git.bitcoin.ninja Git - youtube-dl/blob - test/test_download.py

   1 #!/usr/bin/env python
   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_warnings,
  14     get_params,
  15     gettestcases,
  16     expect_info_dict,
  17     try_rm,
  18     report_warning,
  19 )
  20
  21
  22 import hashlib
  23 import io
  24 import json
  25 import socket
  26
  27 import youtube_dl.YoutubeDL
  28 from youtube_dl.compat import (
  29     compat_http_client,
  30     compat_urllib_error,
  31     compat_HTTPError,
  32 )
  33 from youtube_dl.utils import (
  34     DownloadError,
  35     ExtractorError,
  36     format_bytes,
  37     UnavailableVideoError,
  38 )
  39 from youtube_dl.extractor import get_info_extractor
  40
  41 RETRIES = 3
  42
  43
  44 class YoutubeDL(youtube_dl.YoutubeDL):
  45     def __init__(self, *args, **kwargs):
  46         self.to_stderr = self.to_screen
  47         self.processed_info_dicts = []
  48         super(YoutubeDL, self).__init__(*args, **kwargs)
  49
  50     def report_warning(self, message):
  51         # Don't accept warnings during tests
  52         raise ExtractorError(message)
  53
  54     def process_info(self, info_dict):
  55         self.processed_info_dicts.append(info_dict)
  56         return super(YoutubeDL, self).process_info(info_dict)
  57
  58
  59 def _file_md5(fn):
  60     with open(fn, 'rb') as f:
  61         return hashlib.md5(f.read()).hexdigest()
  62
  63
  64 defs = gettestcases()
  65
  66
  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     def __str__(self):
  75         """Identify each test with the `add_ie` attribute, if available."""
  76
  77         def strclass(cls):
  78             """From 2.7's unittest; 2.6 had _strclass so we can't import it."""
  79             return '%s.%s' % (cls.__module__, cls.__name__)
  80
  81         add_ie = getattr(self, self._testMethodName).add_ie
  82         return '%s (%s)%s:' % (self._testMethodName,
  83                                strclass(self.__class__),
  84                                ' [%s]' % add_ie if add_ie else '')
  85
  86     def setUp(self):
  87         self.defs = defs
  88
  89 # Dynamically generate tests
  90
  91
  92 def generator(test_case, tname):
  93
  94     def test_template(self):
  95         ie = youtube_dl.extractor.get_info_extractor(test_case['name'])()
  96         other_ies = [get_info_extractor(ie_key)() for ie_key in test_case.get('add_ie', [])]
  97         is_playlist = any(k.startswith('playlist') for k in test_case)
  98         test_cases = test_case.get(
  99             'playlist', [] if is_playlist else [test_case])
 100
 101         def print_skipping(reason):
 102             print('Skipping %s: %s' % (test_case['name'], reason))
 103         if not ie.working():
 104             print_skipping('IE marked as not _WORKING')
 105             return
 106
 107         for tc in test_cases:
 108             info_dict = tc.get('info_dict', {})
 109             if not (info_dict.get('id') and info_dict.get('ext')):
 110                 raise Exception('Test definition incorrect. The output file cannot be known. Are both \'id\' and \'ext\' keys present?')
 111
 112         if 'skip' in test_case:
 113             print_skipping(test_case['skip'])
 114             return
 115         for other_ie in other_ies:
 116             if not other_ie.working():
 117                 print_skipping('test depends on %sIE, marked as not WORKING' % other_ie.ie_key())
 118                 return
 119
 120         params = get_params(test_case.get('params', {}))
 121         params['outtmpl'] = tname + '_' + params['outtmpl']
 122         if is_playlist and 'playlist' not in test_case:
 123             params.setdefault('extract_flat', 'in_playlist')
 124             params.setdefault('skip_download', True)
 125
 126         ydl = YoutubeDL(params, auto_init=False)
 127         ydl.add_default_info_extractors()
 128         finished_hook_called = set()
 129
 130         def _hook(status):
 131             if status['status'] == 'finished':
 132                 finished_hook_called.add(status['filename'])
 133         ydl.add_progress_hook(_hook)
 134         expect_warnings(ydl, test_case.get('expected_warnings', []))
 135
 136         def get_tc_filename(tc):
 137             return ydl.prepare_filename(tc.get('info_dict', {}))
 138
 139         res_dict = None
 140
 141         def try_rm_tcs_files(tcs=None):
 142             if tcs is None:
 143                 tcs = test_cases
 144             for tc in tcs:
 145                 tc_filename = get_tc_filename(tc)
 146                 try_rm(tc_filename)
 147                 try_rm(tc_filename + '.part')
 148                 try_rm(os.path.splitext(tc_filename)[0] + '.info.json')
 149         try_rm_tcs_files()
 150         try:
 151             try_num = 1
 152             while True:
 153                 try:
 154                     # We're not using .download here since that is just a shim
 155                     # for outside error handling, and returns the exit code
 156                     # instead of the result dict.
 157                     res_dict = ydl.extract_info(
 158                         test_case['url'],
 159                         force_generic_extractor=params.get('force_generic_extractor', False))
 160                 except (DownloadError, ExtractorError) as err:
 161                     # Check if the exception is not a network related one
 162                     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):
 163                         raise
 164
 165                     if try_num == RETRIES:
 166                         report_warning('%s failed due to network errors, skipping...' % tname)
 167                         return
 168
 169                     print('Retrying: {0} failed tries\n\n##########\n\n'.format(try_num))
 170
 171                     try_num += 1
 172                 else:
 173                     break
 174
 175             if is_playlist:
 176                 self.assertTrue(res_dict['_type'] in ['playlist', 'multi_video'])
 177                 self.assertTrue('entries' in res_dict)
 178                 expect_info_dict(self, res_dict, test_case.get('info_dict', {}))
 179
 180             if 'playlist_mincount' in test_case:
 181                 assertGreaterEqual(
 182                     self,
 183                     len(res_dict['entries']),
 184                     test_case['playlist_mincount'],
 185                     'Expected at least %d in playlist %s, but got only %d' % (
 186                         test_case['playlist_mincount'], test_case['url'],
 187                         len(res_dict['entries'])))
 188             if 'playlist_count' in test_case:
 189                 self.assertEqual(
 190                     len(res_dict['entries']),
 191                     test_case['playlist_count'],
 192                     'Expected %d entries in playlist %s, but got %d.' % (
 193                         test_case['playlist_count'],
 194                         test_case['url'],
 195                         len(res_dict['entries']),
 196                     ))
 197             if 'playlist_duration_sum' in test_case:
 198                 got_duration = sum(e['duration'] for e in res_dict['entries'])
 199                 self.assertEqual(
 200                     test_case['playlist_duration_sum'], got_duration)
 201
 202             # Generalize both playlists and single videos to unified format for
 203             # simplicity
 204             if 'entries' not in res_dict:
 205                 res_dict['entries'] = [res_dict]
 206
 207             for tc_num, tc in enumerate(test_cases):
 208                 tc_res_dict = res_dict['entries'][tc_num]
 209                 # First, check test cases' data against extracted data alone
 210                 expect_info_dict(self, tc_res_dict, tc.get('info_dict', {}))
 211                 # Now, check downloaded file consistency
 212                 tc_filename = get_tc_filename(tc)
 213                 if not test_case.get('params', {}).get('skip_download', False):
 214                     self.assertTrue(os.path.exists(tc_filename), msg='Missing file ' + tc_filename)
 215                     self.assertTrue(tc_filename in finished_hook_called)
 216                     expected_minsize = tc.get('file_minsize', 10000)
 217                     if expected_minsize is not None:
 218                         if params.get('test'):
 219                             expected_minsize = max(expected_minsize, 10000)
 220                         got_fsize = os.path.getsize(tc_filename)
 221                         assertGreaterEqual(
 222                             self, got_fsize, expected_minsize,
 223                             'Expected %s to be at least %s, but it\'s only %s ' %
 224                             (tc_filename, format_bytes(expected_minsize),
 225                                 format_bytes(got_fsize)))
 226                     if 'md5' in tc:
 227                         md5_for_file = _file_md5(tc_filename)
 228                         self.assertEqual(tc['md5'], md5_for_file)
 229                 # Finally, check test cases' data again but this time against
 230                 # extracted data from info JSON file written during processing
 231                 info_json_fn = os.path.splitext(tc_filename)[0] + '.info.json'
 232                 self.assertTrue(
 233                     os.path.exists(info_json_fn),
 234                     'Missing info file %s' % info_json_fn)
 235                 with io.open(info_json_fn, encoding='utf-8') as infof:
 236                     info_dict = json.load(infof)
 237                 expect_info_dict(self, info_dict, tc.get('info_dict', {}))
 238         finally:
 239             try_rm_tcs_files()
 240             if is_playlist and res_dict is not None and res_dict.get('entries'):
 241                 # Remove all other files that may have been extracted if the
 242                 # extractor returns full results even with extract_flat
 243                 res_tcs = [{'info_dict': e} for e in res_dict['entries']]
 244                 try_rm_tcs_files(res_tcs)
 245
 246     return test_template
 247
 248
 249 # And add them to TestDownload
 250 for n, test_case in enumerate(defs):
 251     tname = 'test_' + str(test_case['name'])
 252     i = 1
 253     while hasattr(TestDownload, tname):
 254         tname = 'test_%s_%d' % (test_case['name'], i)
 255         i += 1
 256     test_method = generator(test_case, tname)
 257     test_method.__name__ = str(tname)
 258     ie_list = test_case.get('add_ie')
 259     test_method.add_ie = ie_list and ','.join(ie_list)
 260     setattr(TestDownload, test_method.__name__, test_method)
 261     del test_method
 262
 263
 264 if __name__ == '__main__':
 265     unittest.main()