THIS PROJECT IS ARCHIVED, PLEASE USE py-cordex INSTEAD!

WARNING: This package has no release yet, the API could and will change.

cordex

The cordex python package is supposed to becoma a community effort to collect common data and tools.

Description

Python package for common tools and data in the Cordex community.

Installation

You can install the package directly from github using pip:

pip install git+https://github.com/euro-cordex/cordex

If you want to contribute, I recommend cloning the repository and installing the package in development mode, e.g.

git clone https://github.com/euro-cordex/cordex
cd cordex
pip install -e .

This will install the package but you can still edit it and you don't need the package in your PYTHONPATH

Highlights

Use the domain module to create Cordex domains safe and easy:

from cordex import domain as dm

# available tables
print(dm.tables())
# available domains names
print(dm.names())
# available domains names
print(dm.names('cordex-core'))
# print cordex core table
print(dm.table('cordex-core'))

# show domains with some dummy data (uses cdo topo)
for short_name, domain in dm.domains().items():
    print(domain)
    domain.to_netcdf(short_name+'.nc', dummy='topo')

Using the ESGF module for browsing the file system and creating pandas dataframes of file selections: (DKRZ example).

Note that DKRZ also provides an intake-esm table (https://intake-esm.readthedocs.io/en/latest/) for browsing and searchin MIP data at DKRZ. I highly recommend to have a look at it and create tables for your own data pool. For creating a table, you could use our cordex conventions module!

import datetime as dt
from cordex.ESGF import get_selection, conventions

project_id = 'CMIP5'
root       = '/pool/data/CMIP5/cmip5'

filter   = {'institute'       : 'MPI-M',
            'product'         : 'output1',
            'variable'        : 'tas',
            'model'           : 'MPI-ESM-LR',
            'modeling_realm'  : 'atmos',
            'experiment'      : 'historical',
            'ensemble_member' : 'r1i1p1'}

# get a selection of files (using pandas dataframes)
selection = get_selection(project_id, root=root, filter=filter)
print(selection)

# create a finer selection and convert dates to datetime objects
selection = selection.subset(variable='pr').to_datetime()
# get a timeseries of files
selection = selection.select_timerange([dt.datetime(1990,1,1),dt.datetime(2000,1,1)])
print(selection)

Use the ESGF module to create your filename using an attribute dictionary. Use the CORDEX filenaming convetion (or create your own!):

from cordex import ESGF as esgf

root       = '/my_root'

# define attributes
attributes   = {'institute_id'    : 'GERICS',
                'product'         : 'output',
                'model_id'        : 'GERICS-REMO2015',
                'experiment_id'   : 'evaluation',
                'driving_model_id': 'ECMWF-ERAINT',
                'variable'        : 'pr',
                'rcm_version_id'  : 'v1',
                'date'            : 'v20200221',
                'frequency'       : 'day',
                'CORDEX_domain'   : 'EUR-11',
                'suffix'          : 'nc',
                'ensemble'        : 'r1i1p1'}

# we use the CORDEX convention as example
convention = esgf.CORDEX()
# print the convention patterns
print(convention.path_conv.conv_str)
print(convention.filename_conv.conv_str)
# only filename
filename = convention.filename(**attributes, startdate='20010101', enddate='20010131')
# only path
path     = convention.path(**attributes, startdate='20010101', enddate='20010131')
# only filename with path
file     = convention.pattern(root, **attributes, startdate='20010101', enddate='20010131')

Use the conventions module to create your own filenaming conventions:

from cordex import conventions as conv

# create your own filename convention string and list
filename_conv_str  = 'my_convention_{variable}_{model_id}_{domain_id}.nc'
path_conv_list     = ['model_id','variable']

# create conventions for filename and path
filename_conv = conv.FileNameConvention(filename_conv_str)
path_conv     = conv.FilePathConvention(path_conv_list)


# now define your attributes to fill the templates.
root = '/my_root'
attributes = {'model_id'        : 'GERICS-REMO2015',
              'variable'        : 'pr',
              'domain_id'       : 'EUR-11'}

# create filename and path
filename = filename_conv.pattern(**attributes)
path     = path_conv.pattern(root, **attributes)

# create combined file convention
file_conv = conv.FileConvention(path_conv, filename_conv)

# create full filename with path
file = file_conv.pattern(root, **attributes)

Requirements

• python3.6 or higher
• numpy
• pandas
• (xarray)
• netCDF4
• parse

Note

This project has been set up using PyScaffold 3.2.3. For details and usage information on PyScaffold see https://pyscaffold.org/.