Skip to content

Latest commit

 

History

History
38 lines (29 loc) · 2.22 KB

contributing.md

File metadata and controls

38 lines (29 loc) · 2.22 KB

Contributing

Environment Setup

For code and dataset contribution, please setup a virtual environment and install the requirements using the following commands:

python -m venv venv
source venv/bin/activate
pip install -r requirements-dev.txt

Tests

To run style checkers and tests, run the following command:

python -m black --check --diff .
python -m mypy cropharvest process_labels test benchmarks
python -m mypy benchmarks/dl
python -m pytest

Dataset Contribution

If you would like to contribute a dataset, thank you! Please complete the following steps to do so:

1. Add the dataset to process_labels

The first step is to open a pull request which adds the data to the process_labels folder in this repository. There are 3 places which will need to be changed:

  • Add the raw data to the raw data folder
  • Add a function which takes the raw data and returns a geojson of the appropriate format. Appropriate format means it has all of the required columns, and all the nullable columns possible. These columns are described in columns.py. The expected type of the columns is described (and tested) in test_datasets.py.
  • Update the DATASETS dict to include the newest dataset.

2. Recreate the data, and upload to Zenodo

You will need to recreate the features, and upload the new labels.geojson to Zenodo. Please contact Gabriel Tseng ([email protected]) or Ivan Zvonkov ([email protected]) to do this.

3. Update the Zenodo link in CropHarvest

Add the new Zenodo identifier to the cropharvest config. Because the python package is coupled to the data, this will also require a new release. Once again, please contact Gabriel or Ivan to do this.

4. Update the Readme / datasets to reflect the new dataset

We keep track of certain statistics in the Readme (e.g. dataset size, number of multiclass labels, ...). The final step is to update these numbers so they remain accurate as new datasets are added. In addition, [datasets.md][datasets.md] contains a list of datasets included in CropHarvest - this should also be updated.