Skip to content

cli version update

cli version update #22

Workflow file for this run

# Until this workflow existed, nothing in the repo ran a test: release.yml only publishes.
# Two things had rotted quietly as a result - the docs never mentioned the LLM Judge Scorer
# surface 0.6.36 shipped, and four tests had been asserting on a performance_summary payload
# that 9d45dd1 replaced with real child spans, some of them for over two weeks.
#
# Two jobs, because they answer different questions at different costs:
#
# tests - the base install on both supported Pythons. Fast feedback on the core, and the
# configuration most contributors actually have. Integration tests for optional
# extras skip themselves here via pytest.importorskip.
# extras - every framework integration installed, so those tests actually run instead of
# skipping. Slower (chromadb, onnxruntime and friends come along), so it runs on
# one Python rather than the matrix. autogen-ext is installed explicitly: its
# ReplayChatCompletionClient is a test double, so it belongs in CI rather than in
# the package's "autogen" extra.
name: tests
on:
push:
branches: [main]
pull_request:
workflow_dispatch:
jobs:
tests:
runs-on: ubuntu-latest
timeout-minutes: 15
strategy:
fail-fast: false
matrix:
# The version release.yml builds the published wheel with, and a current one.
python: ["3.9", "3.12"]
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: ${{ matrix.python }}
- run: pip install -e ".[langchain]" pytest
# tests/test_integration.py talks to the hosted API and skips itself without a key;
# no secret is passed here on purpose, so pull requests from forks behave the same.
- run: python -m pytest tests/ -q
extras:
runs-on: ubuntu-latest
timeout-minutes: 30
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: "3.12"
- run: pip install -e ".[all]" autogen-ext pytest
- run: python -m pytest tests/ -q