diff options
| author | Sergey M․ <dstftw@gmail.com> | 2016-08-02 22:55:14 +0700 |
|---|---|---|
| committer | Sergey M․ <dstftw@gmail.com> | 2016-08-02 22:55:14 +0700 |
| commit | b070564efb190e1740c4e9d5fdde25cc436c72c3 (patch) | |
| tree | ef82754ee65b208b9264f2e74de2cc5913806d33 | |
| parent | ce28252c485cea5cdc2b0f5fd9f1c8d50c26c7a2 (diff) | |
[extractor/common] Support multiple properties in _og_search_property
| -rw-r--r-- | test/test_InfoExtractor.py | 3 | ||||
| -rw-r--r-- | youtube_dl/extractor/common.py | 9 |
2 files changed, 10 insertions, 2 deletions
diff --git a/test/test_InfoExtractor.py b/test/test_InfoExtractor.py index 88e8ff904..a98305c74 100644 --- a/test/test_InfoExtractor.py +++ b/test/test_InfoExtractor.py @@ -48,6 +48,9 @@ class TestInfoExtractor(unittest.TestCase): self.assertEqual(ie._og_search_property('foobar', html), 'Foo') self.assertEqual(ie._og_search_property('test1', html), 'foo > < bar') self.assertEqual(ie._og_search_property('test2', html), 'foo >//< bar') + self.assertEqual(ie._og_search_property(('test0', 'test1'), html), 'foo > < bar') + self.assertRaises(RegexNotFoundError, ie._og_search_property, 'test0', html, None, fatal=True) + self.assertRaises(RegexNotFoundError, ie._og_search_property, ('test0', 'test00'), html, None, fatal=True) def test_html_search_meta(self): ie = self.ie diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 53c28f016..3b6a5491d 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -727,9 +727,14 @@ class InfoExtractor(object): [^>]+?content=(["\'])(?P<content>.*?)\2''' % re.escape(prop) def _og_search_property(self, prop, html, name=None, **kargs): + if not isinstance(prop, (list, tuple)): + prop = [prop] if name is None: - name = 'OpenGraph %s' % prop - escaped = self._search_regex(self._og_regexes(prop), html, name, flags=re.DOTALL, **kargs) + name = 'OpenGraph %s' % prop[0] + og_regexes = [] + for p in prop: + og_regexes.extend(self._og_regexes(p)) + escaped = self._search_regex(og_regexes, html, name, flags=re.DOTALL, **kargs) if escaped is None: return None return unescapeHTML(escaped) |