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