Skip to content

0.9.6.24: issued_on and scheduled_for were filled with prose because … #245

0.9.6.24: issued_on and scheduled_for were filled with prose because …

0.9.6.24: issued_on and scheduled_for were filled with prose because … #245

Workflow file for this run

name: CI
on:
push:
branches: [main]
pull_request:
branches: [main]
jobs:
test:
name: Python ${{ matrix.python-version }}
runs-on: ubuntu-latest
strategy:
fail-fast: false
matrix:
python-version: ["3.10", "3.11", "3.12", "3.13"]
steps:
- uses: actions/checkout@v4
- name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v5
with:
python-version: ${{ matrix.python-version }}
cache: pip
- name: Install dependencies
run: |
pip install -e ".[dev]"
pip install pytest-cov
- name: Run tests with coverage
run: |
pytest tests/ -q --tb=short --cov=misata --cov-report=xml --cov-report=term-missing
- name: Run scale test
run: |
pytest tests/test_integration_scale.py -v -m slow --tb=short
# The Gauntlet: 11 tables, 126 independent SQL assertions run by DuckDB
# against the emitted frames. Known-red assertions (named roadmap items)
# are expected failures; any UNEXPECTED failure is a correctness
# regression and fails the build. A known-red that starts passing also
# fails, so the roadmap list can never go stale.
- name: Run the Gauntlet (relational-integrity release gate)
run: |
python -m benchmarks.gauntlet
# The Warren: a second conformance suite with a deliberately different
# shape (multi-tenant, self-referential hierarchy, event-sourced). It
# exists because the Gauntlet hit 100% and a suite at 100% has stopped
# being able to find anything; the Warren found 22 failures on its first
# run, four of which were real engine gaps.
- name: Run the Warren (multi-tenant + hierarchy release gate)
run: |
python -m benchmarks.warren
# The Ledger: bitemporal (two independent time axes) and graph-shaped
# (a DAG edge table plus its transitive closure). Third shape, same
# contract. Together the three suites cover 284 assertions.
- name: Run the Ledger (bitemporal + graph release gate)
run: |
python -m benchmarks.ledger
- name: Audit dependencies for known vulnerabilities
if: matrix.python-version == '3.11'
run: |
pip install pip-audit
pip-audit --skip-editable || true # informational for now; tighten to blocking after a clean baseline
- name: Upload coverage to Codecov
if: matrix.python-version == '3.11'
uses: codecov/codecov-action@v4
with:
token: ${{ secrets.CODECOV_TOKEN }}
files: coverage.xml
fail_ci_if_error: false
stranger:
name: The docs, run as a stranger would
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: "3.12"
# Deliberately NOT `pip install -e .`. An editable install from inside
# the repo is what hid the MCP server's missing extra in 0.9.6.1: the
# dev environment already had the dependency, so the packaging bug was
# invisible until someone installed the published wheel. This job builds
# the wheel and installs it into a clean virtualenv, the way a reader
# following the README does.
- name: Install build tooling only
run: python -m pip install --upgrade pip build
- name: Run every documented example
run: python tools/stranger.py --clean
studio-stranger:
name: A declaration is honoured or refused, never substituted
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: "3.12"
- run: python -m pip install --upgrade pip
- run: pip install -e .
# In-process, so this gates every push without needing the API up. Run it
# with `--api https://api.misata.studio` after a deploy to check the
# served build agrees with this one.
- name: Probe the schema surface the way a caller would
run: python tools/studio_stranger.py --local