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