mirror of
https://github.com/meta-llama/llama-stack.git
synced 2025-12-03 01:48:05 +00:00
40b11efac4
Integration tests can now validate the TypeScript SDK alongside Python tests when running against server-mode stacks. Currently, this only adds a _small_ number of tests. We should extend only if truly needed -- this smoke check may be sufficient. When `RUN_CLIENT_TS_TESTS=1` is set, the test script runs TypeScript tests after Python tests pass. Tests are mapped via `tests/integration/client-typescript/suites.json` which defines which TypeScript test files correspond to each Python suite/setup combination. The fact that we need exact "test_id"s (which are actually generated by pytest) to be hardcoded inside the Typescript tests (so we hit the recorded paths) is a big smell and it might become grating, but maybe the benefit is worth it if we keep this test suite _small_ and targeted. ## Test Plan Run with TypeScript tests enabled: ```bash OPENAI_API_KEY=dummy RUN_CLIENT_TS_TESTS=1 \ scripts/integration-tests.sh --stack-config server:ci-tests --suite responses --setup gpt ```
39 lines
633 B
Text
39 lines
633 B
Text
.env
|
|
__pycache__
|
|
dist
|
|
*.egg-info
|
|
dev_requirements.txt
|
|
build
|
|
.DS_Store
|
|
llama_stack/configs/*
|
|
.cursor/
|
|
xcuserdata/
|
|
*.hmap
|
|
.DS_Store
|
|
.build/
|
|
Package.resolved
|
|
*.pte
|
|
*.ipynb_checkpoints*
|
|
.idea
|
|
.venv/
|
|
.vscode
|
|
_build
|
|
# Sample tool-calling datasets generated by NVIDIA notebooks
|
|
docs/notebooks/nvidia/tool_calling/sample_data/
|
|
pyrightconfig.json
|
|
venv/
|
|
pytest-report.xml
|
|
.coverage
|
|
.python-version
|
|
AGENTS.md
|
|
server.log
|
|
CLAUDE.md
|
|
.claude/
|
|
docs/.docusaurus/
|
|
docs/node_modules/
|
|
docs/static/imported-files/
|
|
docs/docs/api-deprecated/
|
|
docs/docs/api-experimental/
|
|
docs/docs/api/
|
|
tests/integration/client-typescript/node_modules/
|
|
.ts-client-checkout/
|