[camdemy] Simplify and make more robust (#4938)

Do not throw errors if view count or upload date extraction fails.
Dispose of re.MULTILINE, which had absolutely no effect without any ^ or $ in sight.
Follow PEP8 naming conventions.
This commit is contained in:
Philipp Hagemeister 2015-02-12 08:55:06 +01:00
parent 024c53694d
commit 08b38d5401
2 changed files with 40 additions and 34 deletions

View File

@ -1,11 +1,18 @@
# coding: utf-8
from __future__ import unicode_literals
import datetime
import re
from .common import InfoExtractor
from ..compat import (compat_urllib_parse, compat_urlparse)
from ..utils import parse_iso8601
from ..compat import (
compat_urllib_parse,
compat_urlparse,
)
from ..utils import (
parse_iso8601,
str_to_int,
)
class CamdemyIE(InfoExtractor):
@ -23,6 +30,7 @@ class CamdemyIE(InfoExtractor):
'creator': 'ss11spring',
'upload_date': '20130114',
'timestamp': 1358154556,
'view_count': int,
}
}, {
# With non-empty description
@ -55,46 +63,43 @@ class CamdemyIE(InfoExtractor):
def _real_extract(self, url):
video_id = self._match_id(url)
page = self._download_webpage(url, video_id)
srcFrom = self._html_search_regex(
src_from = self._html_search_regex(
r"<div class='srcFrom'>Source: <a title='([^']+)'", page,
'external source', default=None)
if srcFrom:
return self.url_result(srcFrom)
if src_from:
return self.url_result(src_from)
oembed_obj = self._download_json(
'http://www.camdemy.com/oembed/?format=json&url=' + url, video_id)
thumb_url = oembed_obj['thumbnail_url']
video_folder = compat_urlparse.urljoin(thumb_url, 'video/')
fileListXML = self._download_xml(
file_list_doc = self._download_xml(
compat_urlparse.urljoin(video_folder, 'fileList.xml'),
video_id, 'Filelist XML')
fileName = fileListXML.find('./video/item/fileName').text
file_name = file_list_doc.find('./video/item/fileName').text
video_url = compat_urlparse.urljoin(video_folder, file_name)
creation_time = self._html_search_regex(
r"<div class='title'>Posted :</div>[\r\n ]*<div class='value'>([^<>]+)<",
page, 'creation time', flags=re.MULTILINE) + '+08:00'
creation_timestamp = parse_iso8601(creation_time, delimiter=' ')
view_count_str = self._html_search_regex(
r"<div class='title'>Views :</div>[\r\n ]*<div class='value'>([^<>]+)<",
page, 'view count', flags=re.MULTILINE)
views = int(view_count_str.replace(',', ''))
timestamp = parse_iso8601(self._html_search_regex(
r"<div class='title'>Posted\s*:</div>\s*<div class='value'>([^<>]+)<",
page, 'creation time', fatal=False),
delimiter=' ', timezone=datetime.timedelta(hours=8))
view_count = str_to_int(self._html_search_regex(
r"<div class='title'>Views\s*:</div>\s*<div class='value'>([^<>]+)<",
page, 'view count', fatal=False))
return {
'id': video_id,
'url': compat_urlparse.urljoin(video_folder, fileName),
'url': video_url,
'title': oembed_obj['title'],
'thumbnail': thumb_url,
'description': self._html_search_meta('description', page),
'creator': oembed_obj['author_name'],
'duration': oembed_obj['duration'],
'timestamp': creation_timestamp,
'view_count': views,
'timestamp': timestamp,
'view_count': view_count,
}

View File

@ -666,26 +666,27 @@ def https_open(self, req):
req, **kwargs)
def parse_iso8601(date_str, delimiter='T'):
def parse_iso8601(date_str, delimiter='T', timezone=None):
""" Return a UNIX timestamp from the given date """
if date_str is None:
return None
m = re.search(
r'(\.[0-9]+)?(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
date_str)
if not m:
timezone = datetime.timedelta()
else:
date_str = date_str[:-len(m.group(0))]
if not m.group('sign'):
if timezone is None:
m = re.search(
r'(\.[0-9]+)?(?:Z$| ?(?P<sign>\+|-)(?P<hours>[0-9]{2}):?(?P<minutes>[0-9]{2})$)',
date_str)
if not m:
timezone = datetime.timedelta()
else:
sign = 1 if m.group('sign') == '+' else -1
timezone = datetime.timedelta(
hours=sign * int(m.group('hours')),
minutes=sign * int(m.group('minutes')))
date_str = date_str[:-len(m.group(0))]
if not m.group('sign'):
timezone = datetime.timedelta()
else:
sign = 1 if m.group('sign') == '+' else -1
timezone = datetime.timedelta(
hours=sign * int(m.group('hours')),
minutes=sign * int(m.group('minutes')))
date_format = '%Y-%m-%d{0}%H:%M:%S'.format(delimiter)
dt = datetime.datetime.strptime(date_str, date_format) - timezone
return calendar.timegm(dt.timetuple())