feat(server): integrate OpenAI Chat buffered proxy

[Pouyanpi]

Summary

Integrates the authoritative OpenAI Chat projections into the buffered provider-native proxy path and exports their guard contract.

Why

The integration must preserve original provider data while connecting exact guarded targets, checker execution, native errors, route ownership, and OpenAPI documentation. Its exported contract describes the models actually bound to the endpoint.

What Changed

• Adds the guarded POST /v1/chat/completions endpoint and reserves its API-version path family.
• Carries original bytes, validated payload, and exact guarded target through one request state.
• Uses one OpenAI error mapping for runtime rendering and OpenAPI response documentation.
• Preserves non-success provider errors and rejects incompatible successful responses.
• Adds provider revision bindings and mocked HTTP integration coverage.
• Adds a shared buffered exporter that reads the models bound to an existing endpoint, with format validation and an exact drift check for contracts/openai/_generated/chat-completions.buffered.guard.yaml.

Review Notes

Review this PR against #2413. The exported YAML is for inspection and documentation; request handling does not load it. The export contains request and buffered-response policy. The shared CLI consumes an existing endpoint declaration; Python models and bindings remain handwritten.

Streaming requests are recognized but rejected before dispatch. Replacement eligibility does not enable executing replacement outcomes, which remain unsupported here.

create_openai_chat_router(...) returns an APIRouter, not a standalone configured server. The embedding application supplies ContentChecker and HttpDispatch; concrete outbound transport, deployment composition, and trusted Rails configuration resolution remain separate work.

Validation

Offline experimental tests passed, including export and mocked HTTP integration tests. Pre-commit and the exported-contract drift check passed. No live provider calls were used.

AI Assistance

• No AI tools were used.
• AI tools were used; a human reviewed and can explain every change.

Codex assisted with implementation, tests, documentation, and stack maintenance.

Checklist

• I've read the CONTRIBUTING guidelines.
• This is maintainer-led work tracked internally.
• The PR title follows the project commit convention.
• Applicable documentation and tests are included.
• Generated changelog files were not edited manually.
• Automated and human review comments are addressed or answered.
• The responsible reviewer or team is mentioned.

Stack Position

Part 3 of 3.

• Previous: feat(server): add OpenAI Chat buffered projections #2413
• Next: none, final stack PR

Stack Context

The Python projection models and runtime bindings are authoritative. This buffered stack declares their policy alongside the types and exports YAML for inspection and documentation. Request handling does not load the exported contract.

The open chain is #2411 → #2413 → #2414 → #2434 → #2435 → #2436 → #2437. #2412 is closed; its provider provenance and boundary documentation are included in #2413.

Streaming continues in #2434–#2437. The shared exporter currently emits request and buffered-response policy; streaming contract export remains separate work.

Review each PR against its listed base branch.

Order	PR	Branch	Base
1	#2411	pouyanpi/guard-contract-foundation-1	develop
2	#2413	pouyanpi/openai-chat-buffered-projections-3	pouyanpi/guard-contract-foundation-1
3	#2414	pouyanpi/openai-chat-buffered-integration-4	pouyanpi/openai-chat-buffered-projections-3

Summary by CodeRabbit

• New Features
  • Added guarded proxy support for OpenAI Chat Completions, including request and response inspection, provider-compatible error responses, and API revision validation.
  • Added buffered contract export to YAML, with options to write output or check existing artifacts for differences.
  • Added OpenAPI documentation for guarded routes, revision parameters, and provider error responses.
• Documentation
  • Documented contract export options, generated artifact details, and current runtime limitations.

[codecov]

Codecov Report

❌ Patch coverage is 99.05482% with 5 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...ls/server/experimental/provider/contract_export.py	96.70%	3 Missing ⚠️
...uardrails/server/experimental/provider/endpoint.py	97.82%	1 Missing ⚠️
.../server/experimental/provider/projection_policy.py	98.96%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!

[greptile-apps]

Confidence Score: 5/5

[High impact] The PR appears safe to merge based on the reviewed changes.

Summary

The PR connects the OpenAI Chat projections to a buffered guarded proxy and exports their endpoint-bound contract.

• Adds guarded route ownership, request preparation, response inspection, and OpenAI-compatible errors.
• Adds contract export, OpenAPI declarations, and mocked integration coverage.

Diagram

%%{init: {'theme': 'neutral'}}%%
flowchart LR
  A[Chat request] --> B[Validate revision and JSON projection]
  B --> C[Input check]
  C --> D[Provider dispatch]
  D --> E[Validate buffered response projection]
  E --> F[Output check when text is present]
  F --> G[Relay provider response]

Loading

_{Reviews (18) · Last reviewed commit: "test(server): cover buffered integration..." · Reviewed by Greptile}