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

[ytausch]

Description:

This PR reapplies the revert done in #3024.

The error was observed because the HTTP Basic Auth token needs to be base64 encoded.

c54001a applies the necessary fix.

Checklist:

• Pydantic model updated or no update needed

Cross-refs, links to issues, etc:

Cc @pavelzw

[beckermr]

LGTM!

[codecov]

Codecov Report

Attention: Patch coverage is 92.81437% with 36 lines in your changes missing coverage. Please review.

Project coverage is 77.46%. Comparing base (c77cf4a) to head (c54001a).
Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
conda_forge_tick/auto_tick.py	15.78%	16 Missing ⚠️
conda_forge_tick/git_utils.py	89.16%	13 Missing ⚠️
tests/test_git_utils.py	98.01%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3124      +/-   ##
==========================================
+ Coverage   76.60%   77.46%   +0.85%     
==========================================
  Files         124      124              
  Lines       13017    13280     +263     
==========================================
+ Hits         9972    10287     +315     
+ Misses       3045     2993      -52     

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

[beckermr]

@ytausch we need to revert this one again: See the errors in this job:

https://github.com/regro/cf-scripts/actions/runs/11783101890/job/32819561459

[ytausch]

@ytausch we need to revert this one again: See the errors in this job:

https://github.com/regro/cf-scripts/actions/runs/11783101890/job/32819561459

I think this is not a functional bug but just confusing log output. Very likely the message is generated by these lines:

cf-scripts/conda_forge_tick/git_utils.py

Lines 529 to 538 in eedaf0f

	try:
	logger.debug(
	f"Trying to checkout branch {new_branch} without creating a new branch"
	)
	self.checkout_branch(target_dir, new_branch)
	except GitCliError:
	logger.debug(
	f"It seems branch {new_branch} does not exist. Creating it.",
	)
	self.checkout_new_branch(target_dir, new_branch, start_point=base_branch)

(The bot tries to checkout a branch to see if it already exists.)

I can suppress stdout and stderr output for these speculative git invocations, WDYT?

[beckermr]

No there is a bug. See here:

  2024-11-11 17:21:03,988 ERROR    conda_forge_tick.auto_tick || NON GITHUB ERROR
  Traceback (most recent call last):
    File "/home/runner/work/cf-scripts/cf-scripts/cf-scripts/conda_forge_tick/auto_tick.py", line 720, in _run_migrator_on_feedstock_branch
      migrator_uid, pr_json = run_with_tmpdir(
                              ^^^^^^^^^^^^^^^^
    File "/home/runner/work/cf-scripts/cf-scripts/cf-scripts/conda_forge_tick/auto_tick.py", line 451, in run_with_tmpdir
      return run(
             ^^^^
    File "/home/runner/work/cf-scripts/cf-scripts/cf-scripts/conda_forge_tick/auto_tick.py", line 582, in run
      pr_data = git_backend.create_pull_request(
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
    File "/home/runner/work/cf-scripts/cf-scripts/cf-scripts/conda_forge_tick/git_utils.py", line 941, in create_pull_request
      header_fields = {
                      ^
    File "/home/runner/work/cf-scripts/cf-scripts/cf-scripts/conda_forge_tick/git_utils.py", line 942, in <dictcomp>
      k: self._github3_session.last_response_headers[k]
         ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^
    File "/home/runner/micromamba/envs/cf-scripts/lib/python3.11/site-packages/requests/structures.py", line 52, in __getitem__
      return self._store[key.lower()][1]
             ~~~~~~~~~~~^^^^^^^^^^^^^
  KeyError: 'last-modified'

[ytausch]

Just verified: The GitHub API returns the Last-Modified header only if we get a PR with GET, and not when creating one with POST, although the response is otherwise identical. Should not be too hard to fix.