Namyalg / gcbm_preprocessing

Implementation of pseudocode for GCBM Pre-processing

This script follows the pseudocode from Andrew's design to update GCBM configs: https://hackmd.io/@aornugent/B1RluG69c.

Execution

End-Point

• Endpoints supported: /gcbm/upload
• Categories supported: disturbances, classifiers, miscellaneous

# This uploads a classifier, disturbance and a miscellaneous file
curl -F disturbances="@disturbances/disturbances_2011_moja.tiff" -F classifiers="@classifiers/Classifier1_moja.tiff" http://localhost:8080/gcbm/upload

# Try skipping disturbance file, and it will still work :)

Expected Output:

{
  "error": "Missing files for categories: ['config_files', 'input'], they are required for the simulation to run"
}

The outputs for classifier file and disturbance file will be stored in "templates/config/" folder.

GCBM Pre-Processing

Run: python3 gcbm.py and a folder will be generated with the output config files (JSON) for all the categories.

Script:

pip install -r requirements.txt
python3 gcbm.py
cat templates/config/disturbances_2011_moja.json
cat templates/config/Classifier1_moja.json

Please see the main function in the gcbm.py file for more info. A payload is manually added, which checks if it has year and then the has_year key is set to True

Validation

In order to make sure that the generated configs are correct, I've written a script compare_jsons.py. Pass the original folder first and the new folder second if you want (else the default paths are used).

>>> python3 compare_jsons.py compare templates/config/

Note: using compare for original JSONs and templates/config for new JSONs folders.
Failures: 0/11
Success: 11/11