[bvolpato]

Related #30799.

Raise HTTP errors before decoding Hugging Face responses as embeddings, so a failed request reports its status and endpoint instead of failing later in NumPy.

Reuse the authorization established when loading the session. The redundant request header currently replaces a valid HF_TOKEN fallback with Bearer None.

Regression checks fail before the fix and pass afterward using real prepared requests and response objects. The local module tests, Ruff, and YAPF pass. The provider rejection seen in CI still needs separate diagnosis.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 79.93%. Comparing base (a8d6afd) to head (a82805d).
⚠️ Report is 74 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #40779      +/-   ##
==========================================
- Coverage   80.04%   79.93%   -0.12%     
==========================================
  Files         490      493       +3     
  Lines       82916    83474     +558     
==========================================
+ Hits        66373    66721     +348     
- Misses      16543    16753     +210     

Flag	Coverage Δ
python	79.93% <100.00%> (-0.12%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.