diff options
| author | Jaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com> | 2014-12-16 15:48:01 +0100 | 
|---|---|---|
| committer | Jaime Marquínez Ferrándiz <jaime.marquinez.ferrandiz@gmail.com> | 2014-12-16 15:48:01 +0100 | 
| commit | bb18d787b5c0ded4865090e8022338db063520b2 (patch) | |
| tree | 4d910f997693fd24b027fce054c661377a7b7036 | |
| parent | 734ea11e3c57ca4df7d9cfc475b1b99b56c18034 (diff) | |
[aljazeera] Add extractor (closes #4487)
| -rw-r--r-- | youtube_dl/extractor/__init__.py | 1 | ||||
| -rw-r--r-- | youtube_dl/extractor/aljazeera.py | 32 | 
2 files changed, 33 insertions, 0 deletions
diff --git a/youtube_dl/extractor/__init__.py b/youtube_dl/extractor/__init__.py index caf8c04f1..b091505db 100644 --- a/youtube_dl/extractor/__init__.py +++ b/youtube_dl/extractor/__init__.py @@ -5,6 +5,7 @@ from .academicearth import AcademicEarthCourseIE  from .addanime import AddAnimeIE  from .adultswim import AdultSwimIE  from .aftonbladet import AftonbladetIE +from .aljazeera import AlJazeeraIE  from .anitube import AnitubeIE  from .anysex import AnySexIE  from .aol import AolIE diff --git a/youtube_dl/extractor/aljazeera.py b/youtube_dl/extractor/aljazeera.py new file mode 100644 index 000000000..ace5ec19d --- /dev/null +++ b/youtube_dl/extractor/aljazeera.py @@ -0,0 +1,32 @@ +from .common import InfoExtractor + + +class AlJazeeraIE(InfoExtractor): +    _VALID_URL = r'http://www\.aljazeera\.com/programmes/.*?/(?P<id>[^/]+)\.html' + +    _TEST = { +        'url': 'http://www.aljazeera.com/programmes/the-slum/2014/08/deliverance-201482883754237240.html', +        'info_dict': { +            'id': '3792260579001', +            'ext': 'mp4', +            'title': 'The Slum - Episode 1: Deliverance', +            'description': 'As a birth attendant advocating for family planning, Remy is on the frontline of Tondo\'s battle with overcrowding.', +            'uploader': 'Al Jazeera English', +        }, +        'add_ie': ['Brightcove'], +    } + +    def _real_extract(self, url): +        program_name = self._match_id(url) +        webpage = self._download_webpage(url, program_name) +        brightcove_id = self._search_regex( +            r'RenderPagesVideo\(\'(.+?)\'', webpage, 'brightcove id') + +        return { +            '_type': 'url', +            'url': +                'http://c.brightcove.com/services/viewer/federated_f9?' +                '&playerKey=AQ~~%2CAAAAmtVJIFk~%2CTVGOQ5ZTwJbeMWnq5d_H4MOM57xfzApc' +                '&%40videoPlayer={0}'.format(brightcove_id), +            'ie_key': 'Brightcove', +        }  | 
