Merge branch 'master' of github.com:rg3/youtube-dl

This commit is contained in:
Philipp Hagemeister 2014-08-31 23:44:06 +02:00
commit d10548b691
3 changed files with 165 additions and 1 deletions

View File

@ -323,6 +323,7 @@ from .stanfordoc import StanfordOpenClassroomIE
from .steam import SteamIE
from .streamcloud import StreamcloudIE
from .streamcz import StreamCZIE
from .sunporno import SunPornoIE
from .swrmediathek import SWRMediathekIE
from .syfy import SyfyIE
from .sztvhu import SztvHuIE

View File

@ -5,6 +5,7 @@ import re
import json
import base64
import zlib
import xml.etree.ElementTree
from hashlib import sha1
from math import pow, sqrt, floor
@ -17,6 +18,7 @@ from ..utils import (
intlist_to_bytes,
unified_strdate,
clean_html,
urlencode_postdata,
)
from ..aes import (
aes_cbc_decrypt,
@ -51,6 +53,26 @@ class CrunchyrollIE(InfoExtractor):
'1080': ('80', '108'),
}
def _login(self):
(username, password) = self._get_login_info()
if username is None:
return
self.report_login()
login_url = 'https://www.crunchyroll.com/?a=formhandler'
data = urlencode_postdata({
'formname': 'RpcApiUser_Login',
'name': username,
'password': password,
})
login_request = compat_urllib_request.Request(login_url, data)
login_request.add_header('Content-Type', 'application/x-www-form-urlencoded')
self._download_webpage(login_request, None, False, 'Wrong login info')
def _real_initialize(self):
self._login()
def _decrypt_subtitles(self, data, iv, id):
data = bytes_to_intlist(data)
iv = bytes_to_intlist(iv)
@ -97,6 +119,75 @@ class CrunchyrollIE(InfoExtractor):
output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text)
return output
def _convert_subtitles_to_ass(self, subtitles):
output = ''
def ass_bool(strvalue):
assvalue = '0'
if strvalue == '1':
assvalue = '-1'
return assvalue
sub_root = xml.etree.ElementTree.fromstring(subtitles)
if not sub_root:
return output
output = '[Script Info]\n'
output += 'Title: %s\n' % sub_root.attrib["title"]
output += 'ScriptType: v4.00+\n'
output += 'WrapStyle: %s\n' % sub_root.attrib["wrap_style"]
output += 'PlayResX: %s\n' % sub_root.attrib["play_res_x"]
output += 'PlayResY: %s\n' % sub_root.attrib["play_res_y"]
output += """ScaledBorderAndShadow: yes
[V4+ Styles]
Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
"""
for style in sub_root.findall('./styles/style'):
output += 'Style: ' + style.attrib["name"]
output += ',' + style.attrib["font_name"]
output += ',' + style.attrib["font_size"]
output += ',' + style.attrib["primary_colour"]
output += ',' + style.attrib["secondary_colour"]
output += ',' + style.attrib["outline_colour"]
output += ',' + style.attrib["back_colour"]
output += ',' + ass_bool(style.attrib["bold"])
output += ',' + ass_bool(style.attrib["italic"])
output += ',' + ass_bool(style.attrib["underline"])
output += ',' + ass_bool(style.attrib["strikeout"])
output += ',' + style.attrib["scale_x"]
output += ',' + style.attrib["scale_y"]
output += ',' + style.attrib["spacing"]
output += ',' + style.attrib["angle"]
output += ',' + style.attrib["border_style"]
output += ',' + style.attrib["outline"]
output += ',' + style.attrib["shadow"]
output += ',' + style.attrib["alignment"]
output += ',' + style.attrib["margin_l"]
output += ',' + style.attrib["margin_r"]
output += ',' + style.attrib["margin_v"]
output += ',' + style.attrib["encoding"]
output += '\n'
output += """
[Events]
Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
"""
for event in sub_root.findall('./events/event'):
output += 'Dialogue: 0'
output += ',' + event.attrib["start"]
output += ',' + event.attrib["end"]
output += ',' + event.attrib["style"]
output += ',' + event.attrib["name"]
output += ',' + event.attrib["margin_l"]
output += ',' + event.attrib["margin_r"]
output += ',' + event.attrib["margin_v"]
output += ',' + event.attrib["effect"]
output += ',' + event.attrib["text"]
output += '\n'
return output
def _real_extract(self,url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('video_id')
@ -158,6 +249,7 @@ class CrunchyrollIE(InfoExtractor):
})
subtitles = {}
sub_format = self._downloader.params.get('subtitlesformat', 'srt')
for sub_id, sub_name in re.findall(r'\?ssid=([0-9]+)" title="([^"]+)', webpage):
sub_page = self._download_webpage('http://www.crunchyroll.com/xml/?req=RpcApiSubtitle_GetXml&subtitle_script_id='+sub_id,\
video_id, note='Downloading subtitles for '+sub_name)
@ -174,6 +266,9 @@ class CrunchyrollIE(InfoExtractor):
lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False)
if not lang_code:
continue
if sub_format == 'ass':
subtitles[lang_code] = self._convert_subtitles_to_ass(subtitle)
else:
subtitles[lang_code] = self._convert_subtitles_to_srt(subtitle)
return {

View File

@ -0,0 +1,68 @@
from __future__ import unicode_literals
import re
from .common import InfoExtractor
from ..utils import (
parse_duration,
int_or_none,
qualities,
determine_ext,
)
class SunPornoIE(InfoExtractor):
_VALID_URL = r'https?://(?:www\.)?sunporno\.com/videos/(?P<id>\d+)'
_TEST = {
'url': 'http://www.sunporno.com/videos/807778/',
'md5': '6457d3c165fd6de062b99ef6c2ff4c86',
'info_dict': {
'id': '807778',
'ext': 'flv',
'title': 'md5:0a400058e8105d39e35c35e7c5184164',
'description': 'md5:a31241990e1bd3a64e72ae99afb325fb',
'thumbnail': 're:^https?://.*\.jpg$',
'duration': 302,
}
}
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
webpage = self._download_webpage(url, video_id)
title = self._html_search_regex(r'<title>([^<]+)</title>', webpage, 'title')
description = self._html_search_meta('description', webpage, 'description')
thumbnail = self._html_search_regex(
r'poster="([^"]+)"', webpage, 'thumbnail', fatal=False)
duration = parse_duration(self._search_regex(
r'<span>Duration: (\d+:\d+)</span>', webpage, 'duration', fatal=False))
view_count = int_or_none(self._html_search_regex(
r'<span class="views">(\d+)</span>', webpage, 'view count', fatal=False))
comment_count = int_or_none(self._html_search_regex(
r'(\d+)</b> Comments?', webpage, 'comment count', fatal=False))
formats = []
quality = qualities(['mp4', 'flv'])
for video_url in re.findall(r'<source src="([^"]+)"', webpage):
video_ext = determine_ext(video_url)
formats.append({
'url': video_url,
'format_id': video_ext,
'quality': quality(video_ext),
})
self._sort_formats(formats)
return {
'id': video_id,
'title': title,
'description': description,
'thumbnail': thumbnail,
'duration': duration,
'view_count': view_count,
'comment_count': comment_count,
'formats': formats,
}