Subsetting NetCDF4 with xarray and clisops

subset_clisops_xarray.py

"""
Install Anaconda3 https://www.anaconda.com/products/individual (Beginner-level, larger install)
or Miniconda3: https://docs.conda.io/en/latest/miniconda.html (Experienced, minimal install)

In a terminal / command prompt with `conda` installed:

  $ conda create --name data_processing -c conda-forge python=3.7 xarray clisops
  $ conda activate data_processing
  <Modify paths to folders (source files, target_folder)>
  $ python subset_clisops_xarray.py
"""

from pathlib import Path

import xarray as xr
from clisops.core import subset

vector = Path("/Path/to/a/shapefile.shp/or/geojson.json")  # CRS should be WGS84 or NAD83 (decimal degrees).
source_file_folder = Path("/Path/to/the/top-level/folder/of/climate/data/")  # Will subset all files within folder.
source_files = sorted([file for file in source_file_folder.rglob("*.nc")])

target_folder = Path(
    "/Path/to/another/folder/for/newly/subsetted/files/"
)  # Folder must exist and have write access!


def subsetter_function(files, shape, output_folder):
    for file in files:
        # Subset with shapefile/geojson.
        ds = subset.subset_shape(
            xr.open_dataset(file, decode_times=False),  # Open NetCDF4 with xarray.
            shape=shape,  # Supply a geometry to clip with.
        )
        
        # Name file based on location and filename of older file.
        output_name = f"{file.stem}_subset.nc"
        parent_folder = file.parts[-3: -1]  # /scenario/variable/
        output_folder = output_folder.joinpath(parent_folder)  # /output_folder/scenario/variable/
        output_folder.mkdir(exist_ok=True, parents=True)
        
        # Write out object as a new NetCDF4
        ds.to_netcdf(output_folder.joinpath(output_name))


if __name__ == "__main__":
    subsetter_function(source_files, vector, target_folder)