Pytube only works periodically (KeyError: 'assets')

问题

Five out of ten times Pytube will send me this error when attempting to run my small testing script.

Here's the script:

import pytube
import urllib.request


from pytube import YouTube
yt = YouTube('https://www.youtube.com/watch?v=3NCyD3XoJgM')

print('Youtube video title is: ' + yt.title + '! Downloading now!')

Here's what I get:

Traceback (most recent call last):
  File "youtube.py", line 6, in <module>
    yt = YouTube('https://www.youtube.com/watch?v=3NCyD3XoJgM')
  File "C:\Users\test\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.8_qbz5n2kfra8p0\LocalCache\local-packages\Python38\site-packages\pytube\__main__.py", line 91, in __init__
    self.prefetch()
  File "C:\Users\test\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.8_qbz5n2kfra8p0\LocalCache\local-packages\Python38\site-packages\pytube\__main__.py", line 183, in prefetch
    self.js_url = extract.js_url(self.watch_html)
  File "C:\Users\test\AppData\Local\Packages\PythonSoftwareFoundation.Python.3.8_qbz5n2kfra8p0\LocalCache\local-packages\Python38\site-packages\pytube\extract.py", line 143, in js_url
    base_js = get_ytplayer_config(html)["assets"]["js"]
KeyError: 'assets'

I am very confused. I attempted to reinstall Python plus pytube but I can't seem to remedy this issue. It's increasingly perplexing that the script works half of the time, but not the other half.

回答1:

For now fixed 100% with this:

https://github.com/nficano/pytube/pull/767#issuecomment-716184994

With anyone else getting this error or issue, run this command in a terminal or cmd: python -m pip install git+https://github.com/nficano/pytube

An update to pytubeX that hasn't been released with the pip installation yet. The GitHub link is the current dev explaining the situation.

回答2:

I'm in the same trouble, but I guarantee that the top answer doesn't solve anything, just hide the problem until it pops up again. I investigated this scope of "extract.py" file, and found an error. This scope searches for a "string" snippet in the source code of the Youtube page where the video is, through a dictionary search, such as:

#Example ---------------
Vars = {
    'name':'luis'
    'age':'27'
}
print(Vars['name'])

result: 'luis'

#Extract.py Code -------

def js_url(html: str) -> str:
"""Get the base JavaScript url.

Construct the base JavaScript url, which contains 
the decipher
"transforms".

:param str html:
    The html contents of the watch page.
"""
base_js = get_ytplayer_config(html)["assets"]["js"]
return "https://youtube.com" + base_js

The error:

base_js = get_ytplayer_config(html)["assets"]["js"]
KeyError: 'assets'

It is given because this snippet of the source code does not support a search as dicionario, so 'KeyError' key error, because 'assets' is not a valid key, and the source code is not a dictionary. So I did this script, which I believe replace this original, but in mine, particularly, appeared other errors.

def js_url(html: str) -> str:
"""Get the base JavaScript url.

Construct the base JavaScript url, which contains 
the decipher
"transforms".

:param str html:
    The html contents of the watch page.
"""
base_js = html[html.find('js') + 4:html.find('.js') 
+ 4]
return "https://youtube.com" + base_js

The above script searches for what the function wants as a string, not as dictionary.

I hope I have contributed to a more complete future solution :)

回答3:

Add this function to extract.py

def get_ytplayer_js(html: str) -> Any:
    """Get the YouTube player base JavaScript path.

    :param str html
    The html contents of the watch page.
    :rtype: str
    :returns:
    Path to YouTube's base.js file.
    """
    js_url_patterns = [
        r"\"jsUrl\":\"([^\"]*)\"",
    ]
    for pattern in js_url_patterns:
        regex = re.compile(pattern)
        function_match = regex.search(html)
        if function_match:
            logger.debug("finished regex search, matched: %s", pattern)
            yt_player_js = function_match.group(1)
            return yt_player_js

    raise RegexMatchError(
       caller="get_ytplayer_js", pattern="js_url_patterns"
    )

and change the function "js_url" in extract.py from:

def js_url(html: str) -> str:
    """Get the base JavaScript url.

    Construct the base JavaScript url, which contains the decipher
    "transforms".

    :param str html:
        The html contents of the watch page.
    """
    base_js = get_ytplayer_config(html)["assets"]["js"]
    return "https://youtube.com" + base_js

to:

def js_url(html: str) -> str:
    """Get the base JavaScript url.

    Construct the base JavaScript url, which contains the decipher
    "transforms".

    :param str html:
        The html contents of the watch page.
    """
    base_js = get_ytplayer_js(html)
    return "https://youtube.com" + base_js

回答4:

It seems Pytube module is updated.

It works fine for pytube package

i.e try pip install pytube uninstall pytube variations

回答5:

I had the same problem and updating pytube to the latest version available currently the problem disappeared.

pip install pytube==10.0.0

or

pip install --upgrade pytube

回答6:

If you are using the package pytube or pytube3, I would recommend uninstalling that and installing pytubeX. No need to change imports. I have found it works much more reliably.

Edit: From the comments, if none of these work, try pytube4

Edit: pytube is now being maintained again!

回答7:

It's an issue with the pytube library files. You can fix this by manually modifying the "extract.py" file inside the pytube folder. Copy and paste this inside the file instead: https://github.com/nficano/pytube/blob/master/pytube/extract.py

回答8:

Fixed

The extract.py codebase is now updated if you're still getting the error after running this command in a terminal or cmd: python -m pip install git+https://github.com/nficano/pytube is because it didn't updated your pytube/extract.py file.

The fix is Copy all the code from codebase and replace in your extract.py file. I hope this will work.

回答9:

In order to avoid this pytube problem, you can use youtube_dl instead. Here's a code which were tested on Windows and on an Android tablet (with the Pydroid3 app). The purpose is to download the audio track of videos referred to in a public playlist.

import os, re
import youtube_dl
from pytube import Playlist

YOUTUBE_STREAM_AUDIO = '140'
if os.name == 'posix':
    targetAudioDir = '/storage/emulated/0/Download/Audiobooks/test_youtube_dl'
    ydl_opts = {
    'outtmpl': targetAudioDir + '/%(title)s.mp3',
    'format': 'bestaudio/best',
    'quiet': False
    }
else:
    targetAudioDir = 'D:\\Users\\Jean-Pierre\\Downloads\\Audiobooks\\test_youtube_dl'
    ydl_opts = {
    'outtmpl': targetAudioDir + '\\%(title)s.%(ext)s',
    'format': 'bestaudio/best',
    'postprocessors': [{
                        'key': 'FFmpegExtractAudio',
                        'preferredcodec': 'mp3',
                        'preferredquality': '128',
                    }],
    'quiet': False
    }

playlistUrl = 'https://www.youtube.com/playlist?list=PLzwWSJNcZTMSFWGrRGKOypqN29MlyuQvn'
playlistObject = Playlist(playlistUrl)
playlistObject._video_regex = re.compile(r"\"url\":\"(/watch\?v=[\w-]*)")
    
with youtube_dl.YoutubeDL(ydl_opts) as ydl:
    for videoUrl in playlistObject.video_urls:
        meta = ydl.extract_info(videoUrl, download=False)
        videoTitle = meta['title']
        print('Video title: ' + videoTitle)
        ydl.download([videoUrl])

回答10:

Here's a permanent fix to that! You can try tube_dl.

pip install tube_dl
from tube_dl import Youtube
yt = Youtube('url')
yt.Formats()[0].download()

It uses modular approach and is up to date

More about this can be found at : https://github.com/shekharchander/tube_dl/

回答11:

Try to replace the line 143

    base_js = get_ytplayer_config(html)["assets"]["js"]

with

    try:
        base_js = get_ytplayer_config(html)["assets"]["js"]
    except Exception:
        pass

来源：https://stackoverflow.com/questions/64492922/pytube-only-works-periodically-keyerror-assets