X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fcrunchyroll.py;h=4903764f7008ec6f22c054cda2c73557c44b2a91;hb=78272a076e96bea48525e415e42fa4d95d4d66fa;hp=a20b88f0250279abe35b98a6a887b6be11298ec5;hpb=344400951ccddaaa477738957bdbdb86a704c55b;p=youtube-dl diff --git a/youtube_dl/extractor/crunchyroll.py b/youtube_dl/extractor/crunchyroll.py index a20b88f02..4903764f7 100644 --- a/youtube_dl/extractor/crunchyroll.py +++ b/youtube_dl/extractor/crunchyroll.py @@ -2,8 +2,10 @@ from __future__ import unicode_literals import re +import json import base64 import zlib +import xml.etree.ElementTree from hashlib import sha1 from math import pow, sqrt, floor @@ -16,6 +18,7 @@ from ..utils import ( intlist_to_bytes, unified_strdate, clean_html, + urlencode_postdata, ) from ..aes import ( aes_cbc_decrypt, @@ -50,6 +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) @@ -85,9 +108,8 @@ class CrunchyrollIE(InfoExtractor): return zlib.decompress(decrypted_data) def _convert_subtitles_to_srt(self, subtitles): - i = 1 output = '' - for start, end, text in re.findall(r']*?start="([^"]+)" [^>]*?end="([^"]+)" [^>]*?text="([^"]+)"[^>]*?>', subtitles): + for i, (start, end, text) in enumerate(re.findall(r']*?start="([^"]+)" [^>]*?end="([^"]+)" [^>]*?text="([^"]+)"[^>]*?>', subtitles), 1): start = start.replace('.', ',') end = end.replace('.', ',') text = clean_html(text) @@ -95,7 +117,75 @@ class CrunchyrollIE(InfoExtractor): if not text: continue output += '%d\n%s --> %s\n%s\n\n' % (i, start, end, text) - i += 1 + 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): @@ -113,6 +203,12 @@ class CrunchyrollIE(InfoExtractor): if note_m: raise ExtractorError(note_m) + mobj = re.search(r'Page\.messaging_box_controller\.addItems\(\[(?P{.+?})\]\)', webpage) + if mobj: + msg = json.loads(mobj.group('msg')) + if msg.get('type') == 'error': + raise ExtractorError('crunchyroll returned error: %s' % msg['message_body'], expected=True) + video_title = self._html_search_regex(r']*>(.+?)', webpage, 'video_title', flags=re.DOTALL) video_title = re.sub(r' {2,}', ' ', video_title) video_description = self._html_search_regex(r'"description":"([^"]+)', webpage, 'video_description', default='') @@ -153,6 +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) @@ -166,10 +263,13 @@ class CrunchyrollIE(InfoExtractor): data = base64.b64decode(data) subtitle = self._decrypt_subtitles(data, iv, id).decode('utf-8') - lang_code = self._search_regex(r'lang_code=\'([^\']+)', subtitle, 'subtitle_lang_code', fatal=False) + lang_code = self._search_regex(r'lang_code=["\']([^"\']+)', subtitle, 'subtitle_lang_code', fatal=False) if not lang_code: continue - subtitles[lang_code] = self._convert_subtitles_to_srt(subtitle) + if sub_format == 'ass': + subtitles[lang_code] = self._convert_subtitles_to_ass(subtitle) + else: + subtitles[lang_code] = self._convert_subtitles_to_srt(subtitle) return { 'id': video_id, @@ -180,4 +280,4 @@ class CrunchyrollIE(InfoExtractor): 'upload_date': video_upload_date, 'subtitles': subtitles, 'formats': formats, - } \ No newline at end of file + }