[appletrailers] Rework extraction (fixes #1387)

The exraction was broken:
* The includes page contains img elements that need to be fixed.
* Use the 'itunes.inc' page, it contains a json dictionary for each trailer with information.
* Get the formats from 'includes/settings{trailer_name}.json'
* Use urljoin to allow urls with a fragment identifier to work

Removed the thumbnail urls from the tests, they are different now.
This commit is contained in:
Jaime Marquínez Ferrándiz 2013-09-29 20:49:58 +02:00
parent 138a5454b5
commit 843530568f

View File

@ -1,8 +1,10 @@
import re
import xml.etree.ElementTree
import json
from .common import InfoExtractor
from ..utils import (
compat_urlparse,
determine_ext,
)
@ -14,10 +16,9 @@ class AppleTrailersIE(InfoExtractor):
u"playlist": [
{
u"file": u"manofsteel-trailer4.mov",
u"md5": u"11874af099d480cc09e103b189805d5f",
u"md5": u"d97a8e575432dbcb81b7c3acb741f8a8",
u"info_dict": {
u"duration": 111,
u"thumbnail": u"http://trailers.apple.com/trailers/wb/manofsteel/images/thumbnail_11624.jpg",
u"title": u"Trailer 4",
u"upload_date": u"20130523",
u"uploader_id": u"wb",
@ -25,10 +26,9 @@ class AppleTrailersIE(InfoExtractor):
},
{
u"file": u"manofsteel-trailer3.mov",
u"md5": u"07a0a262aae5afe68120eed61137ab34",
u"md5": u"b8017b7131b721fb4e8d6f49e1df908c",
u"info_dict": {
u"duration": 182,
u"thumbnail": u"http://trailers.apple.com/trailers/wb/manofsteel/images/thumbnail_10793.jpg",
u"title": u"Trailer 3",
u"upload_date": u"20130417",
u"uploader_id": u"wb",
@ -36,10 +36,9 @@ class AppleTrailersIE(InfoExtractor):
},
{
u"file": u"manofsteel-trailer.mov",
u"md5": u"e401fde0813008e3307e54b6f384cff1",
u"md5": u"d0f1e1150989b9924679b441f3404d48",
u"info_dict": {
u"duration": 148,
u"thumbnail": u"http://trailers.apple.com/trailers/wb/manofsteel/images/thumbnail_8703.jpg",
u"title": u"Trailer",
u"upload_date": u"20121212",
u"uploader_id": u"wb",
@ -47,10 +46,9 @@ class AppleTrailersIE(InfoExtractor):
},
{
u"file": u"manofsteel-teaser.mov",
u"md5": u"76b392f2ae9e7c98b22913c10a639c97",
u"md5": u"5fe08795b943eb2e757fa95cb6def1cb",
u"info_dict": {
u"duration": 93,
u"thumbnail": u"http://trailers.apple.com/trailers/wb/manofsteel/images/thumbnail_6899.jpg",
u"title": u"Teaser",
u"upload_date": u"20120721",
u"uploader_id": u"wb",
@ -59,87 +57,61 @@ class AppleTrailersIE(InfoExtractor):
]
}
_JSON_RE = r'iTunes.playURL\((.*?)\);'
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
movie = mobj.group('movie')
uploader_id = mobj.group('company')
playlist_url = url.partition(u'?')[0] + u'/includes/playlists/web.inc'
playlist_url = compat_urlparse.urljoin(url, u'includes/playlists/itunes.inc')
playlist_snippet = self._download_webpage(playlist_url, movie)
playlist_cleaned = re.sub(r'(?s)<script>.*?</script>', u'', playlist_snippet)
playlist_cleaned = re.sub(r'(?s)<script[^<]*?>.*?</script>', u'', playlist_snippet)
playlist_cleaned = re.sub(r'<img ([^<]*?)>', r'<img \1/>', playlist_cleaned)
# The ' in the onClick attributes are not escaped, it couldn't be parsed
# with xml.etree.ElementTree.fromstring
# like: http://trailers.apple.com/trailers/wb/gravity/
def _clean_json(m):
return u'iTunes.playURL(%s);' % m.group(1).replace('\'', '&#39;')
playlist_cleaned = re.sub(self._JSON_RE, _clean_json, playlist_cleaned)
playlist_html = u'<html>' + playlist_cleaned + u'</html>'
size_cache = {}
doc = xml.etree.ElementTree.fromstring(playlist_html)
playlist = []
for li in doc.findall('./div/ul/li'):
title = li.find('.//h3').text
on_click = li.find('.//a').attrib['onClick']
trailer_info_json = self._search_regex(self._JSON_RE,
on_click, u'trailer info')
trailer_info = json.loads(trailer_info_json)
title = trailer_info['title']
video_id = movie + '-' + re.sub(r'[^a-zA-Z0-9]', '', title).lower()
thumbnail = li.find('.//img').attrib['src']
upload_date = trailer_info['posted'].replace('-', '')
date_el = li.find('.//p')
upload_date = None
m = re.search(r':\s?(?P<month>[0-9]{2})/(?P<day>[0-9]{2})/(?P<year>[0-9]{2})', date_el.text)
if m:
upload_date = u'20' + m.group('year') + m.group('month') + m.group('day')
runtime_el = date_el.find('./br')
m = re.search(r':\s?(?P<minutes>[0-9]+):(?P<seconds>[0-9]{1,2})', runtime_el.tail)
runtime = trailer_info['runtime']
m = re.search(r'(?P<minutes>[0-9]+):(?P<seconds>[0-9]{1,2})', runtime)
duration = None
if m:
duration = 60 * int(m.group('minutes')) + int(m.group('seconds'))
first_url = trailer_info['url']
trailer_id = first_url.split('/')[-1].rpartition('_')[0]
settings_json_url = compat_urlparse.urljoin(url, 'includes/settings/%s.json' % trailer_id)
settings_json = self._download_webpage(settings_json_url, trailer_id, u'Downloading settings json')
settings = json.loads(settings_json)
formats = []
for formats_el in li.findall('.//a'):
if formats_el.attrib['class'] != 'OverlayPanel':
continue
target = formats_el.attrib['target']
format_code = formats_el.text
if 'Automatic' in format_code:
continue
size_q = formats_el.attrib['href']
size_id = size_q.rpartition('#videos-')[2]
if size_id not in size_cache:
size_url = url + size_q
sizepage_html = self._download_webpage(
size_url, movie,
note=u'Downloading size info %s' % size_id,
errnote=u'Error while downloading size info %s' % size_id,
)
_doc = xml.etree.ElementTree.fromstring(sizepage_html)
size_cache[size_id] = _doc
sizepage_doc = size_cache[size_id]
links = sizepage_doc.findall('.//{http://www.w3.org/1999/xhtml}ul/{http://www.w3.org/1999/xhtml}li/{http://www.w3.org/1999/xhtml}a')
for vid_a in links:
href = vid_a.get('href')
if not href.endswith(target):
continue
detail_q = href.partition('#')[0]
detail_url = url + '/' + detail_q
m = re.match(r'includes/(?P<detail_id>[^/]+)/', detail_q)
detail_id = m.group('detail_id')
detail_html = self._download_webpage(
detail_url, movie,
note=u'Downloading detail %s %s' % (detail_id, size_id),
errnote=u'Error while downloading detail %s %s' % (detail_id, size_id)
)
detail_doc = xml.etree.ElementTree.fromstring(detail_html)
movie_link_el = detail_doc.find('.//{http://www.w3.org/1999/xhtml}a')
assert movie_link_el.get('class') == 'movieLink'
movie_link = movie_link_el.get('href').partition('?')[0].replace('_', '_h')
ext = determine_ext(movie_link)
assert ext == 'mov'
formats.append({
'format': format_code,
'ext': ext,
'url': movie_link,
})
for format in settings['metadata']['sizes']:
# The src is a file pointing to the real video file
format_url = re.sub(r'_(\d*p.mov)', r'_h\1', format['src'])
formats.append({
'url': format_url,
'ext': determine_ext(format_url),
'format': format['type'],
'width': format['width'],
'height': int(format['height']),
})
formats = sorted(formats, key=lambda f: (f['height'], f['width']))
info = {
'_type': 'video',