wasi-master / sponsorblock.py Goto Github PK

Hi,
So overall I'm really satisfied with this library, as it does everything I want.
However, I occasionally get this exception and can't figure out why.
HTTPSConnectionPool(host='sponsor.ajay.app', port=443): Max retries exceeded with url: /api/skipSegments?videoID=gxEPV4kolz0&category=sponsor&category=selfpromo&category=interaction&category=intro&category=outro&category=preview&category=music_offtopic&category=poi_highlight&service=YouTube (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x000001CFA622A310>: Failed to establish a new connection: [WinError 10060] Een verbindingspoging is mislukt omdat de verbonden party niet correct heeft geantwoord na een bepaalde tijd, of de gemaakte verbinding is mislukt omdat de verbonden host niet heeft geantwoord'))
The WinError 10060 says: "An attempt to connect has failed, because the connected party answered incorrectly after a certain time, or the made connection has failed because the connected host hasn't answered'".

Does sponsorblock has a rate limit? If yes, how do I work around it (a sleep() function perhaps) to prevent this?

Description of the bug

Connection errors should be caught and processed, or at least there should be a silent paramater, that says, if an error should be thrown. There is an argument to be made, that this actually shouldn't be the case, but imo the silent parameter should exist.

I also reccomend using logging, then you could just print a warning. When I find the time I am gonna open a pr for this, if you haven't implemented it.

Also sorry if I come of rude, this library really is helpful, but it just crashed my 10 000 lines program. xD
Seriosly keep going.

<33

EDIT: IMO the silent attribute should just return an empty array if anything fails, because yk not segments.

Steps To Reproduce

1. Use the programm
2. Pull out your ethernet cable

Additional Information

No response

Description of the bug

When attempting to import this library on Python lower than 3.8, the following error will appear:

ImportError: cannot import name 'Literal' from 'typing' (/usr/lib/python3.7/typing.py)

This error was reproduced on 2 of my Raspberry Pis running Raspbian GNU/Linux 10 (buster).
There is fortunately a trivial fix. In utils.py change

from typing import Dict, Literal

to

from typing import Dict
from typing_extensions import Literal

Or put the current Literal import code in a try / except with the exception importing the 2nd way. I could make a pull request but figured this would be easier.

Steps To Reproduce

On Python < 3.8:

pip install sponsorblock
python -c "import sponsorblock"

Additional Information

No response

Hi,

I really appreciate your work, and I already contributed once to your project, but I seem to have overlooked some things. After consideration, I decided it would be less work, to implement my own sponsorblock wrapper than to "fix" the issues in your project.

I got a segment in my readme, why I made this decision. If you want you can read this critique, but ofc you don't have to.

Regardless, I consider using your logo as logo for my wrapper. I can see why you might be opposed to this, thus I ask. Would it be fine with you if I used your logo in my own api wrapper, which kinda does the same? If you are opposed to this I ofc. won't do it.

I hope you have a nice day
<333

Hi,
So I want to use get_skip_segments to fetch segments from a YouTube video in JSON format, so JavaScript can use it.
However, the results are apparently Segment objects inside a list wrapper.
Is there a way for me to convert the segment objects into a JSON string, so I can use it for JavaScript?
Because I now get this error: TypeError: Object of type Segment is not JSON serializable

Description of the bug

I wanted to submit segments, however I got a 400 Bad Request. I then modified the client.py Code to show the error by printing response.text in line 391. => "No valid videoID, userID field(s) provided." But these were both valid.
After some searching: You need to send a header that declares you're sending json.

headers = {'Content-type': 'application/json'}
requests.post(url, data=json.dumps(body), headers=headers)

Or alternative you can use the json keyword instead of data

Maybe it would be a good idea to append the request.text to the raised error message so the suer knows what exactly is the problem

Steps To Reproduce

import sponsorblock
client = sponsorblock.Client(user_id="my ID from the settings menu")
client.add_skip_segments(
        "https://www.youtube.com/watch?v=XYZ",
        segment=sponsorblock.Segment(category="intro", start=0, end=6.845000)
    )

=> 400 Bad Request

Additional Information

No response

After updating to 0.2.0, I can't initialize Sponsorblock Client anymore.

Python 3.9.8 (tags/v3.9.8:bb3fdcf, Nov  5 2021, 20:48:33) [MSC v.1929 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import sponsorblock
>>> client = sponsorblock.Client()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "...\lib\site-packages\sponsorblock\utils.py", line 124, in __call__
    if cls not in cls._instances:
AttributeError: type object 'Client' has no attribute '_instances'

Description

The normal getter for segments hides those which are voted negativ etc.
I wanted a method to get all segments of a video, regardless of votes.

This could be achieved by adding a searchSegments Delegation API Method or an actual message which fills in the blanks such you get every segment returned for the given URL.

Additional Information

No response

Description of the bug

The example for the add_skip_segment (https://sponsorblockpy.readthedocs.io/en/latest/api_reference.html?highlight=client#sponsorblock.client.Client.add_skip_segments)
the 5th line of the example code should start with "

Steps To Reproduce

reading the wiki :D

Additional Information

No response