Skip to content

Tools for working with MDTF Diagnostics test datasets

License

Notifications You must be signed in to change notification settings

wrongkindofdoctor/mdtf_test_data

 
 

Repository files navigation

Total alerts Language grade: Python

MDTF Diagnostics Test Data Tools

Package to work with test input datasets for MDTF Diagnostics

Overview

This package is designed to coarsen NetCDF files for generating test datasets and produce synthetic datasets on-the-fly that can be used for testing the MDTF Diagnostics package.

Requirements

cftime

  • envyaml
  • numpy
  • pandas
  • pytest
  • xarray
  • xESMF

Getting the code

The mdtf_test_data package is available on PyPI.

Installation

mkdir mdtf_test_data && cd mdtf_test_data
pip install mdtf-test-data

Usage

This package can be used to generate fully-synthetic datasets based on NCAR's CESM2 and GFDL's CM4 model output that can be used to test the MDTF-Diagnostics package.

usage: mdtf_synthetic.py [-h] [-c CONVENTION] [--startyear year] [--nyears years]
[--dlat latitude resolution in degrees] [--dlon longitude resolution in degrees] [--unittest]

Required arguments:
  -c, --convention      Data convention [NCAR, CESM, GFDL, CMIP]

Optional arguments:
  -h, --help            show this help message and exit
  --startyear           start year of data [default is 1975]
  --nyears              number of years of data to generate [default is 10]
  --dlat                latitude resolution in degrees [default is 20]
  --dlon                longitude resolution in degrees [default is 20]
  --unittest............flag to run unit tests in mdtf_test_data/tests

To generate NCAR CESM output in a directory called NCAR.Synthetic:

mdtf_synthetic.py -c [ NCAR | CESM ] --nyears 7 

Note that "NCAR" and "CESM" are both valid input values that generate output files appended with "NCAR_Synthetic".

To generate GFDL CM4 output in a directory called GFDL.Synthetic:

mdtf_synthetic.py -c GFDL --nyears 10

To generate CMIP output in a directory called CMIP.Synthetic:

mdtf_synthetic.py -c CMIP --nyears 10

To coarsen an existing NetCDF file:

git clone https://github.com/jkrasting/mdtf_test_data.git
cd mdtf_test_data
usage: mdtf_synthetic/util/mdtf-coarsen.py [-h] [-r REGRID_METHOD] [-o OUTFILE] [-O] infile

Coarsen a NetCDF file.

Required arguments:
  infile                Path to input NetCDF file

Optional arguments:
  -h, --help            show this help message and exit
  -r REGRID_METHOD      xESMF regridding method
  -o OUTFILE, --outfile OUTFILE
                        Filename of output NetCDF file
  -O                    Overwrite existing file

Notes:

  • The coarsening tool only supports standard grids with dimensions name lat and lon for now
  • Any xESMF regrid method may be passed with the -r option

Getting Help

Submit a GitHub Issue

About

Tools for working with MDTF Diagnostics test datasets

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%