(Git-6) Use new Git Backend in Entire auto_tick flow

[ytausch]

Description:

Building upon #3003, this replaces the remaining auto_tick logic to use the new git backend.
Notably, this PR ensures that the new backend is used for opening PRs and commenting on PRs.

Checklist:

• Pydantic model updated or no update needed

Cross-refs, links to issues, etc:

#3003 should be merged first.

[ytausch]

@0xbe7a I think this is worth reviewing, even with limited knowledge of the codebase

[0xbe7a]

Why did you relax this?

[ytausch]

In GitHubBackend.create_pull_request, the pr_json pydantic model is directly validated on the JSON returned by the GitHub API. However, we do not want to keep all the information the API returns, mainly since these pydantic models also correspond to the data stored in cf-graph. Hence, I relaxed this to ignore any additional data returned by GitHub.
Since the direct usage of the model also prevents the code from introducing additional extra fields to the pr_json data, it's OK that the model validation test can no longer detect such extra fields when processing the cf-graph data.

[codecov]

Codecov Report

Attention: Patch coverage is 89.92000% with 63 lines in your changes missing coverage. Please review.

Project coverage is 77.40%. Comparing base (a1942f2) to head (d29da4a).

Files with missing lines	Patch %	Lines
conda_forge_tick/auto_tick.py	14.00%	43 Missing ⚠️
conda_forge_tick/git_utils.py	88.79%	13 Missing ⚠️
tests/test_git_utils.py	98.40%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3004      +/-   ##
==========================================
+ Coverage   76.27%   77.40%   +1.12%     
==========================================
  Files         115      115              
  Lines       12532    12948     +416     
==========================================
+ Hits         9559    10022     +463     
+ Misses       2973     2926      -47     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[0xbe7a]

I'm not a huge fan of encoding the credentials in the remote URL and then censoring the logs. I think a better approach would be similar to how GitHub handles it in the checkout action, by setting up the 'http.https://github.com/.extraheader' git config to include the authorization header.

[beckermr]

Yeah putting the credentials on disk is not so great too right, or did I miss something? In any case, I don't actually care either way, but if we do put credentials on disk, we'll want to ensure they get cleaned up before the python process exits. On CI, GitHub will do this for us but potentially people running locally with a token might not.