test(proxy): add property tests for response payload invariants

[iqbalmaulana03]

Summary

Add dev-only Hypothesis property tests for Responses payload sanitization and compact trimming, SSE framing, and historical response.create payload slimming. This implements the maintainer-recommended first, pure-function slice of #1595 without adding lifecycle simulation infrastructure or changing production behavior.

Type of change

• fix: — bug fix (no behavior change beyond the bug)
• feat: — new user-facing feature or capability
• refactor: — internal refactor (no behavior change, no API change)
• docs: — documentation only
• chore: / ci: / build: — tooling, CI, packaging
• test: — test-only change
• Breaking change

Linked issue: Refs #1595

OpenSpec

• This PR includes / updates an OpenSpec change
• Not applicable — test-only/dev-tool change exempt under CONTRIBUTING; no observable behavior, requirement, contract, or schema changes
• This PR touches codex-faithful request/response and SSE paths only through tests and preserves upstream-equivalent behavior

Change directory: N/A

Changes

• Add bounded reusable Hypothesis strategies for arbitrary JSON values, arrays, and objects.
• Property-test request input sanitization, unsupported-field stripping, compact token budgets, marker accounting, state anchors, and tool-call pairing.
• Property-test SSE JSON round trips, CR/LF/CRLF boundaries, Unicode line separators, multiline data, and non-object rejection.
• Property-test historical response.create slimming while preserving the recent suffix, top-level fields, file references, and caller payload immutability.
• Keep Hypothesis in the development dependency group only.

Test plan

uv run pytest tests/unit/test_openai_requests.py tests/unit/test_sse.py tests/unit/test_proxy_utils.py -q
# 1188 passed

uv run ty check
# All checks passed

uv run ruff check tests/unit/test_openai_requests.py tests/unit/test_sse.py tests/unit/test_proxy_utils.py tests/unit/hypothesis_strategies.py
# All checks passed

uv run ruff format --check tests/unit/test_openai_requests.py tests/unit/test_sse.py tests/unit/test_proxy_utils.py tests/unit/hypothesis_strategies.py
# 4 files already formatted

make test-unit
# 5766 passed, 70 skipped

npx --yes @fission-ai/openspec@1.8.0 validate --specs
# 49 passed, 0 failed

Screenshots / output

Not applicable; this is a test-only change with no user-visible surface.

Checklist

• Title is in Conventional Commits format (<type>(<scope>)?: <subject>).
• Linked the related issue above.
• Added or updated tests covering the change.
• Ran the relevant make test-unit target and focused lint/type/test checks locally.
• OpenSpec is not modified; main specs validate successfully.
• Simplicity gates reviewed: no setting, default, README section, .env.example entry, or dashboard surface is added.
• CHANGELOG is not edited by hand.

[Soju06]

@codex review

Thanks — this lands exactly the pure-function slice I suggested in #1595. I've re-run the hung Docker build job (the CI Required failure was infra, not your change). Three small asks while codex runs:

1. The four @given tests without @settings in tests/unit/test_openai_requests.py inherit Hypothesis's default 200 ms deadline, which can flake on loaded CI runners — please add @settings(deadline=None) like the others.
2. Add .hypothesis/ to .gitignore so local runs don't leave an untracked example database.
3. Consider registering a derandomized CI profile (e.g. settings.register_profile("ci", derandomize=True) in a conftest, activated via HYPOTHESIS_PROFILE=ci) so unrelated PRs can't be broken by a freshly generated counterexample — we can loosen this later once the suite has soaked.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 47e8955762

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[iqbalmaulana03]

@codex review

[chatgpt-codex-connector]

To use Codex here, create a Codex account and connect to github.