gval
gval copied to clipboard
A high-level Python framework to evaluate the skill of geospatial datasets by comparing candidates to benchmark maps producing agreement maps and metrics.
There needs to be functionality to convert a STAC query in to a catalog. Catalogs will be represented by a DataFrame.
There is repeated code for homogenization and and attribute tracking for all full compare methods in gval_xarray.py. There should be a decorator for these methods to promote code reuse.
It should be explored if a probabilistic class should be implemented. If so it should be created in the statistics module.
Where appropriate, we should consolidate configuration files to `pyproject.toml` unless some other use is required for the file. Some examples include: - `setup.cfg` - `.coveragerc` - `config.py`: this file may...
Nox sessions have low test coverage (~85%) when ran locally and even lower on GH Actions (~67%). Update `.coveragerc` min fail value to 95 when addressed.
Add nox session or ability to publish package to PyPi via GH actions as an optional task.
Add a version bumper such as those listed [here](https://github.com/c4urself/bump2version/blob/master/RELATED.md) to automate version changing. Add a nox session to do this with the new version possibly as a session parameter.
Add a static type checker such as pyright, mypy, etc as a nox session within the current `PLACEHOLDER` within the `noxfile.py`. This will require adding the type checker as a...
Probabilistic compare should have an example notebook preferably with real data. NOAA produces a variety of ensemble forecasts.