test: type-check test/utils and add it to mypy config

[mustafakhokhar]

Related Issues

• Part of Run typing checks on our test files #10396 (first piece-meal step — enables type-checking for test/utils)

Proposed Changes:

A first incremental step toward typing the whole test suite, as suggested in #10396 ("do it in a piece-meal approach where we only turn on type checking in a small part of tests and grow from there").

• Add test/utils/ to the test:types mypy paths in pyproject.toml.
• Make test/utils pass mypy, preferring real fixes over # type: ignore: proper type annotations, assert isinstance(...) narrowing, and correcting mis-typed test values.
• Fix the init_http_client overloads in haystack/utils/http_client.py to carry defaults (per review). The previous overloads omitted defaults, so calling it with no/partial arguments matched no overload. This is a typing-only change — the implementation and its runtime behaviour are unchanged — and it lets the corresponding test ignores be removed.
• # type: ignore remains only where a test intentionally passes invalid input (e.g. a non-dict http_client_kwargs, error=None) or monkeypatches a method — each scoped to the exact error code.

How did you test it?

• hatch run test:types → clean (now also type-checks test/utils).
• hatch run test:unit test/utils/ → passes.
• hatch run fmt and pre-commit are clean.

Notes for the reviewer

• The only non-test change is the init_http_client overload signatures (typing only, no runtime behaviour change), so no release note is included (the reno CI check passes).
• Addressed all review feedback: corrected the overloads, replaced the attr-defined ignores in test_auth with isinstance narrowing, and gave test_hf's tool_result its correct str type.
• If you run the suite on Python 3.14, you may see one unrelated, pre-existing failure in test_type_serialization (a Union repr change) — independent of this PR.

Checklist

• I have read the contributors guidelines and the code of conduct.
• Added/updated type annotations and typing fixes; no runtime behaviour change.
• I've used a conventional commit type for my PR title (test:).
• I have documented my code.
• Release note — N/A: typing-only change, no runtime behaviour (the reno check passes).
• I have run pre-commit hooks and fixed any issue.

---

Implemented with the help of an AI assistant; I have reviewed all changes and run the relevant tests and quality checks.

[vercel]

@mustafakhokhar is attempting to deploy a commit to the deepset Team on Vercel.

A member of the Team first needs to authorize it.

[github-actions]

Coverage report

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
haystack/utils
http_client.py
Project Total

This report was generated by python-coverage-comment-action

[mustafakhokhar]

Hi @sjrl - gentle ping on this whenever you have a moment. It's a first piece-meal slice of #10396 (type-checking test/utils); CI is green and it merges cleanly. Happy to adjust anything you'd like changed. 🙏

[sjrl]

Hey @mustafakhokhar thanks for opening the PR! I'll go ahead and give a review of this, this morning.

[mustafakhokhar]

Thanks for the review @sjrl, all addressed 🙏

• init_http_client overloads: applied your suggested signatures (defaults on both overloads); removed the 4 now-unnecessary call-overload ignores in the test (kept only the one on the intentionally-invalid "invalid" string, which is still a genuine type error).
• test_auth: added assert isinstance(secret, EnvVarSecret) to narrow the type instead of the attr-defined ignores.
• test_hf: tool_result is now the correctly-typed JSON string instead of a dict, so the arg-type ignores are gone.

hatch run test:types and the unit tests are green, and I've rebased on main. PTAL!

[mustafakhokhar]

Reverted 👍 — restored type(None) in those Dict/List subscriptions (agreed, these tests are specifically about NoneType as a type argument), and kept a scoped # type: ignore[misc] there instead. Pushed.

[sjrl]

Thanks!