]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/nytimes.py
Set upstream metadata fields: Bug-Database, Bug-Submit, Repository, Repository-Browse.
[youtubedl] / youtube_dl / extractor / nytimes.py
index 7bf105d38178a5d59339d81262649b40fa57a46e..fc78ca56c90d37b00c1f396aee7c896d54fb91c9 100644 (file)
+# coding: utf-8
 from __future__ import unicode_literals
 
-import re
+import hmac
+import hashlib
+import base64
 
 from .common import InfoExtractor
-from ..utils import parse_iso8601
+from ..utils import (
+    determine_ext,
+    float_or_none,
+    int_or_none,
+    js_to_json,
+    mimetype2ext,
+    parse_iso8601,
+    remove_start,
+)
 
 
-class NYTimesIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?nytimes\.com/video/(?:[^/]+/)+(?P<id>\d+)'
+class NYTimesBaseIE(InfoExtractor):
+    _SECRET = b'pX(2MbU2);4N{7J8)>YwKRJ+/pQ3JkiU2Q^V>mFYv6g6gYvt6v'
 
-    _TEST = {
+    def _extract_video_from_id(self, video_id):
+        # Authorization generation algorithm is reverse engineered from `signer` in
+        # http://graphics8.nytimes.com/video/vhs/vhs-2.x.min.js
+        path = '/svc/video/api/v3/video/' + video_id
+        hm = hmac.new(self._SECRET, (path + ':vhs').encode(), hashlib.sha512).hexdigest()
+        video_data = self._download_json('http://www.nytimes.com' + path, video_id, 'Downloading video JSON', headers={
+            'Authorization': 'NYTV ' + base64.b64encode(hm.encode()).decode(),
+            'X-NYTV': 'vhs',
+        }, fatal=False)
+        if not video_data:
+            video_data = self._download_json(
+                'http://www.nytimes.com/svc/video/api/v2/video/' + video_id,
+                video_id, 'Downloading video JSON')
+
+        title = video_data['headline']
+
+        def get_file_size(file_size):
+            if isinstance(file_size, int):
+                return file_size
+            elif isinstance(file_size, dict):
+                return int(file_size.get('value', 0))
+            else:
+                return None
+
+        urls = []
+        formats = []
+        for video in video_data.get('renditions', []):
+            video_url = video.get('url')
+            format_id = video.get('type')
+            if not video_url or format_id == 'thumbs' or video_url in urls:
+                continue
+            urls.append(video_url)
+            ext = mimetype2ext(video.get('mimetype')) or determine_ext(video_url)
+            if ext == 'm3u8':
+                formats.extend(self._extract_m3u8_formats(
+                    video_url, video_id, 'mp4', 'm3u8_native',
+                    m3u8_id=format_id or 'hls', fatal=False))
+            elif ext == 'mpd':
+                continue
+            #     formats.extend(self._extract_mpd_formats(
+            #         video_url, video_id, format_id or 'dash', fatal=False))
+            else:
+                formats.append({
+                    'url': video_url,
+                    'format_id': format_id,
+                    'vcodec': video.get('videoencoding') or video.get('video_codec'),
+                    'width': int_or_none(video.get('width')),
+                    'height': int_or_none(video.get('height')),
+                    'filesize': get_file_size(video.get('file_size') or video.get('fileSize')),
+                    'tbr': int_or_none(video.get('bitrate'), 1000) or None,
+                    'ext': ext,
+                })
+        self._sort_formats(formats, ('height', 'width', 'filesize', 'tbr', 'fps', 'format_id'))
+
+        thumbnails = []
+        for image in video_data.get('images', []):
+            image_url = image.get('url')
+            if not image_url:
+                continue
+            thumbnails.append({
+                'url': 'http://www.nytimes.com/' + image_url,
+                'width': int_or_none(image.get('width')),
+                'height': int_or_none(image.get('height')),
+            })
+
+        publication_date = video_data.get('publication_date')
+        timestamp = parse_iso8601(publication_date[:-8]) if publication_date else None
+
+        return {
+            'id': video_id,
+            'title': title,
+            'description': video_data.get('summary'),
+            'timestamp': timestamp,
+            'uploader': video_data.get('byline'),
+            'duration': float_or_none(video_data.get('duration'), 1000),
+            'formats': formats,
+            'thumbnails': thumbnails,
+        }
+
+
+class NYTimesIE(NYTimesBaseIE):
+    _VALID_URL = r'https?://(?:(?:www\.)?nytimes\.com/video/(?:[^/]+/)+?|graphics8\.nytimes\.com/bcvideo/\d+(?:\.\d+)?/iframe/embed\.html\?videoId=)(?P<id>\d+)'
+
+    _TESTS = [{
         'url': 'http://www.nytimes.com/video/opinion/100000002847155/verbatim-what-is-a-photocopier.html?playlistId=100000001150263',
-        'md5': '18a525a510f942ada2720db5f31644c0',
+        'md5': 'd665342765db043f7e225cff19df0f2d',
         'info_dict': {
             'id': '100000002847155',
             'ext': 'mov',
@@ -22,56 +116,108 @@ class NYTimesIE(InfoExtractor):
             'uploader': 'Brett Weiner',
             'duration': 419,
         }
-    }
+    }, {
+        'url': 'http://www.nytimes.com/video/travel/100000003550828/36-hours-in-dubai.html',
+        'only_matching': True,
+    }]
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id')
+        video_id = self._match_id(url)
 
-        video_data = self._download_json(
-            'http://www.nytimes.com/svc/video/api/v2/video/%s' % video_id, video_id, 'Downloading video JSON')
+        return self._extract_video_from_id(video_id)
 
-        title = video_data['headline']
-        description = video_data['summary']
-        duration = video_data['duration'] / 1000.0
 
-        uploader = video_data['byline']
-        timestamp = parse_iso8601(video_data['publication_date'][:-8])
+class NYTimesArticleIE(NYTimesBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?nytimes\.com/(.(?<!video))*?/(?:[^/]+/)*(?P<id>[^.]+)(?:\.html)?'
+    _TESTS = [{
+        'url': 'http://www.nytimes.com/2015/04/14/business/owner-of-gravity-payments-a-credit-card-processor-is-setting-a-new-minimum-wage-70000-a-year.html?_r=0',
+        'md5': 'e2076d58b4da18e6a001d53fd56db3c9',
+        'info_dict': {
+            'id': '100000003628438',
+            'ext': 'mov',
+            'title': 'New Minimum Wage: $70,000 a Year',
+            'description': 'Dan Price, C.E.O. of Gravity Payments, surprised his 120-person staff by announcing that he planned over the next three years to raise the salary of every employee to $70,000 a year.',
+            'timestamp': 1429033037,
+            'upload_date': '20150414',
+            'uploader': 'Matthew Williams',
+        }
+    }, {
+        'url': 'http://www.nytimes.com/2016/10/14/podcasts/revelations-from-the-final-weeks.html',
+        'md5': 'e0d52040cafb07662acf3c9132db3575',
+        'info_dict': {
+            'id': '100000004709062',
+            'title': 'The Run-Up: ‘He Was Like an Octopus’',
+            'ext': 'mp3',
+            'description': 'md5:fb5c6b93b12efc51649b4847fe066ee4',
+            'series': 'The Run-Up',
+            'episode': '‘He Was Like an Octopus’',
+            'episode_number': 20,
+            'duration': 2130,
+        }
+    }, {
+        'url': 'http://www.nytimes.com/2016/10/16/books/review/inside-the-new-york-times-book-review-the-rise-of-hitler.html',
+        'info_dict': {
+            'id': '100000004709479',
+            'title': 'The Rise of Hitler',
+            'ext': 'mp3',
+            'description': 'md5:bce877fd9e3444990cb141875fab0028',
+            'creator': 'Pamela Paul',
+            'duration': 3475,
+        },
+        'params': {
+            'skip_download': True,
+        },
+    }, {
+        'url': 'http://www.nytimes.com/news/minute/2014/03/17/times-minute-whats-next-in-crimea/?_php=true&_type=blogs&_php=true&_type=blogs&_r=1',
+        'only_matching': True,
+    }]
+
+    def _extract_podcast_from_json(self, json, page_id, webpage):
+        podcast_audio = self._parse_json(
+            json, page_id, transform_source=js_to_json)
 
-        def get_file_size(file_size):
-            if isinstance(file_size, int):
-                return file_size
-            elif isinstance(file_size, dict):
-                return int(file_size.get('value', 0))
-            else:
-                return 0
-
-        formats = [
-            {
-                'url': video['url'],
-                'format_id': video['type'],
-                'vcodec': video['video_codec'],
-                'width': video['width'],
-                'height': video['height'],
-                'filesize': get_file_size(video['fileSize']),
-            } for video in video_data['renditions']
-        ]
-        self._sort_formats(formats)
-
-        thumbnails = [
-            {
-                'url': 'http://www.nytimes.com/%s' % image['url'],
-                'resolution': '%dx%d' % (image['width'], image['height']),
-            } for image in video_data['images']
-        ]
+        audio_data = podcast_audio['data']
+        track = audio_data['track']
+
+        episode_title = track['title']
+        video_url = track['source']
+
+        description = track.get('description') or self._html_search_meta(
+            ['og:description', 'twitter:description'], webpage)
+
+        podcast_title = audio_data.get('podcast', {}).get('title')
+        title = ('%s: %s' % (podcast_title, episode_title)
+                 if podcast_title else episode_title)
+
+        episode = audio_data.get('podcast', {}).get('episode') or ''
+        episode_number = int_or_none(self._search_regex(
+            r'[Ee]pisode\s+(\d+)', episode, 'episode number', default=None))
 
         return {
-            'id': video_id,
+            'id': remove_start(podcast_audio.get('target'), 'FT') or page_id,
+            'url': video_url,
             'title': title,
             'description': description,
-            'timestamp': timestamp,
-            'uploader': uploader,
-            'duration': duration,
-            'formats': formats,
-            'thumbnails': thumbnails,
-        }
\ No newline at end of file
+            'creator': track.get('credit'),
+            'series': podcast_title,
+            'episode': episode_title,
+            'episode_number': episode_number,
+            'duration': int_or_none(track.get('duration')),
+        }
+
+    def _real_extract(self, url):
+        page_id = self._match_id(url)
+
+        webpage = self._download_webpage(url, page_id)
+
+        video_id = self._search_regex(
+            r'data-videoid=["\'](\d+)', webpage, 'video id',
+            default=None, fatal=False)
+        if video_id is not None:
+            return self._extract_video_from_id(video_id)
+
+        podcast_data = self._search_regex(
+            (r'NYTD\.FlexTypes\.push\s*\(\s*({.+?})\s*\)\s*;\s*</script',
+             r'NYTD\.FlexTypes\.push\s*\(\s*({.+})\s*\)\s*;'),
+            webpage, 'podcast data')
+        return self._extract_podcast_from_json(podcast_data, page_id, webpage)