]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/openload.py
Merge tag 'upstream/2016.12.01'
[youtubedl] / youtube_dl / extractor / openload.py
index 4e80ca9ff03100e244b157f70c8ff6897256cd55..7f19b1ba5c3c355977c694334694b51f71a9840c 100644 (file)
@@ -1,19 +1,25 @@
 # coding: utf-8
 from __future__ import unicode_literals, division
 
-import math
+import re
 
 from .common import InfoExtractor
-from ..compat import compat_chr
+from ..compat import (
+    compat_chr,
+    compat_ord,
+)
 from ..utils import (
-    decode_png,
     determine_ext,
     ExtractorError,
 )
+from ..jsinterp import (
+    JSInterpreter,
+    _NAME_RE
+)
 
 
 class OpenloadIE(InfoExtractor):
-    _VALID_URL = r'https://openload.(?:co|io)/(?:f|embed)/(?P<id>[a-zA-Z0-9-_]+)'
+    _VALID_URL = r'https?://openload\.(?:co|io)/(?:f|embed)/(?P<id>[a-zA-Z0-9-_]+)'
 
     _TESTS = [{
         'url': 'https://openload.co/f/kUEfGclsU9o',
@@ -24,6 +30,22 @@ class OpenloadIE(InfoExtractor):
             'title': 'skyrim_no-audio_1080.mp4',
             'thumbnail': 're:^https?://.*\.jpg$',
         },
+    }, {
+        'url': 'https://openload.co/embed/rjC09fkPLYs',
+        'info_dict': {
+            'id': 'rjC09fkPLYs',
+            'ext': 'mp4',
+            'title': 'movie.mp4',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'subtitles': {
+                'en': [{
+                    'ext': 'vtt',
+                }],
+            },
+        },
+        'params': {
+            'skip_download': True,  # test subtitles only
+        },
     }, {
         'url': 'https://openload.co/embed/kUEfGclsU9o/skyrim_no-audio_1080.mp4',
         'only_matching': True,
@@ -40,84 +62,98 @@ class OpenloadIE(InfoExtractor):
         'only_matching': True,
     }]
 
+    def openload_decode(self, txt):
+        symbol_dict = {
+            '(゚Д゚) [゚Θ゚]': '_',
+            '(゚Д゚) [゚ω゚ノ]': 'a',
+            '(゚Д゚) [゚Θ゚ノ]': 'b',
+            '(゚Д゚) [\'c\']': 'c',
+            '(゚Д゚) [゚ー゚ノ]': 'd',
+            '(゚Д゚) [゚Д゚ノ]': 'e',
+            '(゚Д゚) [1]': 'f',
+            '(゚Д゚) [\'o\']': 'o',
+            '(o゚ー゚o)': 'u',
+            '(゚Д゚) [\'c\']': 'c',
+            '((゚ー゚) + (o^_^o))': '7',
+            '((o^_^o) +(o^_^o) +(c^_^o))': '6',
+            '((゚ー゚) + (゚Θ゚))': '5',
+            '(-~3)': '4',
+            '(-~-~1)': '3',
+            '(-~1)': '2',
+            '(-~0)': '1',
+            '((c^_^o)-(c^_^o))': '0',
+        }
+        delim = '(゚Д゚)[゚ε゚]+'
+        end_token = '(゚Д゚)[゚o゚]'
+        symbols = '|'.join(map(re.escape, symbol_dict.keys()))
+        txt = re.sub('(%s)\+\s?' % symbols, lambda m: symbol_dict[m.group(1)], txt)
+        ret = ''
+        for aacode in re.findall(r'{0}\+\s?{1}(.*?){0}'.format(re.escape(end_token), re.escape(delim)), txt):
+            for aachar in aacode.split(delim):
+                if aachar.isdigit():
+                    ret += compat_chr(int(aachar, 8))
+                else:
+                    m = re.match(r'^u([\da-f]{4})$', aachar)
+                    if m:
+                        ret += compat_chr(int(m.group(1), 16))
+                    else:
+                        self.report_warning("Cannot decode: %s" % aachar)
+        return ret
+
     def _real_extract(self, url):
         video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
+        webpage = self._download_webpage('https://openload.co/embed/%s/' % video_id, video_id)
 
-        if 'File not found' in webpage:
+        if 'File not found' in webpage or 'deleted by the owner' in webpage:
             raise ExtractorError('File not found', expected=True)
 
-        # The following extraction logic is proposed by @Belderak and @gdkchan
-        # and declared to be used freely in youtube-dl
-        # See https://github.com/rg3/youtube-dl/issues/9706
-
-        numbers_js = self._download_webpage(
-            'https://openload.co/assets/js/obfuscator/n.js', video_id,
-            note='Downloading signature numbers')
-        signums = self._search_regex(
-            r'window\.signatureNumbers\s*=\s*[\'"](?P<data>[a-z]+)[\'"]',
-            numbers_js, 'signature numbers', group='data')
-
-        linkimg_uri = self._search_regex(
-            r'<img[^>]+id="linkimg"[^>]+src="([^"]+)"', webpage, 'link image')
-        linkimg = self._request_webpage(
-            linkimg_uri, video_id, note=False).read()
-
-        width, height, pixels = decode_png(linkimg)
-
-        output = ''
-        for y in range(height):
-            for x in range(width):
-                r, g, b = pixels[y][3 * x:3 * x + 3]
-                if r == 0 and g == 0 and b == 0:
-                    break
-                else:
-                    output += compat_chr(r)
-                    output += compat_chr(g)
-                    output += compat_chr(b)
-
-        img_str_length = len(output) // 200
-        img_str = [[0 for x in range(img_str_length)] for y in range(10)]
-
-        sig_str_length = len(signums) // 260
-        sig_str = [[0 for x in range(sig_str_length)] for y in range(10)]
-
-        for i in range(10):
-            for j in range(img_str_length):
-                begin = i * img_str_length * 20 + j * 20
-                img_str[i][j] = output[begin:begin + 20]
-            for j in range(sig_str_length):
-                begin = i * sig_str_length * 26 + j * 26
-                sig_str[i][j] = signums[begin:begin + 26]
-
-        parts = []
-        # TODO: find better names for str_, chr_ and sum_
-        str_ = ''
-        for i in [2, 3, 5, 7]:
-            str_ = ''
-            sum_ = float(99)
-            for j in range(len(sig_str[i])):
-                for chr_idx in range(len(img_str[i][j])):
-                    if sum_ > float(122):
-                        sum_ = float(98)
-                    chr_ = compat_chr(int(math.floor(sum_)))
-                    if sig_str[i][j][chr_idx] == chr_ and j >= len(str_):
-                        sum_ += float(2.5)
-                        str_ += img_str[i][j][chr_idx]
-            parts.append(str_.replace(',', ''))
-
-        video_url = 'https://openload.co/stream/%s~%s~%s~%s' % (parts[3], parts[1], parts[2], parts[0])
+        # The following decryption algorithm is written by @yokrysty and
+        # declared to be freely used in youtube-dl
+        # See https://github.com/rg3/youtube-dl/issues/10408
+        enc_data = self._html_search_regex(
+            r'<span[^>]*>([^<]+)</span>\s*<span[^>]*>[^<]+</span>\s*<span[^>]+id="streamurl"',
+            webpage, 'encrypted data')
+
+        enc_code = self._html_search_regex(r'<script[^>]+>(゚ω゚[^<]+)</script>',
+                                           webpage, 'encrypted code')
+
+        js_code = self.openload_decode(enc_code)
+        jsi = JSInterpreter(js_code)
+
+        m_offset_fun = self._search_regex(r'slice\(0\s*-\s*(%s)\(\)' % _NAME_RE, js_code, 'javascript offset function')
+        m_diff_fun = self._search_regex(r'charCodeAt\(0\)\s*\+\s*(%s)\(\)' % _NAME_RE, js_code, 'javascript diff function')
+
+        offset = jsi.call_function(m_offset_fun)
+        diff = jsi.call_function(m_diff_fun)
+
+        video_url_chars = []
+
+        for idx, c in enumerate(enc_data):
+            j = compat_ord(c)
+            if j >= 33 and j <= 126:
+                j = ((j + 14) % 94) + 33
+            if idx == len(enc_data) - offset:
+                j += diff
+            video_url_chars += compat_chr(j)
+
+        video_url = 'https://openload.co/stream/%s?mime=true' % ''.join(video_url_chars)
 
         title = self._og_search_title(webpage, default=None) or self._search_regex(
             r'<span[^>]+class=["\']title["\'][^>]*>([^<]+)', webpage,
             'title', default=None) or self._html_search_meta(
             'description', webpage, 'title', fatal=True)
 
-        return {
+        entries = self._parse_html5_media_entries(url, webpage, video_id)
+        subtitles = entries[0]['subtitles'] if entries else None
+
+        info_dict = {
             'id': video_id,
             'title': title,
             'thumbnail': self._og_search_thumbnail(webpage, default=None),
             'url': video_url,
             # Seems all videos have extensions in their titles
             'ext': determine_ext(title),
+            'subtitles': subtitles,
         }
+
+        return info_dict