Merge remote-tracking branch 'jaimemf/format_spec_groups' (closes #6124)

This commit is contained in:
Jaime Marquínez Ferrándiz 2015-08-03 15:22:51 +02:00
commit 03950c90f7
4 changed files with 323 additions and 127 deletions

View file

@ -289,7 +289,7 @@ youtube-dl_test_video_.mp4 # A simple file name
By default youtube-dl tries to download the best quality, but sometimes you may want to download other format.
The simplest case is requesting a specific format, for example `-f 22`. You can get the list of available formats using `--list-formats`, you can also use a file extension (currently it supports aac, m4a, mp3, mp4, ogg, wav, webm) or the special names `best`, `bestvideo`, `bestaudio` and `worst`.
If you want to download multiple videos and they don't have the same formats available, you can specify the order of preference using slashes, as in `-f 22/17/18`. You can also filter the video results by putting a condition in brackets, as in `-f "best[height=720]"` (or `-f "[filesize>10M]"`). This works for filesize, height, width, tbr, abr, vbr, asr, and fps and the comparisons <, <=, >, >=, =, != and for ext, acodec, vcodec, container, and protocol and the comparisons =, != . Formats for which the value is not known are excluded unless you put a question mark (?) after the operator. You can combine format filters, so `-f "[height <=? 720][tbr>500]"` selects up to 720p videos (or videos where the height is not known) with a bitrate of at least 500 KBit/s. Use commas to download multiple formats, such as `-f 136/137/mp4/bestvideo,140/m4a/bestaudio`. You can merge the video and audio of two formats into a single file using `-f <video-format>+<audio-format>` (requires ffmpeg or avconv), for example `-f bestvideo+bestaudio`.
If you want to download multiple videos and they don't have the same formats available, you can specify the order of preference using slashes, as in `-f 22/17/18`. You can also filter the video results by putting a condition in brackets, as in `-f "best[height=720]"` (or `-f "[filesize>10M]"`). This works for filesize, height, width, tbr, abr, vbr, asr, and fps and the comparisons <, <=, >, >=, =, != and for ext, acodec, vcodec, container, and protocol and the comparisons =, != . Formats for which the value is not known are excluded unless you put a question mark (?) after the operator. You can combine format filters, so `-f "[height <=? 720][tbr>500]"` selects up to 720p videos (or videos where the height is not known) with a bitrate of at least 500 KBit/s. Use commas to download multiple formats, such as `-f 136/137/mp4/bestvideo,140/m4a/bestaudio`. You can merge the video and audio of two formats into a single file using `-f <video-format>+<audio-format>` (requires ffmpeg or avconv), for example `-f bestvideo+bestaudio`. Format selectors can also be grouped using parentheses, for example if you want to download the best mp4 and webm formats with a height lower than 480 you can use `-f '(mp4,webm)[height<480]'`.
Since the end of April 2015 and version 2015.04.26 youtube-dl uses `-f bestvideo+bestaudio/best` as default format selection (see #5447, #5456). If ffmpeg or avconv are installed this results in downloading `bestvideo` and `bestaudio` separately and muxing them together into a single file giving the best overall quality available. Otherwise it falls back to `best` and results in downloading best available quality served as a single file. `best` is also needed for videos that don't come from YouTube because they don't provide the audio and video in two different files. If you want to only download some dash formats (for example if you are not interested in getting videos with a resolution higher than 1080p), you can add `-f bestvideo[height<=?1080]+bestaudio/best` to your configuration file. Note that if you use youtube-dl to stream to `stdout` (and most likely to pipe it to your media player then), i.e. you explicitly specify output template as `-o -`, youtube-dl still uses `-f best` format selection in order to start content delivery immediately to your player and not to wait until `bestvideo` and `bestaudio` are downloaded and muxed.

View file

@ -15,7 +15,7 @@ from youtube_dl import YoutubeDL
from youtube_dl.compat import compat_str
from youtube_dl.extractor import YoutubeIE
from youtube_dl.postprocessor.common import PostProcessor
from youtube_dl.utils import match_filter_func
from youtube_dl.utils import ExtractorError, match_filter_func
TEST_URL = 'http://localhost/sample.mp4'
@ -229,21 +229,70 @@ class TestFormatSelection(unittest.TestCase):
'141', '172', '140', '171', '139',
]
for f1id, f2id in zip(order, order[1:]):
f1 = YoutubeIE._formats[f1id].copy()
f1['format_id'] = f1id
f1['url'] = 'url:' + f1id
f2 = YoutubeIE._formats[f2id].copy()
f2['format_id'] = f2id
f2['url'] = 'url:' + f2id
def format_info(f_id):
info = YoutubeIE._formats[f_id].copy()
info['format_id'] = f_id
info['url'] = 'url:' + f_id
return info
formats_order = [format_info(f_id) for f_id in order]
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': 'bestvideo+bestaudio'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded = ydl.downloaded_info_dicts[0]
self.assertEqual(downloaded['format_id'], '137+141')
self.assertEqual(downloaded['ext'], 'mp4')
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': 'bestvideo[height>=999999]+bestaudio/best'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded = ydl.downloaded_info_dicts[0]
self.assertEqual(downloaded['format_id'], '38')
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': 'bestvideo/best,bestaudio'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
self.assertEqual(downloaded_ids, ['137', '141'])
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': '(bestvideo[ext=mp4],bestvideo[ext=webm])+bestaudio'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
self.assertEqual(downloaded_ids, ['137+141', '248+141'])
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': '(bestvideo[ext=mp4],bestvideo[ext=webm])[height<=720]+bestaudio'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
self.assertEqual(downloaded_ids, ['136+141', '247+141'])
info_dict = _make_result(list(formats_order), extractor='youtube')
ydl = YDL({'format': '(bestvideo[ext=none]/bestvideo[ext=webm])+bestaudio'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
self.assertEqual(downloaded_ids, ['248+141'])
for f1, f2 in zip(formats_order, formats_order[1:]):
info_dict = _make_result([f1, f2], extractor='youtube')
ydl = YDL({'format': 'best/bestvideo'})
yie = YoutubeIE(ydl)
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded = ydl.downloaded_info_dicts[0]
self.assertEqual(downloaded['format_id'], f1id)
self.assertEqual(downloaded['format_id'], f1['format_id'])
info_dict = _make_result([f2, f1], extractor='youtube')
ydl = YDL({'format': 'best/bestvideo'})
@ -251,7 +300,17 @@ class TestFormatSelection(unittest.TestCase):
yie._sort_formats(info_dict['formats'])
ydl.process_ie_result(info_dict)
downloaded = ydl.downloaded_info_dicts[0]
self.assertEqual(downloaded['format_id'], f1id)
self.assertEqual(downloaded['format_id'], f1['format_id'])
def test_invalid_format_specs(self):
def assert_syntax_error(format_spec):
ydl = YDL({'format': format_spec})
info_dict = _make_result([{'format_id': 'foo', 'url': TEST_URL}])
self.assertRaises(SyntaxError, ydl.process_ie_result, info_dict)
assert_syntax_error('bestvideo,,best')
assert_syntax_error('+bestaudio')
assert_syntax_error('bestvideo+')
def test_format_filtering(self):
formats = [
@ -308,6 +367,18 @@ class TestFormatSelection(unittest.TestCase):
downloaded = ydl.downloaded_info_dicts[0]
self.assertEqual(downloaded['format_id'], 'G')
ydl = YDL({'format': 'all[width>=400][width<=600]'})
ydl.process_ie_result(info_dict)
downloaded_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
self.assertEqual(downloaded_ids, ['B', 'C', 'D'])
ydl = YDL({'format': 'best[height<40]'})
try:
ydl.process_ie_result(info_dict)
except ExtractorError:
pass
self.assertEqual(ydl.downloaded_info_dicts, [])
class TestYoutubeDL(unittest.TestCase):
def test_subtitles(self):

View file

@ -21,6 +21,7 @@ import subprocess
import socket
import sys
import time
import tokenize
import traceback
if os.name == 'nt':
@ -34,6 +35,7 @@ from .compat import (
compat_http_client,
compat_kwargs,
compat_str,
compat_tokenize_tokenize,
compat_urllib_error,
compat_urllib_request,
)
@ -853,8 +855,8 @@ class YoutubeDL(object):
else:
raise Exception('Invalid result type: %s' % result_type)
def _apply_format_filter(self, format_spec, available_formats):
" Returns a tuple of the remaining format_spec and filtered formats "
def _build_format_filter(self, filter_spec):
" Returns a function to filter the formats according to the filter_spec "
OPERATORS = {
'<': operator.lt,
@ -864,13 +866,13 @@ class YoutubeDL(object):
'=': operator.eq,
'!=': operator.ne,
}
operator_rex = re.compile(r'''(?x)\s*\[
operator_rex = re.compile(r'''(?x)\s*
(?P<key>width|height|tbr|abr|vbr|asr|filesize|fps)
\s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
(?P<value>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)
\]$
$
''' % '|'.join(map(re.escape, OPERATORS.keys())))
m = operator_rex.search(format_spec)
m = operator_rex.search(filter_spec)
if m:
try:
comparison_value = int(m.group('value'))
@ -881,7 +883,7 @@ class YoutubeDL(object):
if comparison_value is None:
raise ValueError(
'Invalid value %r in format specification %r' % (
m.group('value'), format_spec))
m.group('value'), filter_spec))
op = OPERATORS[m.group('op')]
if not m:
@ -889,85 +891,250 @@ class YoutubeDL(object):
'=': operator.eq,
'!=': operator.ne,
}
str_operator_rex = re.compile(r'''(?x)\s*\[
str_operator_rex = re.compile(r'''(?x)
\s*(?P<key>ext|acodec|vcodec|container|protocol)
\s*(?P<op>%s)(?P<none_inclusive>\s*\?)?
\s*(?P<value>[a-zA-Z0-9_-]+)
\s*\]$
\s*$
''' % '|'.join(map(re.escape, STR_OPERATORS.keys())))
m = str_operator_rex.search(format_spec)
m = str_operator_rex.search(filter_spec)
if m:
comparison_value = m.group('value')
op = STR_OPERATORS[m.group('op')]
if not m:
raise ValueError('Invalid format specification %r' % format_spec)
raise ValueError('Invalid filter specification %r' % filter_spec)
def _filter(f):
actual_value = f.get(m.group('key'))
if actual_value is None:
return m.group('none_inclusive')
return op(actual_value, comparison_value)
new_formats = [f for f in available_formats if _filter(f)]
return _filter
new_format_spec = format_spec[:-len(m.group(0))]
if not new_format_spec:
new_format_spec = 'best'
def build_format_selector(self, format_spec):
def syntax_error(note, start):
message = (
'Invalid format specification: '
'{0}\n\t{1}\n\t{2}^'.format(note, format_spec, ' ' * start[1]))
return SyntaxError(message)
return (new_format_spec, new_formats)
PICKFIRST = 'PICKFIRST'
MERGE = 'MERGE'
SINGLE = 'SINGLE'
GROUP = 'GROUP'
FormatSelector = collections.namedtuple('FormatSelector', ['type', 'selector', 'filters'])
def select_format(self, format_spec, available_formats):
while format_spec.endswith(']'):
format_spec, available_formats = self._apply_format_filter(
format_spec, available_formats)
if not available_formats:
return None
def _parse_filter(tokens):
filter_parts = []
for type, string, start, _, _ in tokens:
if type == tokenize.OP and string == ']':
return ''.join(filter_parts)
else:
filter_parts.append(string)
if format_spec in ['best', 'worst', None]:
format_idx = 0 if format_spec == 'worst' else -1
audiovideo_formats = [
f for f in available_formats
if f.get('vcodec') != 'none' and f.get('acodec') != 'none']
if audiovideo_formats:
return audiovideo_formats[format_idx]
# for audio only (soundcloud) or video only (imgur) urls, select the best/worst audio format
elif (all(f.get('acodec') != 'none' for f in available_formats) or
all(f.get('vcodec') != 'none' for f in available_formats)):
return available_formats[format_idx]
elif format_spec == 'bestaudio':
audio_formats = [
f for f in available_formats
if f.get('vcodec') == 'none']
if audio_formats:
return audio_formats[-1]
elif format_spec == 'worstaudio':
audio_formats = [
f for f in available_formats
if f.get('vcodec') == 'none']
if audio_formats:
return audio_formats[0]
elif format_spec == 'bestvideo':
video_formats = [
f for f in available_formats
if f.get('acodec') == 'none']
if video_formats:
return video_formats[-1]
elif format_spec == 'worstvideo':
video_formats = [
f for f in available_formats
if f.get('acodec') == 'none']
if video_formats:
return video_formats[0]
else:
extensions = ['mp4', 'flv', 'webm', '3gp', 'm4a', 'mp3', 'ogg', 'aac', 'wav']
if format_spec in extensions:
filter_f = lambda f: f['ext'] == format_spec
else:
filter_f = lambda f: f['format_id'] == format_spec
matches = list(filter(filter_f, available_formats))
if matches:
return matches[-1]
return None
def _parse_format_selection(tokens, inside_merge=False, inside_choice=False, inside_group=False):
selectors = []
current_selector = None
for type, string, start, _, _ in tokens:
# ENCODING is only defined in python 3.x
if type == getattr(tokenize, 'ENCODING', None):
continue
elif type in [tokenize.NAME, tokenize.NUMBER]:
current_selector = FormatSelector(SINGLE, string, [])
elif type == tokenize.OP:
if string == ')':
if not inside_group:
# ')' will be handled by the parentheses group
tokens.restore_last_token()
break
elif inside_merge and string in ['/', ',']:
tokens.restore_last_token()
break
elif inside_choice and string == ',':
tokens.restore_last_token()
break
elif string == ',':
if not current_selector:
raise syntax_error('"," must follow a format selector', start)
selectors.append(current_selector)
current_selector = None
elif string == '/':
first_choice = current_selector
second_choice = _parse_format_selection(tokens, inside_choice=True)
current_selector = FormatSelector(PICKFIRST, (first_choice, second_choice), [])
elif string == '[':
if not current_selector:
current_selector = FormatSelector(SINGLE, 'best', [])
format_filter = _parse_filter(tokens)
current_selector.filters.append(format_filter)
elif string == '(':
if current_selector:
raise syntax_error('Unexpected "("', start)
group = _parse_format_selection(tokens, inside_group=True)
current_selector = FormatSelector(GROUP, group, [])
elif string == '+':
video_selector = current_selector
audio_selector = _parse_format_selection(tokens, inside_merge=True)
if not video_selector or not audio_selector:
raise syntax_error('"+" must be between two format selectors', start)
current_selector = FormatSelector(MERGE, (video_selector, audio_selector), [])
else:
raise syntax_error('Operator not recognized: "{0}"'.format(string), start)
elif type == tokenize.ENDMARKER:
break
if current_selector:
selectors.append(current_selector)
return selectors
def _build_selector_function(selector):
if isinstance(selector, list):
fs = [_build_selector_function(s) for s in selector]
def selector_function(formats):
for f in fs:
for format in f(formats):
yield format
return selector_function
elif selector.type == GROUP:
selector_function = _build_selector_function(selector.selector)
elif selector.type == PICKFIRST:
fs = [_build_selector_function(s) for s in selector.selector]
def selector_function(formats):
for f in fs:
picked_formats = list(f(formats))
if picked_formats:
return picked_formats
return []
elif selector.type == SINGLE:
format_spec = selector.selector
def selector_function(formats):
formats = list(formats)
if not formats:
return
if format_spec == 'all':
for f in formats:
yield f
elif format_spec in ['best', 'worst', None]:
format_idx = 0 if format_spec == 'worst' else -1
audiovideo_formats = [
f for f in formats
if f.get('vcodec') != 'none' and f.get('acodec') != 'none']
if audiovideo_formats:
yield audiovideo_formats[format_idx]
# for audio only (soundcloud) or video only (imgur) urls, select the best/worst audio format
elif (all(f.get('acodec') != 'none' for f in formats) or
all(f.get('vcodec') != 'none' for f in formats)):
yield formats[format_idx]
elif format_spec == 'bestaudio':
audio_formats = [
f for f in formats
if f.get('vcodec') == 'none']
if audio_formats:
yield audio_formats[-1]
elif format_spec == 'worstaudio':
audio_formats = [
f for f in formats
if f.get('vcodec') == 'none']
if audio_formats:
yield audio_formats[0]
elif format_spec == 'bestvideo':
video_formats = [
f for f in formats
if f.get('acodec') == 'none']
if video_formats:
yield video_formats[-1]
elif format_spec == 'worstvideo':
video_formats = [
f for f in formats
if f.get('acodec') == 'none']
if video_formats:
yield video_formats[0]
else:
extensions = ['mp4', 'flv', 'webm', '3gp', 'm4a', 'mp3', 'ogg', 'aac', 'wav']
if format_spec in extensions:
filter_f = lambda f: f['ext'] == format_spec
else:
filter_f = lambda f: f['format_id'] == format_spec
matches = list(filter(filter_f, formats))
if matches:
yield matches[-1]
elif selector.type == MERGE:
def _merge(formats_info):
format_1, format_2 = [f['format_id'] for f in formats_info]
# The first format must contain the video and the
# second the audio
if formats_info[0].get('vcodec') == 'none':
self.report_error('The first format must '
'contain the video, try using '
'"-f %s+%s"' % (format_2, format_1))
return
output_ext = (
formats_info[0]['ext']
if self.params.get('merge_output_format') is None
else self.params['merge_output_format'])
return {
'requested_formats': formats_info,
'format': '%s+%s' % (formats_info[0].get('format'),
formats_info[1].get('format')),
'format_id': '%s+%s' % (formats_info[0].get('format_id'),
formats_info[1].get('format_id')),
'width': formats_info[0].get('width'),
'height': formats_info[0].get('height'),
'resolution': formats_info[0].get('resolution'),
'fps': formats_info[0].get('fps'),
'vcodec': formats_info[0].get('vcodec'),
'vbr': formats_info[0].get('vbr'),
'stretched_ratio': formats_info[0].get('stretched_ratio'),
'acodec': formats_info[1].get('acodec'),
'abr': formats_info[1].get('abr'),
'ext': output_ext,
}
video_selector, audio_selector = map(_build_selector_function, selector.selector)
def selector_function(formats):
formats = list(formats)
for pair in itertools.product(video_selector(formats), audio_selector(formats)):
yield _merge(pair)
filters = [self._build_format_filter(f) for f in selector.filters]
def final_selector(formats):
for _filter in filters:
formats = list(filter(_filter, formats))
return selector_function(formats)
return final_selector
stream = io.BytesIO(format_spec.encode('utf-8'))
try:
tokens = list(compat_tokenize_tokenize(stream.readline))
except tokenize.TokenError:
raise syntax_error('Missing closing/opening brackets or parenthesis', (0, len(format_spec)))
class TokenIterator(object):
def __init__(self, tokens):
self.tokens = tokens
self.counter = 0
def __iter__(self):
return self
def __next__(self):
if self.counter >= len(self.tokens):
raise StopIteration()
value = self.tokens[self.counter]
self.counter += 1
return value
next = __next__
def restore_last_token(self):
self.counter -= 1
parsed_selector = _parse_format_selection(iter(TokenIterator(tokens)))
return _build_selector_function(parsed_selector)
def _calc_headers(self, info_dict):
res = std_headers.copy()
@ -1111,56 +1278,8 @@ class YoutubeDL(object):
req_format_list.append('bestvideo+bestaudio')
req_format_list.append('best')
req_format = '/'.join(req_format_list)
formats_to_download = []
if req_format == 'all':
formats_to_download = formats
else:
for rfstr in req_format.split(','):
# We can accept formats requested in the format: 34/5/best, we pick
# the first that is available, starting from left
req_formats = rfstr.split('/')
for rf in req_formats:
if re.match(r'.+?\+.+?', rf) is not None:
# Two formats have been requested like '137+139'
format_1, format_2 = rf.split('+')
formats_info = (self.select_format(format_1, formats),
self.select_format(format_2, formats))
if all(formats_info):
# The first format must contain the video and the
# second the audio
if formats_info[0].get('vcodec') == 'none':
self.report_error('The first format must '
'contain the video, try using '
'"-f %s+%s"' % (format_2, format_1))
return
output_ext = (
formats_info[0]['ext']
if self.params.get('merge_output_format') is None
else self.params['merge_output_format'])
selected_format = {
'requested_formats': formats_info,
'format': '%s+%s' % (formats_info[0].get('format'),
formats_info[1].get('format')),
'format_id': '%s+%s' % (formats_info[0].get('format_id'),
formats_info[1].get('format_id')),
'width': formats_info[0].get('width'),
'height': formats_info[0].get('height'),
'resolution': formats_info[0].get('resolution'),
'fps': formats_info[0].get('fps'),
'vcodec': formats_info[0].get('vcodec'),
'vbr': formats_info[0].get('vbr'),
'stretched_ratio': formats_info[0].get('stretched_ratio'),
'acodec': formats_info[1].get('acodec'),
'abr': formats_info[1].get('abr'),
'ext': output_ext,
}
else:
selected_format = None
else:
selected_format = self.select_format(rf, formats)
if selected_format is not None:
formats_to_download.append(selected_format)
break
format_selector = self.build_format_selector(req_format)
formats_to_download = list(format_selector(formats))
if not formats_to_download:
raise ExtractorError('requested format not available',
expected=True)

View file

@ -436,6 +436,11 @@ except TypeError: # Python 2.6
yield n
n += step
if sys.version_info >= (3, 0):
from tokenize import tokenize as compat_tokenize_tokenize
else:
from tokenize import generate_tokens as compat_tokenize_tokenize
__all__ = [
'compat_HTTPError',
'compat_basestring',
@ -457,6 +462,7 @@ __all__ = [
'compat_socket_create_connection',
'compat_str',
'compat_subprocess_get_DEVNULL',
'compat_tokenize_tokenize',
'compat_urllib_error',
'compat_urllib_parse',
'compat_urllib_parse_unquote',