{ "created": 1744679387.346831, "duration": 90.31976795196533, "exitcode": 1, "root": "/Users/erichuang/projects/llama-stack", "environment": {}, "summary": { "passed": 37, "failed": 39, "skipped": 2, "total": 78, "collected": 78 }, "collectors": [ { "nodeid": "", "outcome": "passed", "result": [ { "nodeid": "tests/verifications/openai_api/test_chat_completion.py", "type": "Module" } ] }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py", "outcome": "passed", "result": [ { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "type": "Function", "lineno": 74 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "type": "Function", "lineno": 93 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 117 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 117 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 117 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 136 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 136 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 136 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "type": "Function", "lineno": 160 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "type": "Function", "lineno": 183 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 205 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 205 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 205 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 229 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 229 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 229 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 257 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 257 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 257 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 281 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 281 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 281 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 308 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 308 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 308 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "type": "Function", "lineno": 331 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "type": "Function", "lineno": 331 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "type": "Function", "lineno": 331 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "type": "Function", "lineno": 359 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "type": "Function", "lineno": 450 }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "type": "Function", "lineno": 450 } ] } ], "tests": [ { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "earth" }, "setup": { "duration": 0.1559112500399351, "outcome": "passed" }, "call": { "duration": 0.3692209171131253, "outcome": "passed" }, "teardown": { "duration": 0.00021362490952014923, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "saturn" }, "setup": { "duration": 0.007326166843995452, "outcome": "passed" }, "call": { "duration": 0.49173945817165077, "outcome": "passed" }, "teardown": { "duration": 0.00034487503580749035, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "earth" }, "setup": { "duration": 0.021014458034187555, "outcome": "passed" }, "call": { "duration": 0.36956487502902746, "outcome": "passed" }, "teardown": { "duration": 0.0007119579240679741, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "saturn" }, "setup": { "duration": 0.011922625126317143, "outcome": "passed" }, "call": { "duration": 2.7763332079630345, "outcome": "passed" }, "teardown": { "duration": 0.0004842919297516346, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "earth" }, "setup": { "duration": 0.023896750062704086, "outcome": "passed" }, "call": { "duration": 0.9817597079090774, "outcome": "passed" }, "teardown": { "duration": 0.0004768748767673969, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "lineno": 74, "outcome": "passed", "keywords": [ "test_chat_non_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "saturn" }, "setup": { "duration": 0.07423937506973743, "outcome": "passed" }, "call": { "duration": 0.3721332079730928, "outcome": "passed" }, "teardown": { "duration": 0.00020033284090459347, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "lineno": 93, "outcome": "passed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-earth]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "earth" }, "setup": { "duration": 0.010166750056669116, "outcome": "passed" }, "call": { "duration": 0.41266337502747774, "outcome": "passed" }, "teardown": { "duration": 0.00034358282573521137, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "lineno": 93, "outcome": "passed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "saturn" }, "setup": { "duration": 0.016687541967257857, "outcome": "passed" }, "call": { "duration": 0.7235856249462813, "outcome": "passed" }, "teardown": { "duration": 0.00027179205790162086, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "lineno": 93, "outcome": "failed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-earth]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "earth" }, "setup": { "duration": 0.012556416913866997, "outcome": "passed" }, "call": { "duration": 0.27039612480439246, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'earth', 'input': {'messages': [{'content': 'Which planet do humans live on?', 'role': 'user'}]}, 'output': 'Earth'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_basic\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_basic(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:111: IndexError" }, "teardown": { "duration": 0.0002312080468982458, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "lineno": 93, "outcome": "failed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "saturn" }, "setup": { "duration": 0.006413874914869666, "outcome": "passed" }, "call": { "duration": 0.36463545891456306, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'saturn', 'input': {'messages': [{'content': 'Which planet has rings around it with a name starting with letter S?', 'role': 'user'}]}, 'output': 'Saturn'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_basic\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_basic(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:111: IndexError" }, "teardown": { "duration": 0.00023154192604124546, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "lineno": 93, "outcome": "failed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-earth", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "earth" }, "setup": { "duration": 0.015633082948625088, "outcome": "passed" }, "call": { "duration": 0.8896284159272909, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'earth', 'input': {'messages': [{'content': 'Which planet do humans live on?', 'role': 'user'}]}, 'output': 'Earth'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_basic\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_basic(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:111: IndexError" }, "teardown": { "duration": 0.0006587498355656862, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "lineno": 93, "outcome": "failed", "keywords": [ "test_chat_streaming_basic[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-saturn", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "saturn" }, "setup": { "duration": 0.012669583084061742, "outcome": "passed" }, "call": { "duration": 0.3499396659899503, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 111, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'saturn', 'input': {'messages': [{'content': 'Which planet has rings around it with a name starting with letter S?', 'role': 'user'}]}, 'output': 'Saturn'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_basic\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_basic(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:111: IndexError" }, "teardown": { "duration": 0.00024912506341934204, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 117, "outcome": "skipped", "keywords": [ "test_chat_non_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.0153201250359416, "outcome": "passed" }, "call": { "duration": 0.0001901669893413782, "outcome": "skipped", "longrepr": "('/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py', 126, 'Skipped: Skipping test_chat_non_streaming_image for model meta-llama/Llama-3.3-70B-Instruct-Turbo on provider together based on config.')" }, "teardown": { "duration": 0.00012779212556779385, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 117, "outcome": "passed", "keywords": [ "test_chat_non_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.008855124935507774, "outcome": "passed" }, "call": { "duration": 1.37906050006859, "outcome": "passed" }, "teardown": { "duration": 0.0004904591478407383, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 117, "outcome": "passed", "keywords": [ "test_chat_non_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.017166708130389452, "outcome": "passed" }, "call": { "duration": 4.003400916932151, "outcome": "passed" }, "teardown": { "duration": 0.00042724981904029846, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 136, "outcome": "skipped", "keywords": [ "test_chat_streaming_image[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.007232750067487359, "outcome": "passed" }, "call": { "duration": 0.0001449580304324627, "outcome": "skipped", "longrepr": "('/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py', 145, 'Skipped: Skipping test_chat_streaming_image for model meta-llama/Llama-3.3-70B-Instruct-Turbo on provider together based on config.')" }, "teardown": { "duration": 0.0001349160447716713, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 136, "outcome": "failed", "keywords": [ "test_chat_streaming_image[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.007052165921777487, "outcome": "passed" }, "call": { "duration": 1.4663615000899881, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 154, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 154, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': [{'text': 'What is in this image?', 'type': 'text'}, {'image_url': {...}, 'type': 'image_url'}], 'role': 'user'}]}, 'output': 'llama'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_image\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_image(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:154: IndexError" }, "teardown": { "duration": 0.0005696250591427088, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 136, "outcome": "failed", "keywords": [ "test_chat_streaming_image[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.01214433298446238, "outcome": "passed" }, "call": { "duration": 3.902559082955122, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 154, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 154, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': [{'text': 'What is in this image?', 'type': 'text'}, {'image_url': {...}, 'type': 'image_url'}], 'role': 'user'}]}, 'output': 'llama'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_image\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_image(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n stream=True,\n )\n content = \"\"\n for chunk in response:\n> content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:154: IndexError" }, "teardown": { "duration": 0.000591374933719635, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "calendar" }, "setup": { "duration": 0.01478054211474955, "outcome": "passed" }, "call": { "duration": 0.569845792138949, "outcome": "passed" }, "teardown": { "duration": 0.00038724998012185097, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "math" }, "setup": { "duration": 0.014717916958034039, "outcome": "passed" }, "call": { "duration": 1.1819656670559198, "outcome": "passed" }, "teardown": { "duration": 0.0002410421147942543, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "calendar" }, "setup": { "duration": 0.006486707832664251, "outcome": "passed" }, "call": { "duration": 0.5623017910402268, "outcome": "passed" }, "teardown": { "duration": 0.00032504182308912277, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "math" }, "setup": { "duration": 0.009171125013381243, "outcome": "passed" }, "call": { "duration": 2.6005691669415683, "outcome": "passed" }, "teardown": { "duration": 0.00023995805531740189, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "calendar" }, "setup": { "duration": 0.009700333932414651, "outcome": "passed" }, "call": { "duration": 0.4192442081402987, "outcome": "passed" }, "teardown": { "duration": 0.00040241610258817673, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "lineno": 160, "outcome": "passed", "keywords": [ "test_chat_non_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "math" }, "setup": { "duration": 0.006938542006537318, "outcome": "passed" }, "call": { "duration": 2.1736337919719517, "outcome": "passed" }, "teardown": { "duration": 0.00019279099069535732, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "lineno": 183, "outcome": "passed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "calendar" }, "setup": { "duration": 0.008775749942287803, "outcome": "passed" }, "call": { "duration": 0.5588400410488248, "outcome": "passed" }, "teardown": { "duration": 0.00040091690607368946, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "lineno": 183, "outcome": "passed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-3.3-70B-Instruct-Turbo-math]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "math" }, "setup": { "duration": 0.01844154205173254, "outcome": "passed" }, "call": { "duration": 2.205772665794939, "outcome": "passed" }, "teardown": { "duration": 0.00021091708913445473, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "lineno": 183, "outcome": "failed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "calendar" }, "setup": { "duration": 0.015595750184729695, "outcome": "passed" }, "call": { "duration": 0.6904467919375747, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'calendar', 'input': {'messages': [{'content': 'Extract the event information.', 'role': 'system'}, {'cont...articipants'], 'title': 'CalendarEvent', 'type': 'object'}}, 'type': 'json_schema'}}, 'output': 'valid_calendar_event'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_structured_output\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_structured_output(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n response_format=case[\"input\"][\"response_format\"],\n stream=True,\n )\n maybe_json_content = \"\"\n for chunk in response:\n> maybe_json_content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:202: IndexError" }, "teardown": { "duration": 0.0002907498273998499, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "lineno": 183, "outcome": "failed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-4-Scout-17B-16E-Instruct-math]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "math" }, "setup": { "duration": 0.008272957988083363, "outcome": "passed" }, "call": { "duration": 3.499622541014105, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'math', 'input': {'messages': [{'content': 'You are a helpful math tutor. Guide the user through the solut... ['steps', 'final_answer'], 'title': 'MathReasoning', ...}}, 'type': 'json_schema'}}, 'output': 'valid_math_reasoning'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_structured_output\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_structured_output(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n response_format=case[\"input\"][\"response_format\"],\n stream=True,\n )\n maybe_json_content = \"\"\n for chunk in response:\n> maybe_json_content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:202: IndexError" }, "teardown": { "duration": 0.0005947079043835402, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "lineno": 183, "outcome": "failed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-calendar", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "calendar" }, "setup": { "duration": 0.013340875040739775, "outcome": "passed" }, "call": { "duration": 0.42789591709151864, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'calendar', 'input': {'messages': [{'content': 'Extract the event information.', 'role': 'system'}, {'cont...articipants'], 'title': 'CalendarEvent', 'type': 'object'}}, 'type': 'json_schema'}}, 'output': 'valid_calendar_event'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_structured_output\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_structured_output(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n response_format=case[\"input\"][\"response_format\"],\n stream=True,\n )\n maybe_json_content = \"\"\n for chunk in response:\n> maybe_json_content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:202: IndexError" }, "teardown": { "duration": 0.0003039578441530466, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "lineno": 183, "outcome": "failed", "keywords": [ "test_chat_streaming_structured_output[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-math", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "math" }, "setup": { "duration": 0.01058275019749999, "outcome": "passed" }, "call": { "duration": 5.795635707909241, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 202, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'math', 'input': {'messages': [{'content': 'You are a helpful math tutor. Guide the user through the solut... ['steps', 'final_answer'], 'title': 'MathReasoning', ...}}, 'type': 'json_schema'}}, 'output': 'valid_math_reasoning'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_chat_structured_output\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_structured_output(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n response_format=case[\"input\"][\"response_format\"],\n stream=True,\n )\n maybe_json_content = \"\"\n for chunk in response:\n> maybe_json_content += chunk.choices[0].delta.content or \"\"\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:202: IndexError" }, "teardown": { "duration": 0.0005178749561309814, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 205, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.014336749911308289, "outcome": "passed" }, "call": { "duration": 0.451304541900754, "outcome": "passed" }, "teardown": { "duration": 0.0004718329291790724, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 205, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.01625004201196134, "outcome": "passed" }, "call": { "duration": 0.5111537908669561, "outcome": "passed" }, "teardown": { "duration": 0.00046774977818131447, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 205, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.015832332894206047, "outcome": "passed" }, "call": { "duration": 0.8238586660008878, "outcome": "passed" }, "teardown": { "duration": 0.0006185418460518122, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 229, "outcome": "passed", "keywords": [ "test_chat_streaming_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.007832166040316224, "outcome": "passed" }, "call": { "duration": 0.685583250131458, "outcome": "passed" }, "teardown": { "duration": 0.0004414590075612068, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 229, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.021764083998277783, "outcome": "passed" }, "call": { "duration": 0.35617320891469717, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 247, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_calling(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n stream=True,\n )\n \n> _, tool_calls_buffer = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:247: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0005425831768661737, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 229, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.016708041075617075, "outcome": "passed" }, "call": { "duration": 0.49443637509830296, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 247, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"],\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_calling(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n stream=True,\n )\n \n> _, tool_calls_buffer = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:247: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0002642078325152397, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 257, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.009570583933964372, "outcome": "passed" }, "call": { "duration": 0.5232214999850839, "outcome": "passed" }, "teardown": { "duration": 0.0006591668352484703, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 257, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.01567283389158547, "outcome": "passed" }, "call": { "duration": 0.4465816249139607, "outcome": "passed" }, "teardown": { "duration": 0.0003922500181943178, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 257, "outcome": "passed", "keywords": [ "test_chat_non_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.021711332956328988, "outcome": "passed" }, "call": { "duration": 0.5361095829866827, "outcome": "passed" }, "teardown": { "duration": 0.0003099590539932251, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 281, "outcome": "passed", "keywords": [ "test_chat_streaming_tool_choice_required[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.009334125090390444, "outcome": "passed" }, "call": { "duration": 0.5789772500284016, "outcome": "passed" }, "teardown": { "duration": 0.00037712487392127514, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 281, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.019614499993622303, "outcome": "passed" }, "call": { "duration": 0.444399792002514, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 300, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_choice_required(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"required\", # Force tool call\n stream=True,\n )\n \n> _, tool_calls_buffer = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:300: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0004192921333014965, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 281, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_choice_required[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.012822834076359868, "outcome": "passed" }, "call": { "duration": 0.6777042911853641, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 300, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_choice_required(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"required\", # Force tool call\n stream=True,\n )\n \n> _, tool_calls_buffer = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:300: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0004483328666538, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 308, "outcome": "failed", "keywords": [ "test_chat_non_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.011924332939088345, "outcome": "passed" }, "call": { "duration": 0.4756374170538038, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError: Expected no tool calls when tool_choice='none'\nassert [ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)] is None\n + where [ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]).tool_calls\n + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]), seed=13421903014786785000).message" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_non_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=False,\n )\n \n assert response.choices[0].message.role == \"assistant\"\n> assert response.choices[0].message.tool_calls is None, \"Expected no tool calls when tool_choice='none'\"\nE AssertionError: Expected no tool calls when tool_choice='none'\nE assert [ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)] is None\nE + where [ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]).tool_calls\nE + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_nfd8oz9wmhlwf4mqglcaokrs', function=Function(arguments='{\"location\":\"San Francisco, USA\"}', name='get_weather'), type='function', index=0)]), seed=13421903014786785000).message\n\ntests/verifications/openai_api/test_chat_completion.py:328: AssertionError" }, "teardown": { "duration": 0.0004585420247167349, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 308, "outcome": "failed", "keywords": [ "test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.013246082933619618, "outcome": "passed" }, "call": { "duration": 0.5618870409671217, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError: Expected no tool calls when tool_choice='none'\nassert [ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)] is None\n + where [ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]).tool_calls\n + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]), seed=None).message" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_non_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=False,\n )\n \n assert response.choices[0].message.role == \"assistant\"\n> assert response.choices[0].message.tool_calls is None, \"Expected no tool calls when tool_choice='none'\"\nE AssertionError: Expected no tool calls when tool_choice='none'\nE assert [ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)] is None\nE + where [ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]).tool_calls\nE + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_h5u55eksczab7xtg8oot43k5', function=Function(arguments='{\"location\":\"San Francisco, United States\"}', name='get_weather'), type='function', index=0)]), seed=None).message\n\ntests/verifications/openai_api/test_chat_completion.py:328: AssertionError" }, "teardown": { "duration": 0.00025883293710649014, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 308, "outcome": "failed", "keywords": [ "test_chat_non_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.008055417099967599, "outcome": "passed" }, "call": { "duration": 0.32869591703638434, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError: Expected no tool calls when tool_choice='none'\nassert [ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)] is None\n + where [ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]).tool_calls\n + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]), seed=None).message" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 328, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_non_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n response = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=False,\n )\n \n assert response.choices[0].message.role == \"assistant\"\n> assert response.choices[0].message.tool_calls is None, \"Expected no tool calls when tool_choice='none'\"\nE AssertionError: Expected no tool calls when tool_choice='none'\nE assert [ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)] is None\nE + where [ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]).tool_calls\nE + where ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]) = Choice(finish_reason='tool_calls', index=0, logprobs=None, message=ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_s1bz9v57b8uizqy2i81869pb', function=Function(arguments='{\"location\":\"San Francisco\"}', name='get_weather'), type='function', index=0)]), seed=None).message\n\ntests/verifications/openai_api/test_chat_completion.py:328: AssertionError" }, "teardown": { "duration": 0.0003937501460313797, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "lineno": 331, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_choice_none[meta-llama/Llama-3.3-70B-Instruct-Turbo-case0]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "case0" }, "setup": { "duration": 0.013460749993100762, "outcome": "passed" }, "call": { "duration": 0.35879983310587704, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError: Expected no tool call chunks when tool_choice='none'\nassert not [ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\n + where [ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=True,\n )\n \n content = \"\"\n for chunk in stream:\n delta = chunk.choices[0].delta\n if delta.content:\n content += delta.content\n> assert not delta.tool_calls, \"Expected no tool call chunks when tool_choice='none'\"\nE AssertionError: Expected no tool call chunks when tool_choice='none'\nE assert not [ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\nE + where [ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_q472clmnii99ps1fxqtv8qvr', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls\n\ntests/verifications/openai_api/test_chat_completion.py:355: AssertionError" }, "teardown": { "duration": 0.0002649170346558094, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "lineno": 331, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Scout-17B-16E-Instruct-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "case0" }, "setup": { "duration": 0.0068365419283509254, "outcome": "passed" }, "call": { "duration": 0.5351063329726458, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError: Expected no tool call chunks when tool_choice='none'\nassert not [ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\n + where [ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=True,\n )\n \n content = \"\"\n for chunk in stream:\n delta = chunk.choices[0].delta\n if delta.content:\n content += delta.content\n> assert not delta.tool_calls, \"Expected no tool call chunks when tool_choice='none'\"\nE AssertionError: Expected no tool call chunks when tool_choice='none'\nE assert not [ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\nE + where [ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_l3roc57o2pn9b70f0dcgil53', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls\n\ntests/verifications/openai_api/test_chat_completion.py:355: AssertionError" }, "teardown": { "duration": 0.0004712918307632208, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "lineno": 331, "outcome": "failed", "keywords": [ "test_chat_streaming_tool_choice_none[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-case0", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "case0" }, "setup": { "duration": 0.014073874801397324, "outcome": "passed" }, "call": { "duration": 0.6729549579322338, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError: Expected no tool call chunks when tool_choice='none'\nassert not [ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\n + where [ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 355, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'input': {'messages': [{'content': 'You are a helpful assistant that can use tools to get information.', 'role': 'sys..., 'properties': {...}, 'required': [...], 'type': 'object'}}, 'type': 'function'}]}, 'output': 'get_weather_tool_call'}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases[\"test_tool_calling\"][\"test_params\"][\"case\"], # Reusing existing case for now\n ids=case_id_generator,\n )\n def test_chat_streaming_tool_choice_none(request, openai_client, model, provider, verification_config, case):\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n stream = openai_client.chat.completions.create(\n model=model,\n messages=case[\"input\"][\"messages\"],\n tools=case[\"input\"][\"tools\"],\n tool_choice=\"none\",\n stream=True,\n )\n \n content = \"\"\n for chunk in stream:\n delta = chunk.choices[0].delta\n if delta.content:\n content += delta.content\n> assert not delta.tool_calls, \"Expected no tool call chunks when tool_choice='none'\"\nE AssertionError: Expected no tool call chunks when tool_choice='none'\nE assert not [ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]\nE + where [ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')] = ChoiceDelta(content=None, function_call=None, refusal=None, role=None, tool_calls=[ChoiceDeltaToolCall(index=0, id='call_ktw831i0p838mzvnnaylf6fp', function=ChoiceDeltaToolCallFunction(arguments='', name='get_weather'), type='function')]).tool_calls\n\ntests/verifications/openai_api/test_chat_completion.py:355: AssertionError" }, "teardown": { "duration": 0.000251916004344821, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "lineno": 359, "outcome": "failed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.009340125136077404, "outcome": "passed" }, "call": { "duration": 0.3328715830575675, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 418, "message": "AssertionError: Expected 0 tool calls, but got 1\nassert 1 == 0\n + where 1 = len(([ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)]))\n + where [ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)]).tool_calls" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 418, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_non_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\"\n Test cases for multi-turn tool calling.\n Tool calls are asserted.\n Tool responses are provided in the test case.\n Final response is asserted.\n \"\"\"\n \n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n # Create a copy of the messages list to avoid modifying the original\n messages = []\n tools = case[\"input\"][\"tools\"]\n # Use deepcopy to prevent modification across runs/parametrization\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n # keep going until either\n # 1. we have messages to test in multi-turn\n # 2. no messages but last message is tool response\n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n # do not take new messages if last message is tool response\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n # Ensure new_messages is a list of message objects\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n # If it's a single message object, add it directly\n messages.append(new_messages)\n \n # --- API Call ---\n response = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=False,\n )\n \n # --- Process Response ---\n assistant_message = response.choices[0].message\n messages.append(assistant_message.model_dump(exclude_unset=True))\n \n assert assistant_message.role == \"assistant\"\n \n # Get the expected result data\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n # --- Assertions based on expected result ---\n> assert len(assistant_message.tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(assistant_message.tool_calls or [])}\"\n )\nE AssertionError: Expected 0 tool calls, but got 1\nE assert 1 == 0\nE + where 1 = len(([ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)]))\nE + where [ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_3rr948zuvun0533y4oyyep0z', function=Function(arguments='{\"location\":\"San Francisco, CA\"}', name='get_weather'), type='function', index=0)]).tool_calls\n\ntests/verifications/openai_api/test_chat_completion.py:418: AssertionError" }, "teardown": { "duration": 0.00042020808905363083, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.01490145898424089, "outcome": "passed" }, "call": { "duration": 0.8346118750050664, "outcome": "passed" }, "teardown": { "duration": 0.00034404080361127853, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "add_product_tool" }, "setup": { "duration": 0.014493625145405531, "outcome": "passed" }, "call": { "duration": 0.8973606249783188, "outcome": "passed" }, "teardown": { "duration": 0.00021345820277929306, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.009358166949823499, "outcome": "passed" }, "call": { "duration": 4.5295154170598835, "outcome": "passed" }, "teardown": { "duration": 0.0002461671829223633, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "lineno": 359, "outcome": "failed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.009552374947816133, "outcome": "passed" }, "call": { "duration": 0.34176899981684983, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 429, "message": "AssertionError: Expected arguments '{'month': 1, 'year': 2025}', got '{'month': '1', 'year': '2025'}'\nassert {'month': '1', 'year': '2025'} == {'month': 1, 'year': 2025}\n \n Differing items:\n {'month': '1'} != {'month': 1}\n {'year': '2025'} != {'year': 2025}\n \n Full diff:\n {...\n \n ...Full output truncated (7 lines hidden), use '-vv' to show" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 429, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'compare_monthly_expense_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'month': 1, 'year': ... 'Total expenses for January 2025: $1000'}\"}, {'response': \"{'response': 'Total expenses for February 2024: $2000'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_non_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\"\n Test cases for multi-turn tool calling.\n Tool calls are asserted.\n Tool responses are provided in the test case.\n Final response is asserted.\n \"\"\"\n \n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n # Create a copy of the messages list to avoid modifying the original\n messages = []\n tools = case[\"input\"][\"tools\"]\n # Use deepcopy to prevent modification across runs/parametrization\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n # keep going until either\n # 1. we have messages to test in multi-turn\n # 2. no messages but last message is tool response\n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n # do not take new messages if last message is tool response\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n # Ensure new_messages is a list of message objects\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n # If it's a single message object, add it directly\n messages.append(new_messages)\n \n # --- API Call ---\n response = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=False,\n )\n \n # --- Process Response ---\n assistant_message = response.choices[0].message\n messages.append(assistant_message.model_dump(exclude_unset=True))\n \n assert assistant_message.role == \"assistant\"\n \n # Get the expected result data\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n # --- Assertions based on expected result ---\n assert len(assistant_message.tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(assistant_message.tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n tool_call = assistant_message.tool_calls[0]\n assert tool_call.function.name == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call.function.name}'\"\n )\n # Parse the JSON string arguments before comparing\n actual_arguments = json.loads(tool_call.function.arguments)\n> assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\nE AssertionError: Expected arguments '{'month': 1, 'year': 2025}', got '{'month': '1', 'year': '2025'}'\nE assert {'month': '1', 'year': '2025'} == {'month': 1, 'year': 2025}\nE \nE Differing items:\nE {'month': '1'} != {'month': 1}\nE {'year': '2025'} != {'year': 2025}\nE \nE Full diff:\nE {...\nE \nE ...Full output truncated (7 lines hidden), use '-vv' to show\n\ntests/verifications/openai_api/test_chat_completion.py:429: AssertionError" }, "teardown": { "duration": 0.000527665950357914, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "lineno": 359, "outcome": "failed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.012501416960731149, "outcome": "passed" }, "call": { "duration": 1.585734374821186, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 418, "message": "AssertionError: Expected 0 tool calls, but got 2\nassert 2 == 0\n + where 2 = len(([ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)]))\n + where [ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)]).tool_calls" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 418, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_non_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\"\n Test cases for multi-turn tool calling.\n Tool calls are asserted.\n Tool responses are provided in the test case.\n Final response is asserted.\n \"\"\"\n \n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n # Create a copy of the messages list to avoid modifying the original\n messages = []\n tools = case[\"input\"][\"tools\"]\n # Use deepcopy to prevent modification across runs/parametrization\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n # keep going until either\n # 1. we have messages to test in multi-turn\n # 2. no messages but last message is tool response\n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n # do not take new messages if last message is tool response\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n # Ensure new_messages is a list of message objects\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n # If it's a single message object, add it directly\n messages.append(new_messages)\n \n # --- API Call ---\n response = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=False,\n )\n \n # --- Process Response ---\n assistant_message = response.choices[0].message\n messages.append(assistant_message.model_dump(exclude_unset=True))\n \n assert assistant_message.role == \"assistant\"\n \n # Get the expected result data\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n # --- Assertions based on expected result ---\n> assert len(assistant_message.tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(assistant_message.tool_calls or [])}\"\n )\nE AssertionError: Expected 0 tool calls, but got 2\nE assert 2 == 0\nE + where 2 = len(([ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)]))\nE + where [ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)] = ChatCompletionMessage(content=None, refusal=None, role='assistant', annotations=None, audio=None, function_call=None, tool_calls=[ChatCompletionMessageToolCall(id='call_4fm3kj059swz9no94n6fg54d', function=Function(arguments='{\"location\":\"Sun, NA\"}', name='get_weather'), type='function', index=0), ChatCompletionMessageToolCall(id='call_lzc5lo7y2p7wjyquvmvvzt64', function=Function(arguments='{\"name\":\"Sun\"}', name='get_latin_name'), type='function', index=1)]).tool_calls\n\ntests/verifications/openai_api/test_chat_completion.py:418: AssertionError" }, "teardown": { "duration": 0.0003941669128835201, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.014057958032935858, "outcome": "passed" }, "call": { "duration": 0.7121559998486191, "outcome": "passed" }, "teardown": { "duration": 0.00048266700468957424, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "add_product_tool" }, "setup": { "duration": 0.02072141715325415, "outcome": "passed" }, "call": { "duration": 1.0424797078594565, "outcome": "passed" }, "teardown": { "duration": 0.0004878339823335409, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.018570583080872893, "outcome": "passed" }, "call": { "duration": 3.4340267919469625, "outcome": "passed" }, "teardown": { "duration": 0.00023016706109046936, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.009570334106683731, "outcome": "passed" }, "call": { "duration": 2.2068665840197355, "outcome": "passed" }, "teardown": { "duration": 0.00051837507635355, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "lineno": 359, "outcome": "failed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.01873366697691381, "outcome": "passed" }, "call": { "duration": 0.5193468749057502, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 446, "message": "AssertionError: Expected one of ['sol'] in content, but got: '{\"name\": null, \"parameters\": null}'\nassert False\n + where False = any(. at 0x10e4c0f90>)" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 446, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_non_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\"\n Test cases for multi-turn tool calling.\n Tool calls are asserted.\n Tool responses are provided in the test case.\n Final response is asserted.\n \"\"\"\n \n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n # Create a copy of the messages list to avoid modifying the original\n messages = []\n tools = case[\"input\"][\"tools\"]\n # Use deepcopy to prevent modification across runs/parametrization\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n # keep going until either\n # 1. we have messages to test in multi-turn\n # 2. no messages but last message is tool response\n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n # do not take new messages if last message is tool response\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n # Ensure new_messages is a list of message objects\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n # If it's a single message object, add it directly\n messages.append(new_messages)\n \n # --- API Call ---\n response = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=False,\n )\n \n # --- Process Response ---\n assistant_message = response.choices[0].message\n messages.append(assistant_message.model_dump(exclude_unset=True))\n \n assert assistant_message.role == \"assistant\"\n \n # Get the expected result data\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n # --- Assertions based on expected result ---\n assert len(assistant_message.tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(assistant_message.tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n tool_call = assistant_message.tool_calls[0]\n assert tool_call.function.name == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call.function.name}'\"\n )\n # Parse the JSON string arguments before comparing\n actual_arguments = json.loads(tool_call.function.arguments)\n assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\n \n # Prepare and append the tool response for the next turn\n tool_response = tool_responses.pop(0)\n messages.append(\n {\n \"role\": \"tool\",\n \"tool_call_id\": tool_call.id,\n \"content\": tool_response[\"response\"],\n }\n )\n else:\n assert assistant_message.content is not None, \"Expected content, but none received.\"\n expected_answers = expected[\"answer\"] # This is now a list\n content_lower = assistant_message.content.lower()\n> assert any(ans.lower() in content_lower for ans in expected_answers), (\n f\"Expected one of {expected_answers} in content, but got: '{assistant_message.content}'\"\n )\nE AssertionError: Expected one of ['sol'] in content, but got: '{\"name\": null, \"parameters\": null}'\nE assert False\nE + where False = any(. at 0x10e4c0f90>)\n\ntests/verifications/openai_api/test_chat_completion.py:446: AssertionError" }, "teardown": { "duration": 0.0004933748859912157, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.014272749889642, "outcome": "passed" }, "call": { "duration": 1.911199334077537, "outcome": "passed" }, "teardown": { "duration": 0.00043049990199506283, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "add_product_tool" }, "setup": { "duration": 0.031040542060509324, "outcome": "passed" }, "call": { "duration": 3.0026419160421938, "outcome": "passed" }, "teardown": { "duration": 0.00045104208402335644, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "lineno": 359, "outcome": "failed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.016529500018805265, "outcome": "passed" }, "call": { "duration": 2.7563346249517053, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 429, "message": "AssertionError: Expected arguments '{'name': 'Team Building', 'date': '2025-03-03', 'time': '10:00', 'location': 'Main Conference Room', 'participants': ['Alice', 'Bob', 'Charlie']}', got '{'participants': '[\"Alice\", \"Bob\", \"Charlie\"]', 'location': 'Main Conference Room', 'name': 'Team Building', 'date': '2025-03-03', 'time': '10:00'}'\nassert {'date': '202...arlie\"]', ...} == {'date': '202...harlie'], ...}\n \n Omitting 4 identical items, use -vv to show\n Differing items:\n {'participants': '[\"Alice\", \"Bob\", \"Charlie\"]'} != {'participants': ['Alice', 'Bob', 'Charlie']}\n \n Full diff:\n {...\n \n ...Full output truncated (11 lines hidden), use '-vv' to show" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 429, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'get_then_create_event_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'date': '2025-03-03', ...ents found for 2025-03-03 at 10:00'}\"}, {'response': \"{'response': 'Successfully created new event with id: e_123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_non_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\"\n Test cases for multi-turn tool calling.\n Tool calls are asserted.\n Tool responses are provided in the test case.\n Final response is asserted.\n \"\"\"\n \n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n # Create a copy of the messages list to avoid modifying the original\n messages = []\n tools = case[\"input\"][\"tools\"]\n # Use deepcopy to prevent modification across runs/parametrization\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n # keep going until either\n # 1. we have messages to test in multi-turn\n # 2. no messages but last message is tool response\n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n # do not take new messages if last message is tool response\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n # Ensure new_messages is a list of message objects\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n # If it's a single message object, add it directly\n messages.append(new_messages)\n \n # --- API Call ---\n response = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=False,\n )\n \n # --- Process Response ---\n assistant_message = response.choices[0].message\n messages.append(assistant_message.model_dump(exclude_unset=True))\n \n assert assistant_message.role == \"assistant\"\n \n # Get the expected result data\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n # --- Assertions based on expected result ---\n assert len(assistant_message.tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(assistant_message.tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n tool_call = assistant_message.tool_calls[0]\n assert tool_call.function.name == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call.function.name}'\"\n )\n # Parse the JSON string arguments before comparing\n actual_arguments = json.loads(tool_call.function.arguments)\n> assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\nE AssertionError: Expected arguments '{'name': 'Team Building', 'date': '2025-03-03', 'time': '10:00', 'location': 'Main Conference Room', 'participants': ['Alice', 'Bob', 'Charlie']}', got '{'participants': '[\"Alice\", \"Bob\", \"Charlie\"]', 'location': 'Main Conference Room', 'name': 'Team Building', 'date': '2025-03-03', 'time': '10:00'}'\nE assert {'date': '202...arlie\"]', ...} == {'date': '202...harlie'], ...}\nE \nE Omitting 4 identical items, use -vv to show\nE Differing items:\nE {'participants': '[\"Alice\", \"Bob\", \"Charlie\"]'} != {'participants': ['Alice', 'Bob', 'Charlie']}\nE \nE Full diff:\nE {...\nE \nE ...Full output truncated (11 lines hidden), use '-vv' to show\n\ntests/verifications/openai_api/test_chat_completion.py:429: AssertionError" }, "teardown": { "duration": 0.0005542081780731678, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "lineno": 359, "outcome": "passed", "keywords": [ "test_chat_non_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.013607957866042852, "outcome": "passed" }, "call": { "duration": 3.0105869588442147, "outcome": "passed" }, "teardown": { "duration": 0.0004793750122189522, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.01806124998256564, "outcome": "passed" }, "call": { "duration": 0.3295827910769731, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 500, "message": "AssertionError: Expected 0 tool calls, but got 1\nassert 1 == 0\n + where 1 = len(([{'function': {'arguments': '{\"location\":\"San Francisco, CA\"}', 'name': 'get_weather'}, 'id': 'call_l066e8oey2i8exeodczlv1mh', 'type': 'function'}]))" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 500, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n \n # --- Construct Assistant Message for History ---\n assistant_message_dict = {\"role\": \"assistant\"}\n if accumulated_content:\n assistant_message_dict[\"content\"] = accumulated_content\n if accumulated_tool_calls:\n assistant_message_dict[\"tool_calls\"] = accumulated_tool_calls\n \n messages.append(assistant_message_dict)\n \n # --- Assertions ---\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n> assert len(accumulated_tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(accumulated_tool_calls or [])}\"\n )\nE AssertionError: Expected 0 tool calls, but got 1\nE assert 1 == 0\nE + where 1 = len(([{'function': {'arguments': '{\"location\":\"San Francisco, CA\"}', 'name': 'get_weather'}, 'id': 'call_l066e8oey2i8exeodczlv1mh', 'type': 'function'}]))\n\ntests/verifications/openai_api/test_chat_completion.py:500: AssertionError" }, "teardown": { "duration": 0.0002942080609500408, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.007637625094503164, "outcome": "passed" }, "call": { "duration": 2.021851292112842, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 526, "message": "AssertionError: Expected content, but none received.\nassert ('' is not None and '' != '')" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 526, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'weather_tool_then_text', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'location': 'San Francisco...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n \n # --- Construct Assistant Message for History ---\n assistant_message_dict = {\"role\": \"assistant\"}\n if accumulated_content:\n assistant_message_dict[\"content\"] = accumulated_content\n if accumulated_tool_calls:\n assistant_message_dict[\"tool_calls\"] = accumulated_tool_calls\n \n messages.append(assistant_message_dict)\n \n # --- Assertions ---\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n assert len(accumulated_tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(accumulated_tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n # Use the first accumulated tool call for assertion\n tool_call = accumulated_tool_calls[0]\n assert tool_call[\"function\"][\"name\"] == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call['function']['name']}'\"\n )\n # Parse the accumulated arguments string for comparison\n actual_arguments = json.loads(tool_call[\"function\"][\"arguments\"])\n assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\n \n # Prepare and append the tool response for the next turn\n tool_response = tool_responses.pop(0)\n messages.append(\n {\n \"role\": \"tool\",\n \"tool_call_id\": tool_call[\"id\"],\n \"content\": tool_response[\"response\"],\n }\n )\n else:\n> assert accumulated_content is not None and accumulated_content != \"\", \"Expected content, but none received.\"\nE AssertionError: Expected content, but none received.\nE assert ('' is not None and '' != '')\n\ntests/verifications/openai_api/test_chat_completion.py:526: AssertionError" }, "teardown": { "duration": 0.00036791712045669556, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "lineno": 450, "outcome": "passed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "add_product_tool" }, "setup": { "duration": 0.013031583046540618, "outcome": "passed" }, "call": { "duration": 0.8596610419917852, "outcome": "passed" }, "teardown": { "duration": 0.00042829103767871857, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.015244666952639818, "outcome": "passed" }, "call": { "duration": 1.0227877080906183, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 526, "message": "AssertionError: Expected content, but none received.\nassert ('' is not None and '' != '')" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 526, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'get_then_create_event_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'date': '2025-03-03', ...ents found for 2025-03-03 at 10:00'}\"}, {'response': \"{'response': 'Successfully created new event with id: e_123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n \n # --- Construct Assistant Message for History ---\n assistant_message_dict = {\"role\": \"assistant\"}\n if accumulated_content:\n assistant_message_dict[\"content\"] = accumulated_content\n if accumulated_tool_calls:\n assistant_message_dict[\"tool_calls\"] = accumulated_tool_calls\n \n messages.append(assistant_message_dict)\n \n # --- Assertions ---\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n assert len(accumulated_tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(accumulated_tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n # Use the first accumulated tool call for assertion\n tool_call = accumulated_tool_calls[0]\n assert tool_call[\"function\"][\"name\"] == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call['function']['name']}'\"\n )\n # Parse the accumulated arguments string for comparison\n actual_arguments = json.loads(tool_call[\"function\"][\"arguments\"])\n assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\n \n # Prepare and append the tool response for the next turn\n tool_response = tool_responses.pop(0)\n messages.append(\n {\n \"role\": \"tool\",\n \"tool_call_id\": tool_call[\"id\"],\n \"content\": tool_response[\"response\"],\n }\n )\n else:\n> assert accumulated_content is not None and accumulated_content != \"\", \"Expected content, but none received.\"\nE AssertionError: Expected content, but none received.\nE assert ('' is not None and '' != '')\n\ntests/verifications/openai_api/test_chat_completion.py:526: AssertionError" }, "teardown": { "duration": 0.00024933391250669956, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-3.3-70B-Instruct-Turbo-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-3.3-70B-Instruct-Turbo", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.008626125054433942, "outcome": "passed" }, "call": { "duration": 0.3212552920449525, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 512, "message": "AssertionError: Expected arguments '{'month': 1, 'year': 2025}', got '{'month': '1', 'year': '2025'}'\nassert {'month': '1', 'year': '2025'} == {'month': 1, 'year': 2025}\n \n Differing items:\n {'month': '1'} != {'month': 1}\n {'year': '2025'} != {'year': 2025}\n \n Full diff:\n {...\n \n ...Full output truncated (7 lines hidden), use '-vv' to show" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 512, "message": "AssertionError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-3.3-70B-Instruct-Turbo', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'compare_monthly_expense_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'month': 1, 'year': ... 'Total expenses for January 2025: $1000'}\"}, {'response': \"{'response': 'Total expenses for February 2024: $2000'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n \n # --- Construct Assistant Message for History ---\n assistant_message_dict = {\"role\": \"assistant\"}\n if accumulated_content:\n assistant_message_dict[\"content\"] = accumulated_content\n if accumulated_tool_calls:\n assistant_message_dict[\"tool_calls\"] = accumulated_tool_calls\n \n messages.append(assistant_message_dict)\n \n # --- Assertions ---\n expected = expected_results.pop(0)\n num_tool_calls = expected[\"num_tool_calls\"]\n \n assert len(accumulated_tool_calls or []) == num_tool_calls, (\n f\"Expected {num_tool_calls} tool calls, but got {len(accumulated_tool_calls or [])}\"\n )\n \n if num_tool_calls > 0:\n # Use the first accumulated tool call for assertion\n tool_call = accumulated_tool_calls[0]\n assert tool_call[\"function\"][\"name\"] == expected[\"tool_name\"], (\n f\"Expected tool '{expected['tool_name']}', got '{tool_call['function']['name']}'\"\n )\n # Parse the accumulated arguments string for comparison\n actual_arguments = json.loads(tool_call[\"function\"][\"arguments\"])\n> assert actual_arguments == expected[\"tool_arguments\"], (\n f\"Expected arguments '{expected['tool_arguments']}', got '{actual_arguments}'\"\n )\nE AssertionError: Expected arguments '{'month': 1, 'year': 2025}', got '{'month': '1', 'year': '2025'}'\nE assert {'month': '1', 'year': '2025'} == {'month': 1, 'year': 2025}\nE \nE Differing items:\nE {'month': '1'} != {'month': 1}\nE {'year': '2025'} != {'year': 2025}\nE \nE Full diff:\nE {...\nE \nE ...Full output truncated (7 lines hidden), use '-vv' to show\n\ntests/verifications/openai_api/test_chat_completion.py:512: AssertionError" }, "teardown": { "duration": 0.00020562508143484592, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.007338125025853515, "outcome": "passed" }, "call": { "duration": 0.4175920831039548, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00023462506942451, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.007788832997903228, "outcome": "passed" }, "call": { "duration": 0.45610866602510214, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'weather_tool_then_text', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'location': 'San Francisco...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00021450011990964413, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "add_product_tool" }, "setup": { "duration": 0.006751166889443994, "outcome": "passed" }, "call": { "duration": 0.7053082089405507, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'add_product_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'inStock': True, 'name': 'Widget...}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': 'Successfully added product with id: 123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00021783309057354927, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.008729791967198253, "outcome": "passed" }, "call": { "duration": 0.5665898330044001, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'get_then_create_event_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'date': '2025-03-03', ...ents found for 2025-03-03 at 10:00'}\"}, {'response': \"{'response': 'Successfully created new event with id: e_123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0002288338728249073, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Scout-17B-16E-Instruct-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Scout-17B-16E-Instruct", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.009526000125333667, "outcome": "passed" }, "call": { "duration": 1.1714977910742164, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Scout-17B-16E-Instruct', provider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'compare_monthly_expense_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'month': 1, 'year': ... 'Total expenses for January 2025: $1000'}\"}, {'response': \"{'response': 'Total expenses for February 2024: $2000'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00032483390532433987, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-text_then_weather_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "text_then_weather_tool" }, "setup": { "duration": 0.010107750073075294, "outcome": "passed" }, "call": { "duration": 0.26202141703106463, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'text_then_weather_tool', 'expected': [{'answer': ['sol'], 'num_tool_calls': 0}, {'num_tool_calls': 1, 'to...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00022558285854756832, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-weather_tool_then_text", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "weather_tool_then_text" }, "setup": { "duration": 0.008256082888692617, "outcome": "passed" }, "call": { "duration": 0.3466235001105815, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'weather_tool_then_text', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'location': 'San Francisco...], 'type': 'object'}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': '70 degrees and foggy'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.000535458093509078, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-add_product_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "add_product_tool" }, "setup": { "duration": 0.0180504999589175, "outcome": "passed" }, "call": { "duration": 1.8803812500555068, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'add_product_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'inStock': True, 'name': 'Widget...}}, 'type': 'function'}]}, 'tool_responses': [{'response': \"{'response': 'Successfully added product with id: 123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.00025062495842576027, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-get_then_create_event_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "get_then_create_event_tool" }, "setup": { "duration": 0.00993091706186533, "outcome": "passed" }, "call": { "duration": 0.5258524999953806, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'get_then_create_event_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'date': '2025-03-03', ...ents found for 2025-03-03 at 10:00'}\"}, {'response': \"{'response': 'Successfully created new event with id: e_123'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0002823749091476202, "outcome": "passed" } }, { "nodeid": "tests/verifications/openai_api/test_chat_completion.py::test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "lineno": 450, "outcome": "failed", "keywords": [ "test_chat_streaming_multi_turn_tool_calling[meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool]", "parametrize", "pytestmark", "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-compare_monthly_expense_tool", "test_chat_completion.py", "openai_api", "verifications", "tests", "llama-stack", "" ], "metadata": { "model": "meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8", "case_id": "compare_monthly_expense_tool" }, "setup": { "duration": 0.047535917023196816, "outcome": "passed" }, "call": { "duration": 0.4426498331595212, "outcome": "failed", "crash": { "path": "/Users/erichuang/projects/llama-stack/tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError: list index out of range" }, "traceback": [ { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 485, "message": "" }, { "path": "tests/verifications/openai_api/test_chat_completion.py", "lineno": 587, "message": "IndexError" } ], "longrepr": "request = >\nopenai_client = \nmodel = 'meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8'\nprovider = 'together'\nverification_config = {'providers': {'cerebras': {'api_key_var': 'CEREBRAS_API_KEY', 'base_url': 'https://api.cerebras.ai/v1', 'model_displa...-versatile', 'meta-llama/llama-4-scout-17b-16e-instruct', 'meta-llama/llama-4-maverick-17b-128e-instruct'], ...}, ...}}\ncase = {'case_id': 'compare_monthly_expense_tool', 'expected': [{'num_tool_calls': 1, 'tool_arguments': {'month': 1, 'year': ... 'Total expenses for January 2025: $1000'}\"}, {'response': \"{'response': 'Total expenses for February 2024: $2000'}\"}]}\n\n @pytest.mark.parametrize(\n \"case\",\n chat_completion_test_cases.get(\"test_chat_multi_turn_tool_calling\", {}).get(\"test_params\", {}).get(\"case\", []),\n ids=case_id_generator,\n )\n def test_chat_streaming_multi_turn_tool_calling(request, openai_client, model, provider, verification_config, case):\n \"\"\" \"\"\"\n test_name_base = get_base_test_name(request)\n if should_skip_test(verification_config, provider, model, test_name_base):\n pytest.skip(f\"Skipping {test_name_base} for model {model} on provider {provider} based on config.\")\n \n messages = []\n tools = case[\"input\"][\"tools\"]\n expected_results = copy.deepcopy(case[\"expected\"])\n tool_responses = copy.deepcopy(case.get(\"tool_responses\", []))\n input_messages_turns = copy.deepcopy(case[\"input\"][\"messages\"])\n \n while len(input_messages_turns) > 0 or (len(messages) > 0 and messages[-1][\"role\"] == \"tool\"):\n if len(messages) == 0 or messages[-1][\"role\"] != \"tool\":\n new_messages = input_messages_turns.pop(0)\n if isinstance(new_messages, list):\n messages.extend(new_messages)\n else:\n messages.append(new_messages)\n \n # --- API Call (Streaming) ---\n stream = openai_client.chat.completions.create(\n model=model,\n messages=messages,\n tools=tools,\n stream=True,\n )\n \n # --- Process Stream ---\n> accumulated_content, accumulated_tool_calls = _accumulate_streaming_tool_calls(stream)\n\ntests/verifications/openai_api/test_chat_completion.py:485: \n_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ \n\nstream = \n\n def _accumulate_streaming_tool_calls(stream):\n \"\"\"Accumulates tool calls and content from a streaming ChatCompletion response.\"\"\"\n tool_calls_buffer = {}\n current_id = None\n full_content = \"\" # Initialize content accumulator\n # Process streaming chunks\n for chunk in stream:\n> choice = chunk.choices[0]\nE IndexError: list index out of range\n\ntests/verifications/openai_api/test_chat_completion.py:587: IndexError" }, "teardown": { "duration": 0.0010368749499320984, "outcome": "passed" } } ], "run_timestamp": 1744679294 }