Make HTTP request directly from URL object

[Kilo59]

The urlpath library has this convienent feature. Would it be possible to implement this in yarl?

Maybe have this installable as an extra?

pip install yarl[requests]

Do HTTP requests:

>>> url = URL('https://httpbin.org/get')
>>> url.get()
<Response [200]>

>>> url = URL('https://httpbin.org/post')
>>> url.post(data={'key': 'value'})
<Response [200]>

>>> url = URL('https://httpbin.org/delete')
>>> url.delete()
<Response [200]>

>>> url = URL('https://httpbin.org/patch')
>>> url.patch(data={'key': 'value'})
<Response [200]>

>>> url = URL('https://httpbin.org/put')
>>> url.put(data={'key': 'value'})
<Response [200]>

[aio-libs-bot]

GitMate.io thinks possibly related issues are #33 (Parse URL parameter), #272 (bool(URL()) should be False), #24 (TypeError in URL.with_query), #84 (Incorrect handling of '..' in url path), and #196 (Use custom URL parser).

[kxepal]

I think this is out of the scope of this project. Also, which framework will it use behind the scenes? Why it's synchronous? How it will get manage sessions, keep-alives, cookies and the rest things which aren't related to URL object?

[Kilo59]

I would say Requests with the requests defaults.
I'm not imagining it being anything more complicated than a convience wrapper around requests HTTP calls.

[kxepal]

Ok. Then what stops you from writing:

>>> requests.get(URL('https://github.com'))
<Response [200]>

? In this case you gain full power of http client while URL remains about urls.

[Kilo59]

Absolutely nothing.
Like I said, it would be a convenience feature.
And something I would be willing to work on if I thought it would be accepted.

[kxepal]

That feature will cost a lot for this project with oblivious limitations, while it's all about to save you 9 chars from typing. Does it worth it?

[asvetlov]

@Kilo59 sorry.
I/O is out of the project scope.
Feature requests like yours are never going to land here, unfortunately.