Development¶

Pull requests and issue reports are welcome. For major changes, please open an issue first to discuss what you would like to change.

Setup¶

The project uses uv for everything - use uv run rather than calling python directly.

git clone https://github.com/hbmartin/ambientweather2sqlite.git
cd ambientweather2sqlite
uv sync --dev

Checks¶

Run the formatter and linters after any change:

uv run ruff format ambientweather2sqlite tests
uv run ruff check ambientweather2sqlite tests --fix
uv run pyrefly check ambientweather2sqlite
uv run ty check ambientweather2sqlite

Run the full suite before opening a pull request:

uv run --locked pyroma .                          # package metadata; strict 10/10
uv run deptry ambientweather2sqlite                # dependency hygiene
uv run lizard -Eduplicate ambientweather2sqlite    # complexity and duplication
uv run pytest tests/ --cov=ambientweather2sqlite   # tests; coverage must stay >= 92%

CI runs all of the above against Python 3.13 and 3.14 on every push and pull request.

Package metadata is checked with pyroma621, pinned to 6.2.0 so changes to its checks can be reviewed before upgrading. It provides the pyroma command and Python module. The [tool.pyroma] configuration requires a strict 10/10 rating, including advisory checks, without skipping any checks. uv run --locked pyroma . applies this policy locally and in CI. When checking a built source distribution, pass --strict --min 10 explicitly; archive mode does not load the project's configuration.

Semgrep¶

Project-specific Semgrep rules protect the backup, Home Assistant discovery, and health-metric invariants that generic Python linters cannot express. Run the production scan and its rule fixtures with:

uvx --from semgrep==1.169.0 semgrep scan --metrics=off --error --config semgrep/ambientweather2sqlite.yml ambientweather2sqlite
uvx --from semgrep==1.169.0 semgrep --metrics=off --test semgrep/

Conventions¶

The source uses a flat ambientweather2sqlite/ package layout, not a src/ directory.
Python 3.13+, and the package intentionally has no runtime dependencies - do not add any.
ambientweather2sqlite/mureq.py is vendored. Preserve its public API and compare updates against upstream rather than replacing it wholesale.
Tests are pytest-compatible unittest.TestCase classes. Use Hypothesis for invariant and round-trip coverage where property-based tests fit.
Type hints are first-class; prefer explicitness and small functions.
Use modern Python features - assignment expressions, structural pattern matching - where they improve clarity.
Use a parenthesized tuple of exception classes in every except clause, including for a single exception.
Any user-facing change (e.g. a new CLI flag) must be documented in the README.md.

Documentation¶

The documentation site is built with Zensical from the Markdown files in docs/, configured by zensical.toml.

uv run zensical serve   # live preview on http://localhost:8000
uv run zensical build   # render to site/

Pushes to main that touch the docs are deployed to GitHub Pages automatically by .github/workflows/docs.yml.

Core Architecture¶

Architecture diagram: CLI and configuration feed the daemon, which fetches from the station, parses observations, and writes to SQLite while optionally publishing to MQTT and serving the HTTP API.

Control Flow¶

Control flow diagram: the collection loop fetches live data, parses sensors, evolves the schema, inserts the observation, and sleeps until the next interval.