"""Tests for the ResponsesApiTransport (Codex).""" import json import pytest from types import SimpleNamespace from agent.transports import get_transport from agent.transports.types import NormalizedResponse @pytest.fixture def transport(): import agent.transports.codex # noqa: F401 return get_transport("codex_responses") class TestCodexTransportBasic: def test_api_mode(self, transport): assert transport.api_mode == "codex_responses" def test_registered_on_import(self, transport): assert transport is not None def test_convert_tools(self, transport): tools = [{ "type": "function", "function": { "name": "terminal", "description": "Run a command", "parameters": {"type": "object", "properties": {"command": {"type": "string"}}}, } }] result = transport.convert_tools(tools) assert len(result) == 1 assert result[0]["type"] == "function" assert result[0]["name"] == "terminal" class TestCodexBuildKwargs: def test_cache_key_is_content_addressed_not_session_id(self, transport): """prompt_cache_key is content-addressed from the static prefix (instructions + tools), not the session_id. This keeps recurring cron jobs — whose session_id carries a per-fire timestamp — on a stable warm cache key. The key is a 'pck_' hash and must NOT equal session_id.""" messages = [{"role": "user", "content": "Hi"}] kw = transport.build_kwargs( model="gpt-5.4", messages=messages, tools=[], session_id="cron_job42_20260624_143000", ) pck = kw.get("prompt_cache_key", "") assert pck.startswith("pck_") assert pck != "cron_job42_20260624_143000" def test_cache_key_stable_across_session_ids(self, transport): """Same static prefix + different session_id (e.g. two cron fires of the same job) must yield the same prompt_cache_key — the whole point of the fix: repeated fires reuse the warm prefix instead of going cold.""" messages = [{"role": "user", "content": "Hi"}] kw1 = transport.build_kwargs( model="gpt-5.4", messages=messages, tools=[], session_id="cron_job42_20260624_143000", ) kw2 = transport.build_kwargs( model="gpt-5.4", messages=messages, tools=[], session_id="cron_job42_20260624_143500", ) assert kw1["prompt_cache_key"] == kw2["prompt_cache_key"] def test_github_responses_drops_message_item_id_end_to_end(self, transport): # #32716: Copilot binds codex_message_items ids to a backend # "connection" that doesn't survive credential rotation, a gateway # restart, or load-balancer churn — replaying a stale id gets HTTP # 401 "input item ID does not belong to this connection", even for # ids well under the #27038 64-char length cap. build_kwargs must # thread is_github_responses through to the input converter so the # id never reaches the request. messages = [ {"role": "system", "content": "You are Hermes."}, { "role": "assistant", "content": "pong", "codex_message_items": [ { "type": "message", "role": "assistant", "status": "in_progress", "content": [{"type": "output_text", "text": "pong"}], "id": "msg_short_but_connection_scoped", "phase": "final_answer", } ], }, ] kw = transport.build_kwargs( model="gpt-5.5", messages=messages, tools=[], is_github_responses=True, ) message_item = next(item for item in kw["input"] if item.get("type") == "message") assert "id" not in message_item assert message_item["phase"] == "final_answer" assert message_item["status"] == "in_progress" assert message_item["content"] == [{"type": "output_text", "text": "pong"}] def test_non_github_responses_keeps_message_item_id_end_to_end(self, transport): messages = [ {"role": "system", "content": "You are Hermes."}, { "role": "assistant", "content": "pong", "codex_message_items": [ { "type": "message", "role": "assistant", "status": "completed", "content": [{"type": "output_text", "text": "pong"}], "id": "msg_short_id", } ], }, ] kw = transport.build_kwargs( model="gpt-5.5", messages=messages, tools=[], is_codex_backend=True, ) message_item = next(item for item in kw["input"] if item.get("type") == "message") assert message_item["id"] == "msg_short_id" @pytest.mark.parametrize("model", [ "gpt-5.5", "gpt-5.5-pro", "gpt-5.4", "gpt-5.2", "gpt-5.1-codex-max", "gpt-5.1", "gpt-5.1-codex", "gpt-5.1-codex-mini", "gpt-5.1-chat-latest", "gpt-5", "gpt-5-codex", "gpt-4.1", "openai.gpt-5.5-pro", "openai/gpt-5.1-codex-2026-01-01", ]) def test_extended_cache_models_set_24h_prompt_cache_retention(self, transport, model): messages = [{"role": "user", "content": "Hi"}] kw = transport.build_kwargs( model=model, messages=messages, tools=[], session_id="test-session", base_url="https://bedrock-mantle.us-west-2.api.aws/v1", ) assert kw["prompt_cache_retention"] == "24h" @pytest.mark.parametrize("model", ["gpt-5.6", "gpt-4o", "o3"]) def test_prompt_cache_retention_omitted_for_other_model_families(self, transport, model): kw = transport.build_kwargs( model=model, messages=[{"role": "user", "content": "Hi"}], tools=[], session_id="test-session", base_url="https://bedrock-mantle.us-west-2.api.aws/v1", ) assert "prompt_cache_retention" not in kw @pytest.mark.parametrize("base_url", [ "https://api.openai.com/v1", "https://example.openai.azure.com/openai/v1", "https://api.x.ai/v1", "https://models.github.ai/inference", "https://api.githubcopilot.com", "https://chatgpt.com/backend-api/codex", "https://responses.example.com/v1", "https://bedrock-mantle.us-west-2.api.aws.example/v1", "https://example.com/bedrock-mantle.us-west-2.api.aws/v1", ]) def test_prompt_cache_retention_omitted_for_non_mantle_endpoints(self, transport, base_url): kw = transport.build_kwargs( model="gpt-5.4", messages=[{"role": "user", "content": "Hi"}], tools=[], base_url=base_url, ) assert "prompt_cache_retention" not in kw def test_xai_responses_sends_cache_key_via_extra_body(self, transport): """xAI's Responses API documents ``prompt_cache_key`` as the body-level cache-routing key (the ``x-grok-conv-id`` header is Chat-Completions-only). Passing it via ``extra_body`` is robust against openai SDK builds whose ``Responses.stream()`` kwarg signature ever drops the field — the body field still serializes and reaches xAI either way. The ``x-grok-conv-id`` header is kept as a belt-and-braces fallback so cache routing survives even when the body field would be stripped by an intermediate proxy. Ref: https://docs.x.ai/developers/advanced-api-usage/prompt-caching/maximizing-cache-hits """ messages = [{"role": "user", "content": "Hi"}] kw = transport.build_kwargs( model="grok-4.3", messages=messages, tools=[], session_id="conv-xai-1", is_xai_responses=True, ) assert "prompt_cache_key" not in kw # Body-level prompt_cache_key is content-addressed (pck_ hash), not the # raw session_id, so recurring cron fires stay on a stable warm key. eb_pck = kw.get("extra_body", {}).get("prompt_cache_key", "") assert eb_pck.startswith("pck_") assert eb_pck != "conv-xai-1" # x-grok-conv-id stays the session/transcript id, not the cache key. assert kw.get("extra_headers", {}).get("x-grok-conv-id") == "conv-xai-1" def test_xai_responses_extra_body_preserves_caller_fields(self, transport): """When the caller already supplies ``extra_body`` (e.g. via request_overrides), the xAI cache-key injection must merge into the existing dict instead of overwriting it. Caller-supplied ``prompt_cache_key`` wins (setdefault semantics) so user overrides aren't silently clobbered by the transport.""" messages = [{"role": "user", "content": "Hi"}] kw = transport.build_kwargs( model="grok-4.3", messages=messages, tools=[], session_id="conv-xai-1", is_xai_responses=True, request_overrides={"extra_body": {"prompt_cache_key": "caller-override", "other_field": 42}}, ) eb = kw.get("extra_body", {}) assert eb.get("prompt_cache_key") == "caller-override" assert eb.get("other_field") == 42 @pytest.mark.parametrize("length", [64, 65]) def test_codex_cache_scope_boundary(self, transport, length): session_id = "s" * length scope = transport.build_kwargs( model="gpt-5.4", messages=[{"role": "user", "content": "Hi"}], tools=[], session_id=session_id, is_codex_backend=True, request_overrides={"extra_headers": {"x-test": "1"}}, )["extra_headers"] assert scope["x-test"] == "1" assert len(scope["session_id"]) <= 64 assert scope["x-client-request-id"] == scope["session_id"] if length == 64: assert scope["session_id"] == session_id else: assert scope["session_id"].startswith("pck_") assert scope["session_id"] != session_id def test_xai_injects_native_web_search_when_client_web_search_present(self, transport): """xAI path swaps a client-side ``web_search`` function for xAI's native server-side ``web_search`` built-in so grok server-side search runs to completion (otherwise the turn stalls as reasoning-with-no-answer -> false 'incomplete' -> 3 retries -> fail). Non-conflicting client tools are preserved. """ messages = [{"role": "user", "content": "Find current prices."}] kw = transport.build_kwargs( model="grok-composer-2.5-fast", messages=messages, tools=[ {"type": "function", "function": { "name": "read_file", "description": "Read a file.", "parameters": {"type": "object", "properties": {"path": {"type": "string"}}}}}, {"type": "function", "function": { "name": "web_search", "description": "Search the web.", "parameters": {"type": "object", "properties": {"query": {"type": "string"}}}}}, ], is_xai_responses=True, ) tool_types = [t.get("type") for t in kw.get("tools", [])] assert "web_search" in tool_types, kw.get("tools") # Non-conflicting client-side tools are preserved. names = [t.get("name") for t in kw.get("tools", []) if t.get("type") == "function"] assert "read_file" in names def test_xai_does_not_inject_native_web_search_without_client_web_search(self, transport): """The native ``web_search`` built-in is a 1:1 swap for an already-requested client ``web_search`` — NOT an additive grant. A turn whose toolset has no ``web_search`` (user never enabled the web toolset) must not get Grok server-side search force-injected, which would silently bypass Hermes's web-provider config and tool-trace plumbing for every xai-oauth turn. """ messages = [{"role": "user", "content": "Read this file."}] kw = transport.build_kwargs( model="grok-composer-2.5-fast", messages=messages, tools=[{"type": "function", "function": { "name": "read_file", "description": "Read a file.", "parameters": {"type": "object", "properties": {"path": {"type": "string"}}}}}], is_xai_responses=True, ) tools = kw.get("tools", []) assert not any(t.get("type") == "web_search" for t in tools), tools names = [t.get("name") for t in tools if t.get("type") == "function"] assert "read_file" in names def test_non_xai_path_does_not_inject_native_web_search(self, transport): """Native web_search injection is scoped to xAI — Codex/GitHub paths keep the client-side web_search function untouched.""" messages = [{"role": "user", "content": "Search."}] kw = transport.build_kwargs( model="gpt-5.4", messages=messages, tools=[{"type": "function", "function": { "name": "web_search", "description": "Search the web.", "parameters": {"type": "object", "properties": {"query": {"type": "string"}}}}}], is_xai_responses=False, ) tools = kw.get("tools", []) assert not any(t.get("type") == "web_search" for t in tools) assert any( t.get("type") == "function" and t.get("name") == "web_search" for t in tools ) # --- Grok reasoning-effort capability allowlist --- # api.x.ai 400s with "Model X does not support parameter reasoningEffort" # on grok-4 / grok-4-fast / grok-3 / grok-code-fast / grok-4.20-0309-*. # Those models reason natively but don't expose the dial. The transport # must omit the `reasoning` key for them. As of May 2026 we DO request # ``reasoning.encrypted_content`` back from xAI on every model — # see test_xai_reasoning_effort_passed for the rationale. def test_xai_grok_4_20_0309_variants_omit_reasoning_effort(self, transport): """grok-4.20-0309-(non-)reasoning reject the effort dial. Counterintuitively, only grok-4.20-multi-agent-0309 accepts it. """ messages = [{"role": "user", "content": "Hi"}] for model in ("grok-4.20-0309-reasoning", "grok-4.20-0309-non-reasoning"): kw = transport.build_kwargs( model=model, messages=messages, tools=[], is_xai_responses=True, reasoning_config={"effort": "high"}, ) assert "reasoning" not in kw, f"{model} must not receive reasoning" class TestCodexValidateResponse: def test_none_response(self, transport): assert transport.validate_response(None) is False def test_valid_output(self, transport): r = SimpleNamespace(output=[{"type": "message", "content": []}]) assert transport.validate_response(r) is True @pytest.mark.parametrize("reason", ["max_output_tokens", "length", "", None]) def test_empty_output_other_incomplete_reasons_remain_invalid(self, transport, reason): r = SimpleNamespace( status="incomplete", incomplete_details=SimpleNamespace(reason=reason), output=[], output_text="", ) assert transport.validate_response(r) is False class TestCodexMapFinishReason: def test_completed(self, transport): assert transport.map_finish_reason("completed") == "stop" class TestCodexNormalizeResponse: def test_text_response(self, transport): """Normalize a simple text Codex response.""" r = SimpleNamespace( output=[ SimpleNamespace( type="message", role="assistant", content=[SimpleNamespace(type="output_text", text="Hello world")], status="completed", ), ], status="completed", incomplete_details=None, usage=SimpleNamespace(input_tokens=10, output_tokens=5, input_tokens_details=None, output_tokens_details=None), ) nr = transport.normalize_response(r) assert isinstance(nr, NormalizedResponse) assert nr.content == "Hello world" assert nr.finish_reason == "stop" def test_message_items_preserved_in_provider_data(self, transport): """Codex assistant message item ids/phases must survive transport normalization.""" r = SimpleNamespace( output=[ SimpleNamespace( type="message", role="assistant", id="msg_abc", phase="final_answer", content=[SimpleNamespace(type="output_text", text="Hello world")], status="completed", ), ], status="completed", incomplete_details=None, usage=SimpleNamespace(input_tokens=10, output_tokens=5, input_tokens_details=None, output_tokens_details=None), ) nr = transport.normalize_response(r) assert nr.codex_message_items == [ { "type": "message", "role": "assistant", "status": "completed", "content": [{"type": "output_text", "text": "Hello world"}], "id": "msg_abc", "phase": "final_answer", } ] def test_tool_call_response(self, transport): """Normalize a Codex response with tool calls.""" r = SimpleNamespace( output=[ SimpleNamespace( type="function_call", call_id="call_abc123", name="terminal", arguments=json.dumps({"command": "ls"}), id="fc_abc123", status="completed", ), ], status="completed", incomplete_details=None, usage=SimpleNamespace(input_tokens=10, output_tokens=20, input_tokens_details=None, output_tokens_details=None), ) nr = transport.normalize_response(r) assert nr.finish_reason == "tool_calls" assert len(nr.tool_calls) == 1 tc = nr.tool_calls[0] assert tc.name == "terminal" assert '"command"' in tc.arguments class TestCodexTransportTimeout: """Forward per-request timeout from build_kwargs to the SDK kwargs.""" def test_positive_timeout_preserved(self, transport): kw = transport.build_kwargs( model="gpt-5.5", messages=[{"role": "user", "content": "hi"}], tools=[], timeout=600.0, ) assert kw.get("timeout") == 600.0 def test_inf_timeout_dropped(self, transport): kw = transport.build_kwargs( model="gpt-5.5", messages=[{"role": "user", "content": "hi"}], tools=[], timeout=float("inf"), ) assert "timeout" not in kw class TestCodexTransportXaiServiceTierStrip: """xAI Responses API rejects ``service_tier`` (#28490). ``resolve_fast_mode_overrides`` only returns ``service_tier`` for OpenAI fast-eligible models, so on paper the field should never reach a Grok request. But ``self.service_tier`` lingers across model switches and can also be set directly via ``agent.service_tier`` in config.yaml — both leak paths plumb through ``request_overrides`` and would 400 against xAI's ``/v1/responses``. Strip defensively when targeting xAI. """ @pytest.fixture def transport(self): from agent.transports.codex import ResponsesApiTransport return ResponsesApiTransport() def test_xai_strips_service_tier_from_request_overrides(self, transport): """Headline #28490 case: service_tier=priority leaks through request_overrides, must not reach the xAI request body.""" kw = transport.build_kwargs( model="grok-4.3", messages=[{"role": "user", "content": "hi"}], tools=[], is_xai_responses=True, request_overrides={"service_tier": "priority"}, ) assert "service_tier" not in kw, ( f"service_tier must be stripped on xAI requests, " f"got {kw.get('service_tier')!r}" ) def test_non_xai_codex_preserves_service_tier(self, transport): """The strip is xAI-only — native Codex DOES accept service_tier=priority (OpenAI Priority Processing). Stripping it elsewhere would silently disable the user's fast-mode opt-in. """ kw = transport.build_kwargs( model="gpt-5.5", messages=[{"role": "user", "content": "hi"}], tools=[], is_xai_responses=False, is_codex_backend=True, request_overrides={"service_tier": "priority"}, ) assert kw.get("service_tier") == "priority", ( "non-xAI codex_responses providers must keep service_tier" ) def test_github_responses_preserves_service_tier(self, transport): """GitHub Models (Copilot) is another codex_responses surface that should not be affected by the xAI strip.""" kw = transport.build_kwargs( model="gpt-5.5", messages=[{"role": "user", "content": "hi"}], tools=[], is_github_responses=True, request_overrides={"service_tier": "priority"}, ) assert kw.get("service_tier") == "priority" class TestPreflightSlashEnumStrip: """xAI Responses safety-net: strip slash-containing enum values when the model name indicates a Grok target (#28490). Native Codex accepts ``/``-containing enums; xAI rejects them with HTTP 400 "Invalid arguments passed to the model". The main agent loop and the auxiliary client already sanitize at request-build time; this preflight catches any future code path that bypasses those — gated on model name so we don't unnecessarily strip on non-xAI providers. """ def _make_kwargs(self, model: str, enum_values: list[str]) -> dict: return { "model": model, "instructions": "test", "input": [{"role": "user", "content": "hi"}], "tools": [ { "type": "function", "name": "pick_model", "description": "pick a model", "parameters": { "type": "object", "properties": { "model_id": { "type": "string", "enum": enum_values, }, }, }, }, ], } def test_grok_model_strips_slash_enum_values(self): """When the model name is Grok-family, slash-containing enum values are stripped so xAI doesn't 400 on the tool schema.""" from agent.codex_responses_adapter import _preflight_codex_api_kwargs kwargs = self._make_kwargs( "grok-4.3", ["Qwen/Qwen3.5-0.8B", "openai/gpt-oss-20b", "plain-id"], ) result = _preflight_codex_api_kwargs(kwargs) # The enum keyword itself is stripped (per strip_slash_enum's # semantics — it removes the constraint entirely when any value # contains /). params = result["tools"][0]["parameters"] assert "enum" not in params["properties"]["model_id"], ( "slash-containing enum must be stripped on Grok" ) def test_aggregator_prefixed_grok_also_strips(self): """Aggregator-prefixed (x-ai/grok-*) names hit the same path.""" from agent.codex_responses_adapter import _preflight_codex_api_kwargs kwargs = self._make_kwargs( "x-ai/grok-4.3", ["Qwen/Qwen3.5-0.8B"], ) result = _preflight_codex_api_kwargs(kwargs) assert "enum" not in result["tools"][0]["parameters"]["properties"]["model_id"] def test_non_grok_model_preserves_slash_enum_values(self): """Native Codex / GitHub Models DO accept slash-containing enums. The safety-net must NOT strip there or we silently degrade tool-schema constraints on every codex_responses provider that isn't xAI.""" from agent.codex_responses_adapter import _preflight_codex_api_kwargs kwargs = self._make_kwargs( "gpt-5.5", ["Qwen/Qwen3.5-0.8B", "plain-id"], ) result = _preflight_codex_api_kwargs(kwargs) params = result["tools"][0]["parameters"] # The enum must survive on non-xAI providers. assert params["properties"]["model_id"].get("enum") == [ "Qwen/Qwen3.5-0.8B", "plain-id" ]