fix: use mlflow env variables directly for mlflow auth

[nbowen3]

Relevant issue or PR

Description of changes

In #416, we added support for mlflow auth. Originally, we used tesseract config variables and added it to the url via basic auth. However, mlflow directly outputs the url it receives, which results in leaking this information:

🏃 View run loud-quail-891 at: http://admin/:passwordpassword@

This goes back to the original approach of just using the MLFLOW_USERNAME/MLFLOW_PASSWORD env vars directly. The mlflow client reads them in, and we just have to read them from the env variable for the reachability check.

Testing done

Local testing

[nbowen3]

cc @xalelax

[codecov]

Codecov Report

❌ Patch coverage is 55.55556% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 76.38%. Comparing base (c895f05) to head (8e7ed8f).

Files with missing lines	Patch %	Lines
tesseract_core/runtime/mpa.py	55.55%	3 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #434      +/-   ##
==========================================
- Coverage   76.80%   76.38%   -0.42%     
==========================================
  Files          29       29              
  Lines        3444     3439       -5     
  Branches      537      537              
==========================================
- Hits         2645     2627      -18     
- Misses        569      580      +11     
- Partials      230      232       +2     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[xalelax]

As discussed, unfortunately due to mlflow constraints this is the best compromise. LGTM!