Refactor client internal calls to allow keepalive

[sonneveld]

Internal calls were closing the http connection after every request, which meant that subsequent queries had to recreate the connection. By only closing the request object, I halved the time it took to get my saved tracks:

loading saved tracks before:
real 0m39.799s

loading saved tracks after:
real 0m19.340s

[Harrison97]

Hey! I am hosting a new master branch for this repo since the old creator has been MIA for 2 years. Hoping to get this repo back on it's feet and in good standing. I've been working on updating tests and adding all new functionality and have merged some of the PR's that I thought were sufficient for merging. If your PR didn't make it, resubmit it please.

Thanks!

https://github.com/Harrison97/spotipy

[stephanebruckert]

Hey @sonneveld. Great PR, thanks! Just ran the tests locally, and integration tests are twice as fast:

----------------------------------------------------------------------
Ran 60 tests in 8.069s

OK

vs.

----------------------------------------------------------------------
Ran 60 tests in 20.057s

OK

Going to resolve conflicts and merge it. Will open a new one to fix a warning in the tests

[stephanebruckert]

TIL we are allowed to push on a fork's branch if a PR is open:

You can commit changes on a pull request branch that was created from a fork of your repository with permission from the pull request creator.

https://help.github.com/en/github/collaborating-with-issues-and-pull-requests/committing-changes-to-a-pull-request-branch-created-from-a-fork

[sonneveld]

Nice! Just caught up on the news of the new maintainers. Good luck and thanks for taking it on.

[ritiek]

Thanks for the PR @sonneveld! Looks good to me.

This example code now consistently runs considerably faster:

playlist_url = "https://open.spotify.com/playlist/58mpINjST18mK6qvvPYRzO"
playlist = sp.playlist(playlist_url)

tracks = playlist["tracks"]
while True:
    for item in tracks["items"]:
        print(item["track"]["name"])
    if not tracks["next"]:
        break
    tracks = sp.next(tracks)

Takes ~26s without this PR and ~15s with this PR.