yt_dlp/extractor/neteasemusic.py

@@ -2,6 +2,8 @@ import itertools
 import json
 import re
 import time
+from base64 import b64encode
+from binascii import hexlify
 from hashlib import md5
 from random import randint
 
@@ -11,8 +13,11 @@ from ..compat import compat_urllib_parse_urlencode
 from ..networking import Request
 from ..utils import (
     ExtractorError,
+    bytes_to_intlist,
     clean_html,
+    float_or_none,
     int_or_none,
+    intlist_to_bytes,
     str_or_none,
     strftime_or_none,
     traverse_obj,
@@ -24,39 +29,52 @@ from ..utils import (
 
 
 class NetEaseMusicBaseIE(InfoExtractor):
-    _FORMATS = ['bMusic', 'mMusic', 'hMusic', 'sqMusic', 'hrMusic']
+    _FORMATS = ['bMusic', 'mMusic', 'hMusic']
     _NETEASE_SALT = '3go8&$8*3*3h0k(2)2'
     _API_BASE = 'http://music.163.com/api/'
 
+    @classmethod
+    def _encrypt(cls, dfsid):
+        salt_bytes = bytearray(cls._NETEASE_SALT.encode('utf-8'))
+        string_bytes = bytearray(str(dfsid).encode('ascii'))
+        salt_len = len(salt_bytes)
+        for i in range(len(string_bytes)):
+            string_bytes[i] = string_bytes[i] ^ salt_bytes[i % salt_len]
+        m = md5()
+        m.update(bytes(string_bytes))
+        result = b64encode(m.digest()).decode('ascii')
+        return result.replace('/', '_').replace('+', '-')
+
     def _create_eapi_cipher(self, api_path, query, cookies):
+        KEY = b'e82ckenh8dichen8'
         request_text = json.dumps({**query, 'header': cookies}, separators=(',', ':'))
 
         message = f'nobody{api_path}use{request_text}md5forencrypt'.encode('latin1')
         msg_digest = md5(message).hexdigest()
 
-        data = pkcs7_padding(list(str.encode(
-            f'{api_path}-36cd479b6b5-{request_text}-36cd479b6b5-{msg_digest}')))
-        encrypted = bytes(aes_ecb_encrypt(data, list(b'e82ckenh8dichen8')))
-        return f'params={encrypted.hex().upper()}'.encode()
+        data = pkcs7_padding(bytes_to_intlist(
+            f'{api_path}-36cd479b6b5-{request_text}-36cd479b6b5-{msg_digest}'))
+        encrypted = intlist_to_bytes(aes_ecb_encrypt(data, bytes_to_intlist(KEY)))
+        return b'params=' + hexlify(encrypted).upper()
 
     def _download_eapi_json(self, path, song_id, query, headers={}, **kwargs):
         cookies = {
-            'osver': 'undefined',
-            'deviceId': 'undefined',
+            'osver': None,
+            'deviceId': None,
             'appver': '8.0.0',
             'versioncode': '140',
-            'mobilename': 'undefined',
+            'mobilename': None,
             'buildver': '1623435496',
             'resolution': '1920x1080',
             '__csrf': '',
             'os': 'pc',
-            'channel': 'undefined',
+            'channel': None,
             'requestId': f'{int(time.time() * 1000)}_{randint(0, 1000):04}',
         }
         headers = {
             'Content-Type': 'application/x-www-form-urlencoded',
             'Referer': 'https://music.163.com',
-            'Cookie': '; '.join([f'{k}={v}' for k, v in cookies.items()]),
+            'Cookie': '; '.join([f'{k}={v if v is not None else "undefined"}' for [k, v] in cookies.items()]),
             **headers,
         }
         url = urljoin('https://interface3.music.163.com/', f'/eapi{path}')
@@ -80,22 +98,18 @@ class NetEaseMusicBaseIE(InfoExtractor):
 
             bitrate = int_or_none(details.get('bitrate')) or 999000
             data = self._call_player_api(song_id, bitrate)
-            for song in traverse_obj(data, ('data', ...)):
-                song_url = traverse_obj(song, ('url', {url_or_none}))
+            for song in try_get(data, lambda x: x['data'], list) or []:
+                song_url = try_get(song, lambda x: x['url'])
                 if not song_url:
                     continue
                 if self._is_valid_url(song_url, info['id'], 'song'):
                     formats.append({
                         'url': song_url,
+                        'ext': details.get('extension'),
+                        'abr': float_or_none(song.get('br'), scale=1000),
                         'format_id': song_format,
-                        **traverse_obj(song, {
-                            'ext': ('type', {str}),
-                            'abr': ('br', {lambda i: int_or_none(i, scale=1000)}),
-                            'filesize': ('size', {int_or_none}),
-                        }),
-                        **traverse_obj(details, {
-                            'asr': ('sr', {int_or_none}),
-                        }),
+                        'filesize': int_or_none(song.get('size')),
+                        'asr': int_or_none(details.get('sr')),
                     })
                 elif err == 0:
                     err = try_get(song, lambda x: x['code'], int)
@@ -112,6 +126,10 @@ class NetEaseMusicBaseIE(InfoExtractor):
 
         return formats
 
+    @classmethod
+    def convert_milliseconds(cls, ms):
+        return int(round(ms / 1000.0))
+
     def query_api(self, endpoint, video_id, note):
         req = Request('%s%s' % (self._API_BASE, endpoint))
         req.headers['Referer'] = self._API_BASE
@@ -166,7 +184,6 @@ class NetEaseMusicIE(NetEaseMusicBaseIE):
             'timestamp': 691516800,
             'description': 'md5:1ba2f911a2b0aa398479f595224f2141',
             'duration': 268,
-            'alt_title': '伴唱:现代人乐队 合唱:总政歌舞团',
             'thumbnail': r're:^http.*\.jpg',
         },
     }, {
@@ -238,18 +255,20 @@ class NetEaseMusicIE(NetEaseMusicBaseIE):
             song_id, 'Downloading lyrics data')
         lyrics = self._process_lyrics(lyrics_info)
 
+        alt_title = None
+        if info.get('transNames'):
+            alt_title = '/'.join(info.get('transNames'))
+
         return {
             'id': song_id,
+            'title': info['name'],
+            'alt_title': alt_title,
+            'creator': ' / '.join([artist['name'] for artist in info.get('artists', [])]),
+            'timestamp': self.convert_milliseconds(info.get('album', {}).get('publishTime')),
+            'thumbnail': info.get('album', {}).get('picUrl'),
+            'duration': self.convert_milliseconds(info.get('duration', 0)),
             'description': lyrics,
             'formats': formats,
-            'alt_title': '/'.join(traverse_obj(info, (('transNames', 'alias'), ...))) or None,
-            'creator': ' / '.join(traverse_obj(info, ('artists', ..., 'name'))),
-            **traverse_obj(info, {
-                'title': ('name', {str}),
-                'timestamp': ('album', 'publishTime', {lambda i: int_or_none(i, scale=1000)}),
-                'thumbnail': ('album', 'picUrl', {url_or_none}),
-                'duration': ('duration', {lambda i: int_or_none(i, scale=1000)}),
-            }),
         }
 
 
@@ -331,12 +350,11 @@ class NetEaseMusicSingerIE(NetEaseMusicBaseIE):
             'artist/%s?id=%s' % (singer_id, singer_id),
             singer_id, 'Downloading singer data')
 
-        artist_info = info.get('artist', {})
-        name = artist_info.get('name', '')
-        if artist_info.get('trans'):
+        name = info['artist']['name']
+        if info['artist']['trans']:
             name = '%s - %s' % (name, info['artist']['trans'])
-        if artist_info.get('alias'):
-            name = '%s - %s' % (name, ';'.join(map(str, info['artist']['alias'])))
+        if info['artist']['alias']:
+            name = '%s - %s' % (name, ';'.join(info['artist']['alias']))
 
         entries = [
             self.url_result('http://music.163.com/#/song?id=%s' % song['id'],
@@ -503,10 +521,6 @@ class NetEaseMusicProgramIE(NetEaseMusicBaseIE):
             'id': '10141022',
             'title': '滚滚电台的有声节目',
             'description': 'md5:8d594db46cc3e6509107ede70a4aaa3b',
-            'creator': '滚滚电台ORZ',
-            'timestamp': 1434450733,
-            'upload_date': '20150616',
-            'thumbnail': r're:http.*\.jpg',
         },
         'playlist_count': 4,
     }, {
@@ -535,31 +549,31 @@ class NetEaseMusicProgramIE(NetEaseMusicBaseIE):
             'dj/program/detail?id=%s' % program_id,
             program_id, 'Downloading program info')['program']
 
-        metainfo = traverse_obj(info, {
-            'title': ('name', {str}),
-            'description': ('description', {str}),
-            'creator': ('dj', 'brand', {str}),
-            'thumbnail': ('coverUrl', {url_or_none}),
-            'timestamp': ('createTime', {lambda i: int_or_none(i, scale=1000)}),
-        })
+        name = info['name']
+        description = info['description']
 
         if not self._yes_playlist(info['songs'] and program_id, info['mainSong']['id']):
             formats = self.extract_formats(info['mainSong'])
 
             return {
                 'id': str(info['mainSong']['id']),
+                'title': name,
+                'description': description,
+                'creator': info['dj']['brand'],
+                'timestamp': self.convert_milliseconds(info['createTime']),
+                'thumbnail': info['coverUrl'],
+                'duration': self.convert_milliseconds(info.get('duration', 0)),
                 'formats': formats,
-                'duration': traverse_obj(info, ('mainSong', 'duration', {lambda i: int_or_none(i, scale=1000)})),
-                **metainfo,
             }
 
-        song_ids = traverse_obj(info, ((('mainSong', 'id'), ('songs', ..., 'id')), {int_or_none}))
+        song_ids = [info['mainSong']['id']]
+        song_ids.extend([song['id'] for song in info['songs']])
         entries = [
             self.url_result('http://music.163.com/#/song?id=%s' % song_id,
                             'NetEaseMusic', song_id)
             for song_id in song_ids
         ]
-        return self.playlist_result(entries, program_id, **metainfo)
+        return self.playlist_result(entries, program_id, name, description)
 
 
 class NetEaseMusicDjRadioIE(NetEaseMusicBaseIE):
@@ -580,7 +594,8 @@ class NetEaseMusicDjRadioIE(NetEaseMusicBaseIE):
     def _real_extract(self, url):
         dj_id = self._match_id(url)
 
-        metainfo = {}
+        name = None
+        desc = None
         entries = []
         for offset in itertools.count(start=0, step=self._PAGE_SIZE):
             info = self.query_api(
@@ -594,13 +609,13 @@ class NetEaseMusicDjRadioIE(NetEaseMusicBaseIE):
                     'NetEaseMusicProgram', program['id'])
                 for program in info['programs']
             ])
-            if not metainfo:
-                metainfo = traverse_obj(info, ('programs', 0, 'radio', {
-                    'title': ('name', {str}),
-                    'description': ('desc', {str}),
-                }))
+
+            if name is None:
+                radio = info['programs'][0]['radio']
+                name = radio['name']
+                desc = radio['desc']
 
             if not info['more']:
                 break
 
-        return self.playlist_result(entries, dj_id, **metainfo)
+        return self.playlist_result(entries, dj_id, name, desc)