Add format selection criteria longside/shortside, and tests for both (#30737)

test/formatselection/criteria_longside_shortside.py

@@ -0,0 +1,117 @@
+#!/usr/bin/env python
+# coding: utf-8
+
+from __future__ import unicode_literals
+
+# Allow direct execution
+if __name__ == '__main__':
+    import os
+    import sys
+    repo_dir = os.path.abspath(os.path.join(__file__, '../' * 3))
+    sys.path.insert(0, repo_dir)
+
+import unittest
+
+from youtube_dl.extractor import YoutubeIE
+from test.test_YoutubeDL import YDL, TEST_URL, _make_result
+
+
+default_common_video_properties = {
+    'url': TEST_URL,
+}
+
+
+def prepare_formats_info_dict(sizes, common={}):
+    """Convert sizes (id, width, height) to info_dict."""
+    formats = [{
+        'format_id': id,
+        'width': w,
+        'height': h,
+        **default_common_video_properties,
+        **common,
+    } for (id, w, h) in sizes]
+    info_dict = _make_result(formats)
+    return info_dict
+
+
+def pick_format_ids(sizes, criteria):
+    """Check which size(s) match the criteria. Return their IDs."""
+    ydl = YDL({'format': criteria})
+    yie = YoutubeIE(ydl)
+    info_dict = prepare_formats_info_dict(sizes)
+    yie._sort_formats(info_dict['formats'])
+    ydl.process_ie_result(info_dict.copy())
+    # ^-- :TODO: Do we need this .copy()?
+    picked_ids = [info['format_id'] for info in ydl.downloaded_info_dicts]
+    return picked_ids
+
+
+class TestFormatSelection(unittest.TestCase):
+    """Tests class for longside/shortside format selector."""
+
+    def test_fmtsel_criteria_longside_shortside(self):
+        """Find largest video within upper limits."""
+        # Feature request: https://github.com/ytdl-org/youtube-dl/issues/30737
+
+        crit = {'tendency': '', 'short': '', 'long': ''}
+
+        def verify_fmtsel(sizes, want):
+            crit_all = crit['tendency'] + crit['short'] + crit['long']
+            picked_ids = pick_format_ids(sizes, crit_all)
+            self.assertEqual(','.join(picked_ids), want)
+
+        sizes_h = [
+            ('A', 256, 144,),
+            ('B', 426, 240,),
+            ('C', 640, 360,),
+            ('D', 854, 480,),
+        ]
+        sizes_v = [(id, h, w) for (id, w, h) in sizes_h]
+        self.assertEqual(sizes_v, [
+            # This list is non-authoritative, merely for readers' reference.
+            ('A', 144, 256,),
+            ('B', 240, 426,),
+            ('C', 360, 640,),
+            ('D', 480, 854,),
+        ])
+
+        # def size_by_id(sizes, id):
+        #     return next((s for s in sizes if s[0] == id))
+
+        def verify_all_shapes_same(expected_id):
+            verify_fmtsel(sizes_h, expected_id)
+            verify_fmtsel(sizes_v, expected_id)
+
+        # First, test with no criteria (still empty from initialization above):
+        # verify_fmtsel(sizes_h, 'D')
+        # :TODO: ^-- "requested format not available", why not D?
+
+        crit['tendency'] = 'best'
+        verify_all_shapes_same('D')
+
+        crit['long'] = '[longside<=720]'
+        verify_all_shapes_same('C')
+
+        crit['long'] = '[longside<=420]'
+        verify_all_shapes_same('A')
+
+        def shortside_group_1(long, best):
+            crit['long'] = long
+            crit['short'] = '[shortside<=720]'
+            verify_all_shapes_same(best)
+
+            crit['short'] = '[shortside<=420]'
+            verify_all_shapes_same('C')
+
+            crit['short'] = '[shortside<=360]'
+            verify_all_shapes_same('C')
+
+            crit['short'] = '[shortside<360]'
+            verify_all_shapes_same('B')
+
+        shortside_group_1(long='', best='D')
+        shortside_group_1(long='[longside<=720]', best='C')
+
+
+if __name__ == '__main__':
+    unittest.main()

youtube_dl/YoutubeDL.py

@@ -345,6 +345,7 @@ class YoutubeDL(object):
         'duration', 'view_count', 'like_count', 'dislike_count', 'repost_count',
         'average_rating', 'comment_count', 'age_limit',
         'start_time', 'end_time',
+        'longside', 'shortside',
         'chapter_number', 'season_number', 'episode_number',
         'track_number', 'disc_number', 'release_year',
         'playlist_index',
@@ -1207,7 +1208,7 @@ class YoutubeDL(object):
             '!=': operator.ne,
         }
         operator_rex = re.compile(r'''(?x)\s*
-            (?P<key>width|height|tbr|abr|vbr|asr|filesize|filesize_approx|fps)
+            (?P<key>width|height|shortside|longside|tbr|abr|vbr|asr|filesize|filesize_approx|fps)
             \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
             (?P<value>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)
             $
@@ -1495,6 +1496,8 @@ class YoutubeDL(object):
                                                 formats_info[1].get('format_id')),
                         'width': formats_info[0].get('width'),
                         'height': formats_info[0].get('height'),
+                        'longside': formats_info[0].get('longside'),
+                        'shortside': formats_info[0].get('shortside'),
                         'resolution': formats_info[0].get('resolution'),
                         'fps': formats_info[0].get('fps'),
                         'vcodec': formats_info[0].get('vcodec'),
@@ -1624,6 +1627,17 @@ class YoutubeDL(object):
         sanitize_string_field(info_dict, 'id')
         sanitize_numeric_fields(info_dict)
 
+        def add_calculated_video_proprties(fmt):
+            if type(fmt) is not dict: return
+            dims = [fmt.get(side) for side in ('width', 'height',)]
+            dims = [n for n in dims if n is not None]
+            if len(dims):
+                fmt['shortside'] = min(iter(dims))
+                fmt['longside'] = max(iter(dims))
+
+        for fmt in [info_dict] + (info_dict.get('formats') or []):
+            add_calculated_video_proprties(fmt)
+
         if 'playlist' not in info_dict:
             # It isn't part of a playlist
             info_dict['playlist'] = None