test/test_download.py

   1 #!/usr/bin/env python3
   2
   3 from __future__ import unicode_literals
   4
   5 # Allow direct execution
   6 import os
   7 import sys
   8 import unittest
   9 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
  10
  11 from test.helper import (
  12     assertGreaterEqual,
  13     expect_info_dict,
  14     expect_warnings,
  15     get_params,
  16     gettestcases,
  17     is_download_test,
  18     report_warning,
  19     try_rm,
  20 )
  21
  22
  23 import hashlib
  24 import io
  25 import json
  26 import socket
  27
  28 import yt_dlp.YoutubeDL
  29 from yt_dlp.compat import (
  30     compat_http_client,
  31     compat_urllib_error,
  32     compat_HTTPError,
  33 )
  34 from yt_dlp.utils import (
  35     DownloadError,
  36     ExtractorError,
  37     format_bytes,
  38     UnavailableVideoError,
  39 )
  40 from yt_dlp.extractor import get_info_extractor
  41
  42 RETRIES = 3
  43
  44
  45 class YoutubeDL(yt_dlp.YoutubeDL):
  46     def __init__(self, *args, **kwargs):
  47         self.to_stderr = self.to_screen
  48         self.processed_info_dicts = []
  49         super(YoutubeDL, self).__init__(*args, **kwargs)
  50
  51     def report_warning(self, message):
  52         # Don't accept warnings during tests
  53         raise ExtractorError(message)
  54
  55     def process_info(self, info_dict):
  56         self.processed_info_dicts.append(info_dict)
  57         return super(YoutubeDL, self).process_info(info_dict)
  58
  59
  60 def _file_md5(fn):
  61     with open(fn, 'rb') as f:
  62         return hashlib.md5(f.read()).hexdigest()
  63
  64
  65 defs = gettestcases()
  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     def __str__(self):
  77         """Identify each test with the `add_ie` attribute, if available."""
  78
  79         def strclass(cls):
  80             """From 2.7's unittest; 2.6 had _strclass so we can't import it."""
  81             return '%s.%s' % (cls.__module__, cls.__name__)
  82
  83         add_ie = getattr(self, self._testMethodName).add_ie
  84         return '%s (%s)%s:' % (self._testMethodName,
  85                                strclass(self.__class__),
  86                                ' [%s]' % add_ie if add_ie else '')
  87
  88     def setUp(self):
  89         self.defs = defs
  90
  91 # Dynamically generate tests
  92
  93
  94 def generator(test_case, tname):
  95
  96     def test_template(self):
  97         ie = yt_dlp.extractor.get_info_extractor(test_case['name'])()
  98         other_ies = [get_info_extractor(ie_key)() for ie_key in test_case.get('add_ie', [])]
  99         is_playlist = any(k.startswith('playlist') for k in test_case)
 100         test_cases = test_case.get(
 101             'playlist', [] if is_playlist else [test_case])
 102
 103         def print_skipping(reason):
 104             print('Skipping %s: %s' % (test_case['name'], reason))
 105         if not ie.working():
 106             print_skipping('IE marked as not _WORKING')
 107             return
 108
 109         for tc in test_cases:
 110             info_dict = tc.get('info_dict', {})
 111             if not (info_dict.get('id') and info_dict.get('ext')):
 112                 raise Exception('Test definition incorrect. The output file cannot be known. Are both \'id\' and \'ext\' keys present?')
 113
 114         if 'skip' in test_case:
 115             print_skipping(test_case['skip'])
 116             return
 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                 return
 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('playlist_mincount'))
 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(tc.get('info_dict', {}))
 141
 142         res_dict = None
 143
 144         def try_rm_tcs_files(tcs=None):
 145             if tcs is None:
 146                 tcs = test_cases
 147             for tc in tcs:
 148                 tc_filename = get_tc_filename(tc)
 149                 try_rm(tc_filename)
 150                 try_rm(tc_filename + '.part')
 151                 try_rm(os.path.splitext(tc_filename)[0] + '.info.json')
 152         try_rm_tcs_files()
 153         try:
 154             try_num = 1
 155             while True:
 156                 try:
 157                     # We're not using .download here since that is just a shim
 158                     # for outside error handling, and returns the exit code
 159                     # instead of the result dict.
 160                     res_dict = ydl.extract_info(
 161                         test_case['url'],
 162                         force_generic_extractor=params.get('force_generic_extractor', False))
 163                 except (DownloadError, ExtractorError) as err:
 164                     # Check if the exception is not a network related one
 165                     if not err.exc_info[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError, compat_http_client.BadStatusLine) or (err.exc_info[0] == compat_HTTPError and err.exc_info[1].code == 503):
 166                         raise
 167
 168                     if try_num == RETRIES:
 169                         report_warning('%s failed due to network errors, skipping...' % tname)
 170                         return
 171
 172                     print('Retrying: {0} failed tries\n\n##########\n\n'.format(try_num))
 173
 174                     try_num += 1
 175                 else:
 176                     break
 177
 178             if is_playlist:
 179                 self.assertTrue(res_dict['_type'] in ['playlist', 'multi_video'])
 180                 self.assertTrue('entries' in res_dict)
 181                 expect_info_dict(self, res_dict, test_case.get('info_dict', {}))
 182
 183             if 'playlist_mincount' in test_case:
 184                 assertGreaterEqual(
 185                     self,
 186                     len(res_dict['entries']),
 187                     test_case['playlist_mincount'],
 188                     'Expected at least %d in playlist %s, but got only %d' % (
 189                         test_case['playlist_mincount'], test_case['url'],
 190                         len(res_dict['entries'])))
 191             if 'playlist_count' in test_case:
 192                 self.assertEqual(
 193                     len(res_dict['entries']),
 194                     test_case['playlist_count'],
 195                     'Expected %d entries in playlist %s, but got %d.' % (
 196                         test_case['playlist_count'],
 197                         test_case['url'],
 198                         len(res_dict['entries']),
 199                     ))
 200             if 'playlist_duration_sum' in test_case:
 201                 got_duration = sum(e['duration'] for e in res_dict['entries'])
 202                 self.assertEqual(
 203                     test_case['playlist_duration_sum'], got_duration)
 204
 205             # Generalize both playlists and single videos to unified format for
 206             # simplicity
 207             if 'entries' not in res_dict:
 208                 res_dict['entries'] = [res_dict]
 209
 210             for tc_num, tc in enumerate(test_cases):
 211                 tc_res_dict = res_dict['entries'][tc_num]
 212                 # First, check test cases' data against extracted data alone
 213                 expect_info_dict(self, tc_res_dict, tc.get('info_dict', {}))
 214                 # Now, check downloaded file consistency
 215                 tc_filename = get_tc_filename(tc)
 216                 if not test_case.get('params', {}).get('skip_download', False):
 217                     self.assertTrue(os.path.exists(tc_filename), msg='Missing file ' + tc_filename)
 218                     self.assertTrue(tc_filename in finished_hook_called)
 219                     expected_minsize = tc.get('file_minsize', 10000)
 220                     if expected_minsize is not None:
 221                         if params.get('test'):
 222                             expected_minsize = max(expected_minsize, 10000)
 223                         got_fsize = os.path.getsize(tc_filename)
 224                         assertGreaterEqual(
 225                             self, got_fsize, expected_minsize,
 226                             'Expected %s to be at least %s, but it\'s only %s ' %
 227                             (tc_filename, format_bytes(expected_minsize),
 228                                 format_bytes(got_fsize)))
 229                     if 'md5' in tc:
 230                         md5_for_file = _file_md5(tc_filename)
 231                         self.assertEqual(tc['md5'], md5_for_file)
 232                 # Finally, check test cases' data again but this time against
 233                 # extracted data from info JSON file written during processing
 234                 info_json_fn = os.path.splitext(tc_filename)[0] + '.info.json'
 235                 self.assertTrue(
 236                     os.path.exists(info_json_fn),
 237                     'Missing info file %s' % info_json_fn)
 238                 with io.open(info_json_fn, encoding='utf-8') as infof:
 239                     info_dict = json.load(infof)
 240                 expect_info_dict(self, info_dict, tc.get('info_dict', {}))
 241         finally:
 242             try_rm_tcs_files()
 243             if is_playlist and res_dict is not None and res_dict.get('entries'):
 244                 # Remove all other files that may have been extracted if the
 245                 # extractor returns full results even with extract_flat
 246                 res_tcs = [{'info_dict': e} for e in res_dict['entries']]
 247                 try_rm_tcs_files(res_tcs)
 248
 249     return test_template
 250
 251
 252 # And add them to TestDownload
 253 for n, test_case in enumerate(defs):
 254     tname = 'test_' + str(test_case['name'])
 255     i = 1
 256     while hasattr(TestDownload, tname):
 257         tname = 'test_%s_%d' % (test_case['name'], i)
 258         i += 1
 259     test_method = generator(test_case, tname)
 260     test_method.__name__ = str(tname)
 261     ie_list = test_case.get('add_ie')
 262     test_method.add_ie = ie_list and ','.join(ie_list)
 263     setattr(TestDownload, test_method.__name__, test_method)
 264     del test_method
 265
 266
 267 if __name__ == '__main__':
 268     unittest.main()