]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/bandcamp.py
Prepare new release.
[youtubedl] / youtube_dl / extractor / bandcamp.py
index c13446665d2fb0e202973a26f5c7499c325719d1..056e06376667e02b34c8efa7b2565be51e4625a4 100644 (file)
@@ -1,13 +1,22 @@
 from __future__ import unicode_literals
 
 import json
+import random
 import re
+import time
 
 from .common import InfoExtractor
-from ..utils import (
+from ..compat import (
     compat_str,
     compat_urlparse,
+)
+from ..utils import (
     ExtractorError,
+    float_or_none,
+    int_or_none,
+    parse_filesize,
+    unescapeHTML,
+    update_url_query,
 )
 
 
@@ -25,7 +34,7 @@ class BandcampIE(InfoExtractor):
         '_skip': 'There is a limit of 200 free downloads / month for the test song'
     }, {
         'url': 'http://benprunty.bandcamp.com/track/lanius-battle',
-        'md5': '2b68e5851514c20efdff2afc5603b8b4',
+        'md5': '73d0b3171568232574e45652f8720b5c',
         'info_dict': {
             'id': '2650410135',
             'ext': 'mp3',
@@ -44,115 +53,198 @@ class BandcampIE(InfoExtractor):
             if m_trackinfo:
                 json_code = m_trackinfo.group(1)
                 data = json.loads(json_code)[0]
+                track_id = compat_str(data['id'])
+
+                if not data.get('file'):
+                    raise ExtractorError('Not streamable', video_id=track_id, expected=True)
 
                 formats = []
                 for format_id, format_url in data['file'].items():
                     ext, abr_str = format_id.split('-', 1)
                     formats.append({
                         'format_id': format_id,
-                        'url': format_url,
+                        'url': self._proto_relative_url(format_url, 'http:'),
                         'ext': ext,
                         'vcodec': 'none',
                         'acodec': ext,
-                        'abr': int(abr_str),
+                        'abr': int_or_none(abr_str),
                     })
 
                 self._sort_formats(formats)
 
                 return {
-                    'id': compat_str(data['id']),
+                    'id': track_id,
                     'title': data['title'],
                     'formats': formats,
-                    'duration': float(data['duration']),
+                    'duration': float_or_none(data.get('duration')),
                 }
             else:
                 raise ExtractorError('No free songs found')
 
         download_link = m_download.group(1)
         video_id = self._search_regex(
-            r'var TralbumData = {.*?id: (?P<id>\d+),?$',
-            webpage, 'video id', flags=re.MULTILINE | re.DOTALL)
-
-        download_webpage = self._download_webpage(download_link, video_id, 'Downloading free downloads page')
-        # We get the dictionary of the track from some javascript code
-        info = re.search(r'items: (.*?),$', download_webpage, re.MULTILINE).group(1)
-        info = json.loads(info)[0]
-        # We pick mp3-320 for now, until format selection can be easily implemented.
-        mp3_info = info['downloads']['mp3-320']
-        # If we try to use this url it says the link has expired
-        initial_url = mp3_info['url']
-        re_url = r'(?P<server>http://(.*?)\.bandcamp\.com)/download/track\?enc=mp3-320&fsig=(?P<fsig>.*?)&id=(?P<id>.*?)&ts=(?P<ts>.*)$'
-        m_url = re.match(re_url, initial_url)
-        #We build the url we will use to get the final track url
-        # This url is build in Bandcamp in the script download_bunde_*.js
-        request_url = '%s/statdownload/track?enc=mp3-320&fsig=%s&id=%s&ts=%s&.rand=665028774616&.vrs=1' % (m_url.group('server'), m_url.group('fsig'), video_id, m_url.group('ts'))
-        final_url_webpage = self._download_webpage(request_url, video_id, 'Requesting download url')
-        # If we could correctly generate the .rand field the url would be
-        #in the "download_url" key
-        final_url = re.search(r'"retry_url":"(.*?)"', final_url_webpage).group(1)
+            r'(?ms)var TralbumData = .*?[{,]\s*id: (?P<id>\d+),?$',
+            webpage, 'video id')
+
+        download_webpage = self._download_webpage(
+            download_link, video_id, 'Downloading free downloads page')
+
+        blob = self._parse_json(
+            self._search_regex(
+                r'data-blob=(["\'])(?P<blob>{.+?})\1', download_webpage,
+                'blob', group='blob'),
+            video_id, transform_source=unescapeHTML)
+
+        info = blob['digital_items'][0]
+
+        downloads = info['downloads']
+        track = info['title']
+
+        artist = info.get('artist')
+        title = '%s - %s' % (artist, track) if artist else track
+
+        download_formats = {}
+        for f in blob['download_formats']:
+            name, ext = f.get('name'), f.get('file_extension')
+            if all(isinstance(x, compat_str) for x in (name, ext)):
+                download_formats[name] = ext.strip('.')
+
+        formats = []
+        for format_id, f in downloads.items():
+            format_url = f.get('url')
+            if not format_url:
+                continue
+            # Stat URL generation algorithm is reverse engineered from
+            # download_*_bundle_*.js
+            stat_url = update_url_query(
+                format_url.replace('/download/', '/statdownload/'), {
+                    '.rand': int(time.time() * 1000 * random.random()),
+                })
+            format_id = f.get('encoding_name') or format_id
+            stat = self._download_json(
+                stat_url, video_id, 'Downloading %s JSON' % format_id,
+                transform_source=lambda s: s[s.index('{'):s.rindex('}') + 1],
+                fatal=False)
+            if not stat:
+                continue
+            retry_url = stat.get('retry_url')
+            if not isinstance(retry_url, compat_str):
+                continue
+            formats.append({
+                'url': self._proto_relative_url(retry_url, 'http:'),
+                'ext': download_formats.get(format_id),
+                'format_id': format_id,
+                'format_note': f.get('description'),
+                'filesize': parse_filesize(f.get('size_mb')),
+                'vcodec': 'none',
+            })
+        self._sort_formats(formats)
 
         return {
             'id': video_id,
-            'title': info['title'],
-            'ext': 'mp3',
-            'vcodec': 'none',
-            'url': final_url,
+            'title': title,
             'thumbnail': info.get('thumb_url'),
             'uploader': info.get('artist'),
+            'artist': artist,
+            'track': track,
+            'formats': formats,
         }
 
 
 class BandcampAlbumIE(InfoExtractor):
     IE_NAME = 'Bandcamp:album'
-    _VALID_URL = r'https?://(?:(?P<subdomain>[^.]+)\.)?bandcamp\.com(?:/album/(?P<title>[^?#]+))'
+    _VALID_URL = r'https?://(?:(?P<subdomain>[^.]+)\.)?bandcamp\.com(?:/album/(?P<album_id>[^?#]+)|/?(?:$|[?#]))'
 
     _TESTS = [{
         'url': 'http://blazo.bandcamp.com/album/jazz-format-mixtape-vol-1',
         'playlist': [
             {
-                'file': '1353101989.mp3',
                 'md5': '39bc1eded3476e927c724321ddf116cf',
                 'info_dict': {
+                    'id': '1353101989',
+                    'ext': 'mp3',
                     'title': 'Intro',
                 }
             },
             {
-                'file': '38097443.mp3',
                 'md5': '1a2c32e2691474643e912cc6cd4bffaa',
                 'info_dict': {
+                    'id': '38097443',
+                    'ext': 'mp3',
                     'title': 'Kero One - Keep It Alive (Blazo remix)',
                 }
             },
         ],
+        'info_dict': {
+            'title': 'Jazz Format Mixtape vol.1',
+            'id': 'jazz-format-mixtape-vol-1',
+            'uploader_id': 'blazo',
+        },
         'params': {
             'playlistend': 2
         },
-        'skip': 'Bandcamp imposes download limits. See test_playlists:test_bandcamp_album for the playlist test'
+        'skip': 'Bandcamp imposes download limits.'
     }, {
         'url': 'http://nightbringer.bandcamp.com/album/hierophany-of-the-open-grave',
         'info_dict': {
             'title': 'Hierophany of the Open Grave',
+            'uploader_id': 'nightbringer',
+            'id': 'hierophany-of-the-open-grave',
         },
         'playlist_mincount': 9,
+    }, {
+        'url': 'http://dotscale.bandcamp.com',
+        'info_dict': {
+            'title': 'Loom',
+            'id': 'dotscale',
+            'uploader_id': 'dotscale',
+        },
+        'playlist_mincount': 7,
+    }, {
+        # with escaped quote in title
+        'url': 'https://jstrecords.bandcamp.com/album/entropy-ep',
+        'info_dict': {
+            'title': '"Entropy" EP',
+            'uploader_id': 'jstrecords',
+            'id': 'entropy-ep',
+        },
+        'playlist_mincount': 3,
+    }, {
+        # not all tracks have songs
+        'url': 'https://insulters.bandcamp.com/album/we-are-the-plague',
+        'info_dict': {
+            'id': 'we-are-the-plague',
+            'title': 'WE ARE THE PLAGUE',
+            'uploader_id': 'insulters',
+        },
+        'playlist_count': 2,
     }]
 
     def _real_extract(self, url):
         mobj = re.match(self._VALID_URL, url)
-        playlist_id = mobj.group('subdomain')
-        title = mobj.group('title')
-        display_id = title or playlist_id
-        webpage = self._download_webpage(url, display_id)
-        tracks_paths = re.findall(r'<a href="(.*?)" itemprop="url">', webpage)
-        if not tracks_paths:
+        uploader_id = mobj.group('subdomain')
+        album_id = mobj.group('album_id')
+        playlist_id = album_id or uploader_id
+        webpage = self._download_webpage(url, playlist_id)
+        track_elements = re.findall(
+            r'(?s)<div[^>]*>(.*?<a[^>]+href="([^"]+?)"[^>]+itemprop="url"[^>]*>.*?)</div>', webpage)
+        if not track_elements:
             raise ExtractorError('The page doesn\'t contain any tracks')
+        # Only tracks with duration info have songs
         entries = [
             self.url_result(compat_urlparse.urljoin(url, t_path), ie=BandcampIE.ie_key())
-            for t_path in tracks_paths]
-        title = self._search_regex(r'album_title : "(.*?)"', webpage, 'title')
+            for elem_content, t_path in track_elements
+            if self._html_search_meta('duration', elem_content, default=None)]
+
+        title = self._html_search_regex(
+            r'album_title\s*:\s*"((?:\\.|[^"\\])+?)"',
+            webpage, 'title', fatal=False)
+        if title:
+            title = title.replace(r'\"', '"')
         return {
             '_type': 'playlist',
+            'uploader_id': uploader_id,
             'id': playlist_id,
-            'display_id': display_id,
             'title': title,
             'entries': entries,
         }