mlcommons / mlcommons/endpoints
test: SGLang adapter integration tests are permanently skipped
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 21
- Forks
- 28
- Avg merge
- 3d 17h
- Merged PRs (30d)
- 13
Description
Problem
Two SGLang adapter integration tests in tests/integration/ are marked with @pytest.mark.skip with no resolution plan or linked issue. The SGLang adapter is production code under src/inference_endpoint/sglang/, but it has no running integration test coverage.
This means regressions in the SGLang adapter will not be caught by CI.
Expected Behavior
The SGLang integration tests should be un-skipped and run against the echo/oracle test server fixtures (or a mock SGLang server if the protocol differs from OpenAI).
If the tests cannot be un-skipped due to a known blocker, the @pytest.mark.skip should include a reason= referencing this issue.
Files to Update
- Skipped test files under
tests/integration/for SGLang src/inference_endpoint/sglang/— verify adapter implementation is complete
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the two SGLang tests marked with @pytest.mark.skip under tests/integration/ and read the adapter in src/inference_endpoint/sglang/. Run the tests with the echo/oracle fixtures to determine whether the existing protocol works or a mock server is needed. Done means the tests run in CI, or each unavoidable skip has a reason= that references issue 220.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend, testing
- Issue type
- Feature
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 55/100