ff7865..640b8c99d75d6215b232c0d2d678dd903f41e8bb 100755 (executable)
@@ -4,8 +4,10 @@
 from __future__ import absolute_import, unicode_literals
 
 import collections
+import contextlib
 import datetime
 import errno
+import fileinput
 import io
 import itertools
 import json
@@ -28,6 +30,7 @@ from .compat import (
     compat_basestring,
     compat_cookiejar,
     compat_expanduser,
+    compat_get_terminal_size,
     compat_http_client,
     compat_kwargs,
     compat_str,
@@ -46,18 +49,19 @@ from .utils import (
     ExtractorError,
     format_bytes,
     formatSeconds,
-    get_term_width,
     locked_file,
     make_HTTPS_handler,
     MaxDownloadsReached,
     PagedList,
     parse_filesize,
+    PerRequestProxyHandler,
     PostProcessingError,
     platform_name,
     preferredencoding,
     render_table,
     SameFileError,
     sanitize_filename,
+    sanitize_path,
     std_headers,
     subtitles_filename,
     takewhile_inclusive,
@@ -154,7 +158,7 @@ class YoutubeDL(object):
     allsubtitles:      Downloads all the subtitles of the video
                        (requires writesubtitles or writeautomaticsub)
     listsubtitles:     Lists all available subtitles for the video
-    subtitlesformat:   Subtitle format [srt/sbv/vtt] (default=srt)
+    subtitlesformat:   The format code for subtitles
     subtitleslangs:    List of languages of the subtitles to download
     keepvideo:         Keep the video file after post-processing
     daterange:         A DateRange object, download only if the upload_date is in the range.
@@ -181,6 +185,8 @@ class YoutubeDL(object):
     prefer_insecure:   Use HTTP instead of HTTPS to retrieve information.
                        At the moment, this is only supported by YouTube.
     proxy:             URL of the proxy server to use
+    cn_verification_proxy:  URL of the proxy to use for IP address verification
+                       on Chinese sites. (Experimental)
     socket_timeout:    Time to wait for unresponsive hosts, in seconds
     bidi_workaround:   Work around buggy terminals without bidirectional text
                        support, using fridibi
@@ -247,10 +253,10 @@ class YoutubeDL(object):
     hls_prefer_native: Use the native HLS downloader instead of ffmpeg/avconv.
 
     The following parameters are not used by YoutubeDL itself, they are used by
-    the FileDownloader:
+    the downloader (see youtube_dl/downloader/common.py):
     nopart, updatetime, buffersize, ratelimit, min_filesize, max_filesize, test,
     noresizebuffer, retries, continuedl, noprogress, consoletitle,
-    xattr_set_filesize.
+    xattr_set_filesize, external_downloader_args.
 
     The following options are used by the post processors:
     prefer_ffmpeg:     If True, use ffmpeg instead of avconv if both are available,
@@ -284,7 +290,7 @@ class YoutubeDL(object):
             try:
                 import pty
                 master, slave = pty.openpty()
-                width = get_term_width()
+                width = compat_get_terminal_size().columns
                 if width is None:
                     width_args = []
                 else:
@@ -317,8 +323,10 @@ class YoutubeDL(object):
                 'Set the LC_ALL environment variable to fix this.')
             self.params['restrictfilenames'] = True
 
-        if '%(stitle)s' in self.params.get('outtmpl', ''):
-            self.report_warning('%(stitle)s is deprecated. Use the %(title)s and the --restrict-filenames flag(which also secures %(uploader)s et al) instead.')
+        if isinstance(params.get('outtmpl'), bytes):
+            self.report_warning(
+                'Parameter outtmpl is bytes, but should be a unicode string. '
+                'Put  from __future__ import unicode_literals  at the top of your code file or consider switching to Python 3.x.')
 
         self._setup_opener()
 
@@ -557,7 +565,7 @@ class YoutubeDL(object):
                                  if v is not None)
             template_dict = collections.defaultdict(lambda: 'NA', template_dict)
 
-            outtmpl = self.params.get('outtmpl', DEFAULT_OUTTMPL)
+            outtmpl = sanitize_path(self.params.get('outtmpl', DEFAULT_OUTTMPL))
             tmpl = compat_expanduser(outtmpl)
             filename = tmpl % template_dict
             # Temporary fix for #4787
@@ -624,7 +632,7 @@ class YoutubeDL(object):
         Returns a list with a dictionary for each video we find.
         If 'download', also downloads the videos.
         extra_info is a dict containing the extra values to add to each result
-         '''
+        '''
 
         if ie_key:
             ies = [self.get_info_extractor(ie_key)]
@@ -1008,6 +1016,15 @@ class YoutubeDL(object):
                 info_dict['timestamp'])
             info_dict['upload_date'] = upload_date.strftime('%Y%m%d')
 
+        if self.params.get('listsubtitles', False):
+            if 'automatic_captions' in info_dict:
+                self.list_subtitles(info_dict['id'], info_dict.get('automatic_captions'), 'automatic captions')
+            self.list_subtitles(info_dict['id'], info_dict.get('subtitles'), 'subtitles')
+            return
+        info_dict['requested_subtitles'] = self.process_subtitles(
+            info_dict['id'], info_dict.get('subtitles'),
+            info_dict.get('automatic_captions'))
+
         # This extractors handle format selection themselves
         if info_dict['extractor'] in ['Youku']:
             if download:
@@ -1071,8 +1088,7 @@ class YoutubeDL(object):
         if req_format is None:
             req_format = 'best'
         formats_to_download = []
-        # The -1 is for supporting YoutubeIE
-        if req_format in ('-1', 'all'):
+        if req_format == 'all':
             formats_to_download = formats
         else:
             for rfstr in req_format.split(','):
@@ -1136,6 +1152,55 @@ class YoutubeDL(object):
         info_dict.update(formats_to_download[-1])
         return info_dict
 
+    def process_subtitles(self, video_id, normal_subtitles, automatic_captions):
+        """Select the requested subtitles and their format"""
+        available_subs = {}
+        if normal_subtitles and self.params.get('writesubtitles'):
+            available_subs.update(normal_subtitles)
+        if automatic_captions and self.params.get('writeautomaticsub'):
+            for lang, cap_info in automatic_captions.items():
+                if lang not in available_subs:
+                    available_subs[lang] = cap_info
+
+        if (not self.params.get('writesubtitles') and not
+                self.params.get('writeautomaticsub') or not
+                available_subs):
+            return None
+
+        if self.params.get('allsubtitles', False):
+            requested_langs = available_subs.keys()
+        else:
+            if self.params.get('subtitleslangs', False):
+                requested_langs = self.params.get('subtitleslangs')
+            elif 'en' in available_subs:
+                requested_langs = ['en']
+            else:
+                requested_langs = [list(available_subs.keys())[0]]
+
+        formats_query = self.params.get('subtitlesformat', 'best')
+        formats_preference = formats_query.split('/') if formats_query else []
+        subs = {}
+        for lang in requested_langs:
+            formats = available_subs.get(lang)
+            if formats is None:
+                self.report_warning('%s subtitles not available for %s' % (lang, video_id))
+                continue
+            for ext in formats_preference:
+                if ext == 'best':
+                    f = formats[-1]
+                    break
+                matches = list(filter(lambda f: f['ext'] == ext, formats))
+                if matches:
+                    f = matches[-1]
+                    break
+            else:
+                f = formats[-1]
+                self.report_warning(
+                    'No subtitle format found matching "%s" for language %s, '
+                    'using %s' % (formats_query, lang, f['ext']))
+            subs[lang] = f
+        return subs
+
     def process_info(self, info_dict):
         """Process a single resolved IE result."""
 
@@ -1150,9 +1215,6 @@ class YoutubeDL(object):
         if len(info_dict['title']) > 200:
             info_dict['title'] = info_dict['title'][:197] + '...'
 
-        # Keep for backwards compatibility
-        info_dict['stitle'] = info_dict['title']
-
         if 'format' not in info_dict:
             info_dict['format'] = info_dict['ext']
 
@@ -1198,7 +1260,7 @@ class YoutubeDL(object):
             return
 
         try:
-            dn = os.path.dirname(encodeFilename(filename))
+            dn = os.path.dirname(sanitize_path(encodeFilename(filename)))
             if dn and not os.path.exists(dn):
                 os.makedirs(dn)
         except (OSError, IOError) as err:
@@ -1238,15 +1300,23 @@ class YoutubeDL(object):
         subtitles_are_requested = any([self.params.get('writesubtitles', False),
                                        self.params.get('writeautomaticsub')])
 
-        if subtitles_are_requested and 'subtitles' in info_dict and info_dict['subtitles']:
+        if subtitles_are_requested and info_dict.get('requested_subtitles'):
             # subtitles download errors are already managed as troubles in relevant IE
             # that way it will silently go on when used with unsupporting IE
-            subtitles = info_dict['subtitles']
-            sub_format = self.params.get('subtitlesformat', 'srt')
-            for sub_lang in subtitles.keys():
-                sub = subtitles[sub_lang]
-                if sub is None:
-                    continue
+            subtitles = info_dict['requested_subtitles']
+            ie = self.get_info_extractor(info_dict['extractor_key'])
+            for sub_lang, sub_info in subtitles.items():
+                sub_format = sub_info['ext']
+                if sub_info.get('data') is not None:
+                    sub_data = sub_info['data']
+                else:
+                    try:
+                        sub_data = ie._download_webpage(
+                            sub_info['url'], info_dict['id'], note=False)
+                    except ExtractorError as err:
+                        self.report_warning('Unable to download subtitle for "%s": %s' %
+                                            (sub_lang, compat_str(err.cause)))
+                        continue
                 try:
                     sub_filename = subtitles_filename(filename, sub_lang, sub_format)
                     if self.params.get('nooverwrites', False) and os.path.exists(encodeFilename(sub_filename)):
@@ -1254,7 +1324,7 @@ class YoutubeDL(object):
                     else:
                         self.to_screen('[info] Writing video subtitles to: ' + sub_filename)
                         with io.open(encodeFilename(sub_filename), 'w', encoding='utf-8') as subfile:
-                            subfile.write(sub)
+                            subfile.write(sub_data)
                 except (OSError, IOError):
                     self.report_error('Cannot write subtitles file ' + sub_filename)
                     return
@@ -1386,8 +1456,11 @@ class YoutubeDL(object):
         return self._download_retcode
 
     def download_with_info_file(self, info_filename):
-        with io.open(info_filename, 'r', encoding='utf-8') as f:
-            info = json.load(f)
+        with contextlib.closing(fileinput.FileInput(
+                [info_filename], mode='r',
+                openhook=fileinput.hook_encoded('utf-8'))) as f:
+            # FileInput doesn't have a read method, we can't call json.load
+            info = json.loads('\n'.join(f))
         try:
             self.process_ie_result(info, download=True)
         except DownloadError:
@@ -1564,6 +1637,17 @@ class YoutubeDL(object):
             ['ID', 'width', 'height', 'URL'],
             [[t['id'], t.get('width', 'unknown'), t.get('height', 'unknown'), t['url']] for t in thumbnails]))
 
+    def list_subtitles(self, video_id, subtitles, name='subtitles'):
+        if not subtitles:
+            self.to_screen('%s has no %s' % (video_id, name))
+            return
+        self.to_screen(
+            'Available %s for %s:' % (name, video_id))
+        self.to_screen(render_table(
+            ['Language', 'formats'],
+            [[lang, ', '.join(f['ext'] for f in reversed(formats))]
+                for lang, formats in subtitles.items()]))
+
     def urlopen(self, req):
         """ Start an HTTP download """
 
@@ -1617,10 +1701,10 @@ class YoutubeDL(object):
             out = out.decode().strip()
             if re.match('[0-9a-f]+', out):
                 self._write_string('[debug] Git HEAD: ' + out + '\n')
-        except:
+        except Exception:
             try:
                 sys.exc_clear()
-            except:
+            except Exception:
                 pass
         self._write_string('[debug] Python version %s - %s\n' % (
             platform.python_version(), platform_name()))
@@ -1680,13 +1764,14 @@ class YoutubeDL(object):
             # Set HTTPS proxy to HTTP one if given (https://github.com/rg3/youtube-dl/issues/805)
             if 'http' in proxies and 'https' not in proxies:
                 proxies['https'] = proxies['http']
-        proxy_handler = compat_urllib_request.ProxyHandler(proxies)
+        proxy_handler = PerRequestProxyHandler(proxies)
 
         debuglevel = 1 if self.params.get('debug_printtraffic') else 0
         https_handler = make_HTTPS_handler(self.params, debuglevel=debuglevel)
         ydlh = YoutubeDLHandler(self.params, debuglevel=debuglevel)
         opener = compat_urllib_request.build_opener(
-            https_handler, proxy_handler, cookie_processor, ydlh)
+            proxy_handler, https_handler, cookie_processor, ydlh)
+
         # Delete the default user-agent header, which would otherwise apply in
         # cases where our custom HTTP handler doesn't come into play
         # (See https://github.com/rg3/youtube-dl/issues/1309 for details)