Adding New Features

If you have created new feature collection and wish for the individual features to be included in Geometric Data, you should:

clone the full GitHub repository

split the feature collection into its individual features

update the manifest of all features and tags

commit the changes and make a pull request

import os
from geometric_features import GeometricFeatures, read_feature_collection
from geometric_features.utils import write_feature_names_and_tags

# A new feature colleciton has been constructed and tagged using the
# naming convetions of geometric_features
fc = read_feature_collection('my_new_features.geojson')

# make a geometric features object that points to geometry in the local
# cache in ./geometric_data
gf = GeometricFeatures(cacheLocation='./geometric_data')

# split the feature collection into individual features within
# ./geometric_data
gf.split(fc)

# write a file features_and_tags.json with features and tags from the cache
write_feature_names_and_tags(gf.cacheLocation)

# move features_and_tags.json into geometric_features to replace the old
# manifest
os.rename('features_and_tags.json',
          'geometric_features/features_and_tags.json')

After this, you can git add and git commit the changes, and make a pull request to have them added to the repository.

It is not recommended that you modified features directly in geometric_data, but if you have already done so, you can update the manifest of all features and tags based on your changes first, then merge your features into a feature collection and then split it back out into individual features to ensure consistent formatting.

import os
from geometric_features import GeometricFeatures
from geometric_features.utils import write_feature_names_and_tags


# Write a file features_and_tags.json with features and tags from the cache.
# This updates the file names, feature names and tags that geometric_features
# knows about.
write_feature_names_and_tags('./geometric_data')

# move features_and_tags.json into geometric_features to replace the old
# manifest
os.rename('features_and_tags.json',
          'geometric_features/features_and_tags.json')

# Make a geometric features object that gets data from a local cache in
# ./geometric_data.  (The remote branch won't matter.)
gf = GeometricFeatures(cacheLocation='./geometric_data')

# Make a feature colleciton with the standrd transport sections
fc = gf.read(componentName='ocean', objectType='transect',
             tags=['standard_transport_sections'])

# split the feature collection back into individual features within
# ./geometric_data to clean things up
gf.split(fc)

Co-developing geometric_features with Other Packages

If you are developing geometric_features alongside another package (such as MPAS-Analysis), you can set up a single development environment for both projects. This is useful for testing changes in both packages together.

Create a development environment with dependencies from both packages:
```
conda create -n gf-dev --file /path/to/geometric_features/dev-spec.txt --file /path/to/MPAS-Analysis/dev-spec.txt
```
Replace the paths above with the actual locations of the two repositories.

Install both packages in editable mode:

conda activate gf-dev
pip install --no-deps --no-build-isolation -e /path/to/geometric_features
pip install --no-deps --no-build-isolation -e /path/to/MPAS-Analysis

Set the GEOMETRIC_DATA_DIR environment variable:

You must set the $GEOMETRIC_DATA_DIR environment variable to the absolute path of the geometric_data subdirectory in your geometric_features repository. This is required so that other packages (like MPAS-Analysis) can find the geometric data during development.
```
export GEOMETRIC_DATA_DIR=/absolute/path/to/geometric_features/geometric_data
```
You will need to run this command every time you activate your development environment. If you are running jobs (e.g., with MPAS-Analysis), make sure to include this export in your job scripts as well.