[YoutubeDL] Add generic video filtering (Fixes #4916)

This functionality is intended to eventually encompass the current format filtering.

test/test_utils.py

@@ -53,6 +53,7 @@
     version_tuple,
     xpath_with_ns,
     render_table,
+    match_str,
 )
 
 
@@ -459,6 +460,37 @@ def test_render_table(self):
             '123  4\n'
             '9999 51')
 
+    def test_match_str(self):
+        self.assertRaises(ValueError, match_str, 'xy>foobar', {})
+        self.assertFalse(match_str('xy', {'x': 1200}))
+        self.assertTrue(match_str('!xy', {'x': 1200}))
+        self.assertTrue(match_str('x', {'x': 1200}))
+        self.assertFalse(match_str('!x', {'x': 1200}))
+        self.assertTrue(match_str('x', {'x': 0}))
+        self.assertFalse(match_str('x>0', {'x': 0}))
+        self.assertFalse(match_str('x>0', {}))
+        self.assertTrue(match_str('x>?0', {}))
+        self.assertTrue(match_str('x>1K', {'x': 1200}))
+        self.assertFalse(match_str('x>2K', {'x': 1200}))
+        self.assertTrue(match_str('x>=1200 & x < 1300', {'x': 1200}))
+        self.assertFalse(match_str('x>=1100 & x < 1200', {'x': 1200}))
+        self.assertFalse(match_str('y=a212', {'y': 'foobar42'}))
+        self.assertTrue(match_str('y=foobar42', {'y': 'foobar42'}))
+        self.assertFalse(match_str('y!=foobar42', {'y': 'foobar42'}))
+        self.assertTrue(match_str('y!=foobar2', {'y': 'foobar42'}))
+        self.assertFalse(match_str(
+            'like_count > 100 & dislike_count <? 50 & description',
+            {'like_count': 90, 'description': 'foo'}))
+        self.assertTrue(match_str(
+            'like_count > 100 & dislike_count <? 50 & description',
+            {'like_count': 190, 'description': 'foo'}))
+        self.assertFalse(match_str(
+            'like_count > 100 & dislike_count <? 50 & description',
+            {'like_count': 190, 'dislike_count': 60, 'description': 'foo'}))
+        self.assertFalse(match_str(
+            'like_count > 100 & dislike_count <? 50 & description',
+            {'like_count': 190, 'dislike_count': 10}))
+
 
 if __name__ == '__main__':
     unittest.main()

youtube_dl/YoutubeDL.py

@@ -228,6 +228,11 @@ class YoutubeDL(object):
     external_downloader:  Executable of the external downloader to call.
     listformats:       Print an overview of available video formats and exit.
     list_thumbnails:   Print a table of all thumbnails and exit.
+    match_filter:      A function that gets called with the info_dict of
+                       every video.
+                       If it returns a message, the video is ignored.
+                       If it returns None, the video is downloaded.
+                       match_filter_func in utils.py is one example for this.
 
 
     The following parameters are not used by YoutubeDL itself, they are used by
@@ -583,9 +588,16 @@ def _match_entry(self, info_dict):
             if max_views is not None and view_count > max_views:
                 return 'Skipping %s, because it has exceeded the maximum view count (%d/%d)' % (video_title, view_count, max_views)
         if age_restricted(info_dict.get('age_limit'), self.params.get('age_limit')):
-            return 'Skipping "%s" because it is age restricted' % title
+            return 'Skipping "%s" because it is age restricted' % video_title
         if self.in_download_archive(info_dict):
             return '%s has already been recorded in archive' % video_title
+
+        match_filter = self.params.get('match_filter')
+        if match_filter is not None:
+            ret = match_filter(info_dict)
+            if ret is not None:
+                return ret
+
         return None
 
     @staticmethod

youtube_dl/__init__.py

@@ -23,9 +23,10 @@
 )
 from .utils import (
     DateRange,
-    DEFAULT_OUTTMPL,
     decodeOption,
+    DEFAULT_OUTTMPL,
     DownloadError,
+    match_filter_func,
     MaxDownloadsReached,
     preferredencoding,
     read_batch_urls,
@@ -247,6 +248,9 @@ def _real_main(argv=None):
             xattr  # Confuse flake8
         except ImportError:
             parser.error('setting filesize xattr requested but python-xattr is not available')
+    match_filter = (
+        None if opts.match_filter is None
+        else match_filter_func(opts.match_filter))
 
     ydl_opts = {
         'usenetrc': opts.usenetrc,
@@ -344,6 +348,7 @@ def _real_main(argv=None):
         'list_thumbnails': opts.list_thumbnails,
         'playlist_items': opts.playlist_items,
         'xattr_set_filesize': opts.xattr_set_filesize,
+        'match_filter': match_filter,
     }
 
     with YoutubeDL(ydl_opts) as ydl:

youtube_dl/options.py

@@ -244,6 +244,25 @@ def _hide_login_info(opts):
         '--max-views',
         metavar='COUNT', dest='max_views', default=None, type=int,
         help='Do not download any videos with more than COUNT views')
+    selection.add_option(
+        '--match-filter',
+        metavar='FILTER', dest='match_filter', default=None,
+        help=(
+            '(Experimental) Generic video filter. '
+            'Specify any key (see help for -o for a list of available keys) to'
+            ' match if the key is present, '
+            '!key to check if the key is not present,'
+            'key > NUMBER (like "comment_count > 12", also works with '
+            '>=, <, <=, !=, =) to compare against a number, and '
+            '& to require multiple matches. '
+            'Values which are not known are excluded unless you'
+            ' put a question mark (?) after the operator.'
+            'For example, to only match videos that have been liked more than '
+            '100 times and disliked less than 50 times (or the dislike '
+            'functionality is not available at the given service), but who '
+            'also have a description, use  --match-filter '
+            '"like_count > 100 & dislike_count <? 50 & description" .'
+        ))
     selection.add_option(
         '--no-playlist',
         action='store_true', dest='noplaylist', default=False,

youtube_dl/utils.py

@@ -17,6 +17,7 @@
 import json
 import locale
 import math
+import operator
 import os
 import pipes
 import platform
@@ -1678,3 +1679,79 @@ def render_table(header_row, data):
     max_lens = [max(len(compat_str(v)) for v in col) for col in zip(*table)]
     format_str = ' '.join('%-' + compat_str(ml + 1) + 's' for ml in max_lens[:-1]) + '%s'
     return '\n'.join(format_str % tuple(row) for row in table)
+
+
+def _match_one(filter_part, dct):
+    COMPARISON_OPERATORS = {
+        '<': operator.lt,
+        '<=': operator.le,
+        '>': operator.gt,
+        '>=': operator.ge,
+        '=': operator.eq,
+        '!=': operator.ne,
+    }
+    operator_rex = re.compile(r'''(?x)\s*
+        (?P<key>[a-z_]+)
+        \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
+        (?:
+            (?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
+            (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
+        )
+        \s*$
+        ''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
+    m = operator_rex.search(filter_part)
+    if m:
+        op = COMPARISON_OPERATORS[m.group('op')]
+        if m.group('strval') is not None:
+            if m.group('op') not in ('=', '!='):
+                raise ValueError(
+                    'Operator %s does not support string values!' % m.group('op'))
+            comparison_value = m.group('strval')
+        else:
+            try:
+                comparison_value = int(m.group('intval'))
+            except ValueError:
+                comparison_value = parse_filesize(m.group('intval'))
+                if comparison_value is None:
+                    comparison_value = parse_filesize(m.group('intval') + 'B')
+                if comparison_value is None:
+                    raise ValueError(
+                        'Invalid integer value %r in filter part %r' % (
+                            m.group('intval'), filter_part))
+        actual_value = dct.get(m.group('key'))
+        if actual_value is None:
+            return m.group('none_inclusive')
+        return op(actual_value, comparison_value)
+
+    UNARY_OPERATORS = {
+        '': lambda v: v is not None,
+        '!': lambda v: v is None,
+    }
+    operator_rex = re.compile(r'''(?x)\s*
+        (?P<op>%s)\s*(?P<key>[a-z_]+)
+        \s*$
+        ''' % '|'.join(map(re.escape, UNARY_OPERATORS.keys())))
+    m = operator_rex.search(filter_part)
+    if m:
+        op = UNARY_OPERATORS[m.group('op')]
+        actual_value = dct.get(m.group('key'))
+        return op(actual_value)
+
+    raise ValueError('Invalid filter part %r' % filter_part)
+
+
+def match_str(filter_str, dct):
+    """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
+
+    return all(
+        _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
+
+
+def match_filter_func(filter_str):
+    def _match_func(info_dict):
+        if match_str(filter_str, info_dict):
+            return None
+        else:
+            video_title = info_dict.get('title', info_dict.get('id', 'video'))
+            return '%s does not pass filter %s, skipping ..' % (video_title, filter_str)
+    return _match_func