菜单
Run the Retail Shopping Assistant test suites under tests/, including offline pytest unit tests in tests/unit and live integration scripts in tests/integration for conversation, quality, timing, and plots.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
基于 SOC 职业分类
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/NVIDIA-AI-Blueprints/retail-shopping-assistant --skill retail-test-runner命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | retail-test-runner |
| description | Run the Retail Shopping Assistant test suites under tests/, including offline pytest unit tests in tests/unit and live integration scripts in tests/integration for conversation, quality, timing, and plots. |
| metadata | {"short-description":"Run retail unit and integration tests"} |
Use this skill when the user asks to run, validate, debug, or explain the Retail Shopping Assistant tests under this repository's tests/ directory.
Use the deterministic runner from the repo root:
python skills/retail-test-runner/scripts/run_retail_tests.py all --test-path shopping
Common invocations:
python skills/retail-test-runner/scripts/run_retail_tests.py unit
python skills/retail-test-runner/scripts/run_retail_tests.py integration --test-path shopping
python skills/retail-test-runner/scripts/run_retail_tests.py integration --test-path rails --skip-quality
The runner:
.env before checking integration credentials..local-run/dev-venv/bin/python when it exists, then .venv-tests/bin/python, then the current Python.tests/ so tests/pytest.ini is applied.tests/integration/ so their relative conversations/<TEST_PATH> paths resolve correctly.http://localhost:8009/query/timing by default.TEST_PATH for integration runs.The unit suite lives under tests/unit and is offline. It should not require Docker, Milvus, NIMs, guardrails services, or network calls.
If Python dev dependencies are missing, install them into the repo-local dev venv:
python skills/retail-local-runner/scripts/local_runner.py install-dev
Targeted unit examples:
cd tests
../.local-run/dev-venv/bin/python -m pytest -q unit/chain_server
../.local-run/dev-venv/bin/python -m pytest -q unit/chain_server/test_cart.py
../.local-run/dev-venv/bin/python -m pytest -q -k test_add_to_cart
The integration suite lives under tests/integration and drives live HTTP endpoints. Before running it:
http://localhost:8009.tests/integration/conversations/, usually shopping or rails.NVIDIA_API_KEY in the repo-root .env or export it in the launching shell, unless using --skip-quality; response_quality.py requires it for LLM-as-judge scoring.The runner preflights the selected conversation directory and service URL. Use --no-preflight only when intentionally testing a nonstandard setup.
Integration outputs are written under:
tests/integration/conversations/<TEST_PATH>/results/tests/integration/conversations/<TEST_PATH>/judge/Do not delete or overwrite these outputs unless the user asks for a clean run.
After editing this skill, validate both skill folders:
uv run --with pyyaml python "$CODEX_HOME/skills/.system/skill-creator/scripts/quick_validate.py" skills/retail-test-runner
uv run --with pyyaml python "$CODEX_HOME/skills/.system/skill-creator/scripts/quick_validate.py" .agents/skills/retail-test-runner
Start, stop, configure, inspect, and troubleshoot the Retail Shopping Assistant locally with backend Python and React code outside containers, local Milvus infra containers, and remote NIM endpoints from docker-compose-nim-local.yaml.
Start, stop, configure, inspect, and troubleshoot the Retail Shopping Assistant locally with backend Python and React code outside containers, local Milvus infra containers, and remote NIM endpoints from docker-compose-nim-local.yaml.
Run the Retail Shopping Assistant test suites under tests/, including offline pytest unit tests in tests/unit and live integration scripts in tests/integration for conversation, quality, timing, and plots.