def gettestcases(include_onlymatching=False):
for ie in youtube_dl.extractor.gen_extractors():
- t = getattr(ie, '_TEST', None)
- if t:
- assert not hasattr(ie, '_TESTS'), \
- '%s has _TEST and _TESTS' % type(ie).__name__
- tests = [t]
- else:
- tests = getattr(ie, '_TESTS', [])
- for t in tests:
- if not include_onlymatching and t.get('only_matching', False):
- continue
- t['name'] = type(ie).__name__[:-len('IE')]
- yield t
+ for tc in ie.get_testcases(include_onlymatching):
+ yield tc
md5 = lambda s: hashlib.md5(s.encode('utf-8')).hexdigest()
-def expect_info_dict(self, expected_dict, got_dict):
+def expect_info_dict(self, got_dict, expected_dict):
for info_field, expected in expected_dict.items():
if isinstance(expected, compat_str) and expected.startswith('re:'):
got = got_dict.get(info_field)
else:
if isinstance(expected, compat_str) and expected.startswith('md5:'):
got = 'md5:' + md5(got_dict.get(info_field))
+ elif isinstance(expected, compat_str) and expected.startswith('mincount:'):
+ got = got_dict.get(info_field)
+ self.assertTrue(
+ isinstance(got, list),
+ 'Expected field %s to be a list, but it is of type %s' % (
+ info_field, type(got).__name__))
+ expected_num = int(expected.partition(':')[2])
+ assertGreaterEqual(
+ self, len(got), expected_num,
+ 'Expected %d items in field %s, but only got %d' % (
+ expected_num, info_field, len(got)
+ )
+ )
+ continue
else:
got = got_dict.get(info_field)
self.assertEqual(expected, got,
else:
m = re.match(regexp, text)
if not m:
- note = 'Regexp didn\'t match: %r not found in %r' % (regexp, text)
+ note = 'Regexp didn\'t match: %r not found' % (regexp)
+ if len(text) < 1000:
+ note += ' in %r' % text
if msg is None:
msg = note
else: