phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-03 18:00:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
llama-stack-mirror/tests/client-sdk
History
Xi Yan 66d7e15c93
perf: ensure ToolCall in ChatCompletionResponse is subset of ChatCompletionRequest.tools (#1041) 
# What does this PR do?

**Problem**
- Using script:
https://gist.github.com/thoraxe/6163b2145ce7b1c24c6026b64cf90085

- This hits an issue on server with `code_interpreter` not found, as we
do not pass "builtin::code_interpreter" in AgentConfig's `toolgroups`.

This is a general issue where model always tries to output
`code_interpreter` in `ToolCall` even when we do not have
`code_interpreter` available for execution.

**Reproduce Deeper Problem in chat-completion**
- Use script:
https://gist.github.com/yanxi0830/163a9ad7b5db10556043fbfc7ecd7603

1. We currently always populate `code_interpreter` in `ToolCall` in
ChatCompletionResponse if the model's response begins with
`<|python_tag|>`. See
c5f5958498/models/llama3/api/chat_format.py (L200-L213)

<img width="913" alt="image"
src="https://github.com/user-attachments/assets/328d313d-0a0b-495c-8715-61cca9ccc4a6"
/>

2. This happens even if we do not pass the `code_interpreter` as a
`tools` in ChatCompletionRequest.

**This PR**

Explicitly make sure that the tools returned in
`ChatCompletionResponse.tool_calls` is always a tool requested by
`ChatCompletionRequest.tools`.

[//]: # (If resolving an issue, uncomment and update the line below)
[//]: # (Closes #[issue-number])

## Test Plan

**Before**
<img width="913" alt="image"
src="https://github.com/user-attachments/assets/328d313d-0a0b-495c-8715-61cca9ccc4a6"
/>
<img width="997" alt="image"
src="https://github.com/user-attachments/assets/d3e82b62-b142-4939-954c-62843bec7110"
/>


**After**
<img width="856" alt="image"
src="https://github.com/user-attachments/assets/2c70ce55-c8d0-45ea-b10f-f70adc50d3d9"
/>
<img width="1000" alt="image"
src="https://github.com/user-attachments/assets/b5e81826-c35b-4052-bf81-7afff93ce2ef"
/>



**Unit Test**
```
LLAMA_STACK_BASE_URL=http://localhost:8321 pytest -v tests/client-sdk/inference/test_text_inference.py::test_text_chat_completion_tool_calling_tools_not_in_request --inference-model "meta-llama/Llama-3.3-70B-Instruct"
```

```
LLAMA_STACK_BASE_URL=http://localhost:8321 pytest -v tests/client-sdk/agents/
```
<img width="1002" alt="image"
src="https://github.com/user-attachments/assets/04808517-eded-4122-97f5-7e5142de9779"
/>



**Streaming**
- Chat Completion
<img width="902" alt="image"
src="https://github.com/user-attachments/assets/f477bc86-bd38-4729-b49e-a0a6ed3f835a"
/>

- Agent
<img width="916" alt="image"
src="https://github.com/user-attachments/assets/f4cc3417-23cd-46b1-953d-3a2271e79bbb"
/>


[//]: # (## Documentation)
[//]: # (- [ ] Added a Changelog entry if the change is significant)
2025-02-11 18:31:35 -08:00
..
agents fix: agent config validation (#1053) 2025-02-11 14:48:42 -08:00
inference perf: ensure ToolCall in ChatCompletionResponse is subset of ChatCompletionRequest.tools (#1041) 2025-02-11 18:31:35 -08:00
safety Fix precommit check after moving to ruff (#927) 2025-02-02 06:46:45 -08:00
tool_runtime Fix precommit check after moving to ruff (#927) 2025-02-02 06:46:45 -08:00
vector_io Fix precommit check after moving to ruff (#927) 2025-02-02 06:46:45 -08:00
__init__.py [tests] add client-sdk pytests & delete client.py (#638) 2024-12-16 12:04:56 -08:00
conftest.py Update client-sdk test config option handling 2025-01-31 15:30:07 -08:00
metadata.py Report generation minor fixes (#884) 2025-01-28 04:58:12 -08:00
README.md test: Split inference tests to text and vision (#1008) 2025-02-07 09:35:49 -08:00
report.py Fix precommit check after moving to ruff (#927) 2025-02-02 06:46:45 -08:00

README.md

Llama Stack Integration Tests

You can run llama stack integration tests on either a Llama Stack Library or a Llama Stack endpoint.

To test on a Llama Stack library with certain configuration, run

LLAMA_STACK_CONFIG=./llama_stack/templates/cerebras/run.yaml
pytest -s -v tests/client-sdk/inference/

or just the template name

LLAMA_STACK_CONFIG=together
pytest -s -v tests/client-sdk/inference/

To test on a Llama Stack endpoint, run

LLAMA_STACK_BASE_URL=http//localhost:8089
pytest -s -v tests/client-sdk/inference

Report Generation

To generate a report, run with --report option

LLAMA_STACK_CONFIG=together pytest -s -v report.md tests/client-sdk/ --report

Common options

Depending on the API, there are custom options enabled

  • For tests in inference/ and agents/, we support --inference-model(to be used in text inference tests) and--vision-inference-model` (only used in image inference tests) overrides
  • For tests in vector_io/, we support --embedding-model override
  • For tests in safety/, we support --safety-shield override
  • The param can be --report or --report <path> If path is not provided, we do a best effort to infer based on the config / template name. For url endpoints, path is required.