test/test_download.py

   1 #!/usr/bin/env python3
   2
   3 # Allow direct execution
   4 import os
   5 import sys
   6 import unittest
   7
   8 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
   9
  10
  11 import collections
  12 import hashlib
  13 import json
  14
  15 from test.helper import (
  16     assertGreaterEqual,
  17     expect_info_dict,
  18     expect_warnings,
  19     get_params,
  20     gettestcases,
  21     getwebpagetestcases,
  22     is_download_test,
  23     report_warning,
  24     try_rm,
  25 )
  26
  27 import yt_dlp.YoutubeDL  # isort: split
  28 from yt_dlp.extractor import get_info_extractor
  29 from yt_dlp.networking.exceptions import HTTPError, TransportError
  30 from yt_dlp.utils import (
  31     DownloadError,
  32     ExtractorError,
  33     UnavailableVideoError,
  34     YoutubeDLError,
  35     format_bytes,
  36     join_nonempty,
  37 )
  38
  39 RETRIES = 3
  40
  41
  42 class YoutubeDL(yt_dlp.YoutubeDL):
  43     def __init__(self, *args, **kwargs):
  44         self.to_stderr = self.to_screen
  45         self.processed_info_dicts = []
  46         super().__init__(*args, **kwargs)
  47
  48     def report_warning(self, message, *args, **kwargs):
  49         # Don't accept warnings during tests
  50         raise ExtractorError(message)
  51
  52     def process_info(self, info_dict):
  53         self.processed_info_dicts.append(info_dict.copy())
  54         return super().process_info(info_dict)
  55
  56
  57 def _file_md5(fn):
  58     with open(fn, 'rb') as f:
  59         return hashlib.md5(f.read()).hexdigest()
  60
  61
  62 normal_test_cases = gettestcases()
  63 webpage_test_cases = getwebpagetestcases()
  64 tests_counter = collections.defaultdict(collections.Counter)
  65
  66
  67 @is_download_test
  68 class TestDownload(unittest.TestCase):
  69     # Parallel testing in nosetests. See
  70     # http://nose.readthedocs.org/en/latest/doc_tests/test_multiprocess/multiprocess.html
  71     _multiprocess_shared_ = True
  72
  73     maxDiff = None
  74
  75     COMPLETED_TESTS = {}
  76
  77     def __str__(self):
  78         """Identify each test with the `add_ie` attribute, if available."""
  79         cls, add_ie = type(self), getattr(self, self._testMethodName).add_ie
  80         return f'{self._testMethodName} ({cls.__module__}.{cls.__name__}){f" [{add_ie}]" if add_ie else ""}:'
  81
  82
  83 # Dynamically generate tests
  84
  85 def generator(test_case, tname):
  86     def test_template(self):
  87         if self.COMPLETED_TESTS.get(tname):
  88             return
  89         self.COMPLETED_TESTS[tname] = True
  90         ie = yt_dlp.extractor.get_info_extractor(test_case['name'])()
  91         other_ies = [get_info_extractor(ie_key)() for ie_key in test_case.get('add_ie', [])]
  92         is_playlist = any(k.startswith('playlist') for k in test_case)
  93         test_cases = test_case.get(
  94             'playlist', [] if is_playlist else [test_case])
  95
  96         def print_skipping(reason):
  97             print('Skipping %s: %s' % (test_case['name'], reason))
  98             self.skipTest(reason)
  99
 100         if not ie.working():
 101             print_skipping('IE marked as not _WORKING')
 102
 103         for tc in test_cases:
 104             if tc.get('expected_exception'):
 105                 continue
 106             info_dict = tc.get('info_dict', {})
 107             params = tc.get('params', {})
 108             if not info_dict.get('id'):
 109                 raise Exception(f'Test {tname} definition incorrect - "id" key is not present')
 110             elif not info_dict.get('ext') and info_dict.get('_type', 'video') == 'video':
 111                 if params.get('skip_download') and params.get('ignore_no_formats_error'):
 112                     continue
 113                 raise Exception(f'Test {tname} definition incorrect - "ext" key must be present to define the output file')
 114
 115         if 'skip' in test_case:
 116             print_skipping(test_case['skip'])
 117
 118         for other_ie in other_ies:
 119             if not other_ie.working():
 120                 print_skipping('test depends on %sIE, marked as not WORKING' % other_ie.ie_key())
 121
 122         params = get_params(test_case.get('params', {}))
 123         params['outtmpl'] = tname + '_' + params['outtmpl']
 124         if is_playlist and 'playlist' not in test_case:
 125             params.setdefault('extract_flat', 'in_playlist')
 126             params.setdefault('playlistend', test_case.get(
 127                 'playlist_mincount', test_case.get('playlist_count', -2) + 1))
 128             params.setdefault('skip_download', True)
 129
 130         ydl = YoutubeDL(params, auto_init=False)
 131         ydl.add_default_info_extractors()
 132         finished_hook_called = set()
 133
 134         def _hook(status):
 135             if status['status'] == 'finished':
 136                 finished_hook_called.add(status['filename'])
 137         ydl.add_progress_hook(_hook)
 138         expect_warnings(ydl, test_case.get('expected_warnings', []))
 139
 140         def get_tc_filename(tc):
 141             return ydl.prepare_filename(dict(tc.get('info_dict', {})))
 142
 143         res_dict = None
 144
 145         def match_exception(err):
 146             expected_exception = test_case.get('expected_exception')
 147             if not expected_exception:
 148                 return False
 149             if err.__class__.__name__ == expected_exception:
 150                 return True
 151             for exc in err.exc_info:
 152                 if exc.__class__.__name__ == expected_exception:
 153                     return True
 154             return False
 155
 156         def try_rm_tcs_files(tcs=None):
 157             if tcs is None:
 158                 tcs = test_cases
 159             for tc in tcs:
 160                 tc_filename = get_tc_filename(tc)
 161                 try_rm(tc_filename)
 162                 try_rm(tc_filename + '.part')
 163                 try_rm(os.path.splitext(tc_filename)[0] + '.info.json')
 164         try_rm_tcs_files()
 165         try:
 166             try_num = 1
 167             while True:
 168                 try:
 169                     # We're not using .download here since that is just a shim
 170                     # for outside error handling, and returns the exit code
 171                     # instead of the result dict.
 172                     res_dict = ydl.extract_info(
 173                         test_case['url'],
 174                         force_generic_extractor=params.get('force_generic_extractor', False))
 175                 except (DownloadError, ExtractorError) as err:
 176                     # Check if the exception is not a network related one
 177                     if not isinstance(err.exc_info[1], (TransportError, UnavailableVideoError)) or (isinstance(err.exc_info[1], HTTPError) and err.exc_info[1].status == 503):
 178                         if match_exception(err):
 179                             return
 180                         err.msg = f'{getattr(err, "msg", err)} ({tname})'
 181                         raise
 182
 183                     if try_num == RETRIES:
 184                         report_warning('%s failed due to network errors, skipping...' % tname)
 185                         return
 186
 187                     print(f'Retrying: {try_num} failed tries\n\n##########\n\n')
 188
 189                     try_num += 1
 190                 except YoutubeDLError as err:
 191                     if match_exception(err):
 192                         return
 193                     raise
 194                 else:
 195                     break
 196
 197             if is_playlist:
 198                 self.assertTrue(res_dict['_type'] in ['playlist', 'multi_video'])
 199                 self.assertTrue('entries' in res_dict)
 200                 expect_info_dict(self, res_dict, test_case.get('info_dict', {}))
 201
 202             if 'playlist_mincount' in test_case:
 203                 assertGreaterEqual(
 204                     self,
 205                     len(res_dict['entries']),
 206                     test_case['playlist_mincount'],
 207                     'Expected at least %d in playlist %s, but got only %d' % (
 208                         test_case['playlist_mincount'], test_case['url'],
 209                         len(res_dict['entries'])))
 210             if 'playlist_count' in test_case:
 211                 self.assertEqual(
 212                     len(res_dict['entries']),
 213                     test_case['playlist_count'],
 214                     'Expected %d entries in playlist %s, but got %d.' % (
 215                         test_case['playlist_count'],
 216                         test_case['url'],
 217                         len(res_dict['entries']),
 218                     ))
 219             if 'playlist_duration_sum' in test_case:
 220                 got_duration = sum(e['duration'] for e in res_dict['entries'])
 221                 self.assertEqual(
 222                     test_case['playlist_duration_sum'], got_duration)
 223
 224             # Generalize both playlists and single videos to unified format for
 225             # simplicity
 226             if 'entries' not in res_dict:
 227                 res_dict['entries'] = [res_dict]
 228
 229             for tc_num, tc in enumerate(test_cases):
 230                 tc_res_dict = res_dict['entries'][tc_num]
 231                 # First, check test cases' data against extracted data alone
 232                 expect_info_dict(self, tc_res_dict, tc.get('info_dict', {}))
 233                 if tc_res_dict.get('_type', 'video') != 'video':
 234                     continue
 235                 # Now, check downloaded file consistency
 236                 tc_filename = get_tc_filename(tc)
 237                 if not test_case.get('params', {}).get('skip_download', False):
 238                     self.assertTrue(os.path.exists(tc_filename), msg='Missing file ' + tc_filename)
 239                     self.assertTrue(tc_filename in finished_hook_called)
 240                     expected_minsize = tc.get('file_minsize', 10000)
 241                     if expected_minsize is not None:
 242                         if params.get('test'):
 243                             expected_minsize = max(expected_minsize, 10000)
 244                         got_fsize = os.path.getsize(tc_filename)
 245                         assertGreaterEqual(
 246                             self, got_fsize, expected_minsize,
 247                             'Expected %s to be at least %s, but it\'s only %s ' %
 248                             (tc_filename, format_bytes(expected_minsize),
 249                                 format_bytes(got_fsize)))
 250                     if 'md5' in tc:
 251                         md5_for_file = _file_md5(tc_filename)
 252                         self.assertEqual(tc['md5'], md5_for_file)
 253                 # Finally, check test cases' data again but this time against
 254                 # extracted data from info JSON file written during processing
 255                 info_json_fn = os.path.splitext(tc_filename)[0] + '.info.json'
 256                 self.assertTrue(
 257                     os.path.exists(info_json_fn),
 258                     'Missing info file %s' % info_json_fn)
 259                 with open(info_json_fn, encoding='utf-8') as infof:
 260                     info_dict = json.load(infof)
 261                 expect_info_dict(self, info_dict, tc.get('info_dict', {}))
 262         finally:
 263             try_rm_tcs_files()
 264             if is_playlist and res_dict is not None and res_dict.get('entries'):
 265                 # Remove all other files that may have been extracted if the
 266                 # extractor returns full results even with extract_flat
 267                 res_tcs = [{'info_dict': e} for e in res_dict['entries']]
 268                 try_rm_tcs_files(res_tcs)
 269             ydl.close()
 270     return test_template
 271
 272
 273 # And add them to TestDownload
 274 def inject_tests(test_cases, label=''):
 275     for test_case in test_cases:
 276         name = test_case['name']
 277         tname = join_nonempty('test', name, label, tests_counter[name][label], delim='_')
 278         tests_counter[name][label] += 1
 279
 280         test_method = generator(test_case, tname)
 281         test_method.__name__ = tname
 282         test_method.add_ie = ','.join(test_case.get('add_ie', []))
 283         setattr(TestDownload, test_method.__name__, test_method)
 284
 285
 286 inject_tests(normal_test_cases)
 287
 288 # TODO: disable redirection to the IE to ensure we are actually testing the webpage extraction
 289 inject_tests(webpage_test_cases, 'webpage')
 290
 291
 292 def batch_generator(name):
 293     def test_template(self):
 294         for label, num_tests in tests_counter[name].items():
 295             for i in range(num_tests):
 296                 test_name = join_nonempty('test', name, label, i, delim='_')
 297                 try:
 298                     getattr(self, test_name)()
 299                 except unittest.SkipTest:
 300                     print(f'Skipped {test_name}')
 301
 302     return test_template
 303
 304
 305 for name in tests_counter:
 306     test_method = batch_generator(name)
 307     test_method.__name__ = f'test_{name}_all'
 308     test_method.add_ie = ''
 309     setattr(TestDownload, test_method.__name__, test_method)
 310 del test_method
 311
 312
 313 if __name__ == '__main__':
 314     unittest.main()