]> jfr.im git - yt-dlp.git/blobdiff - yt_dlp/YoutubeDL.py
[cleanup] Misc fixes (see desc)
[yt-dlp.git] / yt_dlp / YoutubeDL.py
index b1753241def99f120f8c6dfcc2886cdaf91e3200..b8c250d736668da5d18bc67e7b271f9e02423e42 100644 (file)
@@ -27,6 +27,7 @@
 
 from .cache import Cache
 from .compat import (
+    HAS_LEGACY as compat_has_legacy,
     compat_get_terminal_size,
     compat_os_name,
     compat_shlex_quote,
@@ -194,13 +195,6 @@ class YoutubeDL:
                        For compatibility, a single list is also accepted
     print_to_file:     A dict with keys WHEN (same as forceprint) mapped to
                        a list of tuples with (template, filename)
-    forceurl:          Force printing final URL. (Deprecated)
-    forcetitle:        Force printing title. (Deprecated)
-    forceid:           Force printing ID. (Deprecated)
-    forcethumbnail:    Force printing thumbnail URL. (Deprecated)
-    forcedescription:  Force printing description. (Deprecated)
-    forcefilename:     Force printing final filename. (Deprecated)
-    forceduration:     Force printing duration. (Deprecated)
     forcejson:         Force printing info_dict as JSON.
     dump_single_json:  Force printing the info_dict of the whole playlist
                        (or video) as a single JSON line.
@@ -277,9 +271,6 @@ class YoutubeDL:
     writedesktoplink:  Write a Linux internet shortcut file (.desktop)
     writesubtitles:    Write the video subtitles to a file
     writeautomaticsub: Write the automatically generated subtitles to a file
-    allsubtitles:      Deprecated - Use subtitleslangs = ['all']
-                       Downloads all the subtitles of the video
-                       (requires writesubtitles or writeautomaticsub)
     listsubtitles:     Lists all available subtitles for the video
     subtitlesformat:   The format code for subtitles
     subtitleslangs:    List of languages of the subtitles to download (can be regex).
@@ -333,7 +324,6 @@ class YoutubeDL:
     bidi_workaround:   Work around buggy terminals without bidirectional text
                        support, using fridibi
     debug_printtraffic:Print out sent and received HTTP traffic
-    include_ads:       Download ads as well (deprecated)
     default_search:    Prepend this string if an input url is not valid.
                        'auto' for elaborate guessing
     encoding:          Use this encoding instead of the system-specified.
@@ -349,10 +339,6 @@ class YoutubeDL:
                        * when: When to run the postprocessor. Allowed values are
                                the entries of utils.POSTPROCESS_WHEN
                                Assumed to be 'post_process' if not given
-    post_hooks:        Deprecated - Register a custom postprocessor instead
-                       A list of functions that get called as the final step
-                       for each video file, after all postprocessors have been
-                       called. The filename will be passed as the only argument.
     progress_hooks:    A list of functions that get called on download
                        progress, with a dictionary with the entries
                        * status: One of "downloading", "error", or "finished".
@@ -397,8 +383,6 @@ class YoutubeDL:
                        - "detect_or_warn": check whether we can do anything
                                            about it, warn otherwise (default)
     source_address:    Client-side IP address to bind to.
-    call_home:         Boolean, true iff we are allowed to contact the
-                       yt-dlp servers for debugging. (BROKEN)
     sleep_interval_requests: Number of seconds to sleep between requests
                        during extraction
     sleep_interval:    Number of seconds to sleep before each download when
@@ -433,17 +417,10 @@ class YoutubeDL:
     geo_bypass_ip_block:
                        IP range in CIDR notation that will be used similarly to
                        geo_bypass_country
-
-    The following options determine which downloader is picked:
     external_downloader: A dictionary of protocol keys and the executable of the
                        external downloader to use for it. The allowed protocols
                        are default|http|ftp|m3u8|dash|rtsp|rtmp|mms.
                        Set the value to 'native' to use the native downloader
-    hls_prefer_native: Deprecated - Use external_downloader = {'m3u8': 'native'}
-                       or {'m3u8': 'ffmpeg'} instead.
-                       Use the native HLS downloader instead of ffmpeg/avconv
-                       if True, otherwise use ffmpeg/avconv if False, otherwise
-                       use downloader suggested by extractor if None.
     compat_opts:       Compatibility options. See "Differences in default behavior".
                        The following options do not work when used through the API:
                        filename, abort-on-error, multistreams, no-live-chat, format-sort
@@ -456,6 +433,13 @@ class YoutubeDL:
     retry_sleep_functions: Dictionary of functions that takes the number of attempts
                        as argument and returns the time to sleep in seconds.
                        Allowed keys are 'http', 'fragment', 'file_access'
+    download_ranges:   A function that gets called for every video with the signature
+                       (info_dict, *, ydl) -> Iterable[Section].
+                       Only the returned sections will be downloaded. Each Section contains:
+                       * start_time: Start time of the section in seconds
+                       * end_time: End time of the section in seconds
+                       * title: Section title (Optional)
+                       * index: Section number (Optional)
 
     The following parameters are not used by YoutubeDL itself, they are used by
     the downloader (see yt_dlp/downloader/common.py):
@@ -465,8 +449,6 @@ class YoutubeDL:
     external_downloader_args, concurrent_fragment_downloads.
 
     The following options are used by the post processors:
-    prefer_ffmpeg:     If False, use avconv instead of ffmpeg if both are available,
-                       otherwise prefer ffmpeg. (avconv support is deprecated)
     ffmpeg_location:   Location of the ffmpeg/avconv binary; either the path
                        to the binary or its containing directory.
     postprocessor_args: A dictionary of postprocessor/executable keys (in lower case)
@@ -486,12 +468,48 @@ class YoutubeDL:
                        See "EXTRACTOR ARGUMENTS" for details.
                        Eg: {'youtube': {'skip': ['dash', 'hls']}}
     mark_watched:      Mark videos watched (even with --simulate). Only for YouTube
-    youtube_include_dash_manifest: Deprecated - Use extractor_args instead.
+
+    The following options are deprecated and may be removed in the future:
+
+    forceurl:          - Use forceprint
+                       Force printing final URL.
+    forcetitle:        - Use forceprint
+                       Force printing title.
+    forceid:           - Use forceprint
+                       Force printing ID.
+    forcethumbnail:    - Use forceprint
+                       Force printing thumbnail URL.
+    forcedescription:  - Use forceprint
+                       Force printing description.
+    forcefilename:     - Use forceprint
+                       Force printing final filename.
+    forceduration:     - Use forceprint
+                       Force printing duration.
+    allsubtitles:      - Use subtitleslangs = ['all']
+                       Downloads all the subtitles of the video
+                       (requires writesubtitles or writeautomaticsub)
+    include_ads:       - Doesn't work
+                       Download ads as well
+    call_home:         - Not implemented
+                       Boolean, true iff we are allowed to contact the
+                       yt-dlp servers for debugging.
+    post_hooks:        - Register a custom postprocessor
+                       A list of functions that get called as the final step
+                       for each video file, after all postprocessors have been
+                       called. The filename will be passed as the only argument.
+    hls_prefer_native: - Use external_downloader = {'m3u8': 'native'} or {'m3u8': 'ffmpeg'}.
+                       Use the native HLS downloader instead of ffmpeg/avconv
+                       if True, otherwise use ffmpeg/avconv if False, otherwise
+                       use downloader suggested by extractor if None.
+    prefer_ffmpeg:     - avconv support is deprecated
+                       If False, use avconv instead of ffmpeg if both are available,
+                       otherwise prefer ffmpeg.
+    youtube_include_dash_manifest: - Use extractor_args
                        If True (default), DASH manifests and related
                        data will be downloaded and processed by extractor.
                        You can reduce network I/O by disabling it if you don't
                        care about DASH. (only for youtube)
-    youtube_include_hls_manifest: Deprecated - Use extractor_args instead.
+    youtube_include_hls_manifest: - Use extractor_args
                        If True (default), HLS manifests and related
                        data will be downloaded and processed by extractor.
                        You can reduce network I/O by disabling it if you don't
@@ -558,7 +576,7 @@ def __init__(self, params=None, auto_init=True):
         )
         self._allow_colors = Namespace(**{
             type_: not self.params.get('no_color') and supports_terminal_sequences(stream)
-            for type_, stream in self._out_files if type_ != 'console'
+            for type_, stream in self._out_files.items_ if type_ != 'console'
         })
 
         if sys.version_info < (3, 6):
@@ -591,7 +609,10 @@ def check_deprecated(param, option, suggestion):
         for msg in self.params.get('_deprecation_warnings', []):
             self.deprecation_warning(msg)
 
-        if 'list-formats' in self.params.get('compat_opts', []):
+        self.params['compat_opts'] = set(self.params.get('compat_opts', ()))
+        if not compat_has_legacy:
+            self.params['compat_opts'].add('no-compat-legacy')
+        if 'list-formats' in self.params['compat_opts']:
             self.params['listformats_table'] = False
 
         if 'overwrites' not in self.params and self.params.get('nooverwrites') is not None:
@@ -788,9 +809,9 @@ def to_stdout(self, message, skip_eol=False, quiet=None):
         """Print message to stdout"""
         if quiet is not None:
             self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument quiet. Use "YoutubeDL.to_screen" instead')
-        self._write_string(
-            '%s%s' % (self._bidi_workaround(message), ('' if skip_eol else '\n')),
-            self._out_files.out)
+        if skip_eol is not False:
+            self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument skip_eol. Use "YoutubeDL.to_screen" instead')
+        self._write_string(f'{self._bidi_workaround(message)}\n', self._out_files.out)
 
     def to_screen(self, message, skip_eol=False, quiet=None):
         """Print message to screen if not in quiet mode"""
@@ -942,7 +963,7 @@ def write_debug(self, message, only_once=False):
         '''Log debug message or Print message to stderr'''
         if not self.params.get('verbose', False):
             return
-        message = '[debug] %s' % message
+        message = f'[debug] {message}'
         if self.params.get('logger'):
             self.params['logger'].debug(message)
         else:
@@ -1136,7 +1157,7 @@ def get_value(mdict):
         def filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames')):
             return sanitize_filename(str(value), restricted=restricted, is_id=(
                 bool(re.search(r'(^|[_.])id(\.|$)', key))
-                if 'filename-sanitization' in self.params.get('compat_opts', [])
+                if 'filename-sanitization' in self.params['compat_opts']
                 else NO_DEFAULT))
 
         sanitizer = sanitize if callable(sanitize) else filename_sanitizer
@@ -1775,7 +1796,7 @@ def get_entry(i):
         max_failures = self.params.get('skip_playlist_after_errors') or float('inf')
         for i, entry_tuple in enumerate(entries, 1):
             playlist_index, entry = entry_tuple
-            if 'playlist-index' in self.params.get('compat_opts', []):
+            if 'playlist-index' in self.params['compat_opts']:
                 playlist_index = playlistitems[i - 1] if playlistitems else i + playliststart - 1
             self.to_screen('[download] Downloading video %s of %s' % (
                 self._format_screen(i, self.Styles.ID), self._format_screen(n_entries, self.Styles.EMPHASIS)))
@@ -1906,7 +1927,7 @@ def _check_formats(self, formats):
             temp_file.close()
             try:
                 success, _ = self.dl(temp_file.name, f, test=True)
-            except (DownloadError, IOError, OSError, ValueError) + network_exceptions:
+            except (DownloadError, OSError, ValueError) + network_exceptions:
                 success = False
             finally:
                 if os.path.exists(temp_file.name):
@@ -1935,7 +1956,7 @@ def can_merge():
         compat = (
             prefer_best
             or self.params.get('allow_multiple_audio_streams', False)
-            or 'format-spec' in self.params.get('compat_opts', []))
+            or 'format-spec' in self.params['compat_opts'])
 
         return (
             'best/bestvideo+bestaudio' if prefer_best
@@ -2549,7 +2570,7 @@ def is_wellformed(f):
                 format['dynamic_range'] = 'SDR'
             if (info_dict.get('duration') and format.get('tbr')
                     and not format.get('filesize') and not format.get('filesize_approx')):
-                format['filesize_approx'] = info_dict['duration'] * format['tbr'] * (1024 / 8)
+                format['filesize_approx'] = int(info_dict['duration'] * format['tbr'] * (1024 / 8))
 
             # Add HTTP headers, so that external programs can use them from the
             # json output
@@ -2637,16 +2658,34 @@ def is_wellformed(f):
             # Process what we can, even without any available formats.
             formats_to_download = [{}]
 
-        best_format = formats_to_download[-1]
+        requested_ranges = self.params.get('download_ranges')
+        if requested_ranges:
+            requested_ranges = tuple(requested_ranges(info_dict, self))
+
+        best_format, downloaded_formats = formats_to_download[-1], []
         if download:
             if best_format:
-                self.to_screen(
-                    f'[info] {info_dict["id"]}: Downloading {len(formats_to_download)} format(s): '
-                    + ', '.join([f['format_id'] for f in formats_to_download]))
+                def to_screen(*msg):
+                    self.to_screen(f'[info] {info_dict["id"]}: {" ".join(", ".join(variadic(m)) for m in msg)}')
+
+                to_screen(f'Downloading {len(formats_to_download)} format(s):',
+                          (f['format_id'] for f in formats_to_download))
+                if requested_ranges:
+                    to_screen(f'Downloading {len(requested_ranges)} time ranges:',
+                              (f'{int(c["start_time"])}-{int(c["end_time"])}' for c in requested_ranges))
             max_downloads_reached = False
-            for i, fmt in enumerate(formats_to_download):
-                formats_to_download[i] = new_info = self._copy_infodict(info_dict)
+
+            for fmt, chapter in itertools.product(formats_to_download, requested_ranges or [{}]):
+                new_info = self._copy_infodict(info_dict)
                 new_info.update(fmt)
+                if chapter:
+                    new_info.update({
+                        'section_start': chapter.get('start_time'),
+                        'section_end': chapter.get('end_time', 0),
+                        'section_title': chapter.get('title'),
+                        'section_number': chapter.get('index'),
+                    })
+                downloaded_formats.append(new_info)
                 try:
                     self.process_info(new_info)
                 except MaxDownloadsReached:
@@ -2659,12 +2698,12 @@ def is_wellformed(f):
                 if max_downloads_reached:
                     break
 
-            write_archive = {f.get('__write_download_archive', False) for f in formats_to_download}
+            write_archive = {f.get('__write_download_archive', False) for f in downloaded_formats}
             assert write_archive.issubset({True, False, 'ignore'})
             if True in write_archive and False not in write_archive:
                 self.record_download_archive(info_dict)
 
-            info_dict['requested_downloads'] = formats_to_download
+            info_dict['requested_downloads'] = downloaded_formats
             info_dict = self.run_all_pps('after_video', info_dict)
             if max_downloads_reached:
                 raise MaxDownloadsReached()
@@ -3020,6 +3059,16 @@ def existing_video_file(*filepaths):
                     return file
 
                 success = True
+                merger, fd = FFmpegMergerPP(self), None
+                if info_dict.get('url'):
+                    fd = get_suitable_downloader(info_dict, self.params, to_stdout=temp_filename == '-')
+                    if fd is not FFmpegFD and (
+                            info_dict.get('section_start') or info_dict.get('section_end')):
+                        msg = ('This format cannot be partially downloaded' if merger.available
+                               else 'You have requested downloading the video partially, but ffmpeg is not installed')
+                        self.report_error(f'{msg}. Aborting')
+                        return
+
                 if info_dict.get('requested_formats') is not None:
 
                     def compatible_formats(formats):
@@ -3075,9 +3124,6 @@ def correct_ext(filename, ext=new_ext):
                     info_dict['__real_download'] = False
 
                     downloaded = []
-                    merger = FFmpegMergerPP(self)
-
-                    fd = get_suitable_downloader(info_dict, self.params, to_stdout=temp_filename == '-')
                     if dl_filename is not None:
                         self.report_file_already_downloaded(dl_filename)
                     elif fd:
@@ -3625,7 +3671,7 @@ def get_encoding(stream):
             sys.getfilesystemencoding(),
             self.get_encoding(),
             ', '.join(
-                f'{key} {get_encoding(stream)}' for key, stream in self._out_files
+                f'{key} {get_encoding(stream)}' for key, stream in self._out_files.items_
                 if stream is not None and key != 'console')
         )
 
@@ -3652,8 +3698,8 @@ def get_encoding(stream):
             write_debug('Plugins: %s' % [
                 '%s%s' % (klass.__name__, '' if klass.__name__ == name else f' as {name}')
                 for name, klass in itertools.chain(plugin_extractors.items(), plugin_postprocessors.items())])
-        if self.params.get('compat_opts'):
-            write_debug('Compatibility options: %s' % ', '.join(self.params.get('compat_opts')))
+        if self.params['compat_opts']:
+            write_debug('Compatibility options: %s' % ', '.join(self.params['compat_opts']))
 
         if source == 'source':
             try: