Skip to main content

UCLALES utilities

Project description

UCLALES utils in python

This package contains utilities for working with output from the UCLALES model.

Installation

Install uclales-utils with pip from pypi to get the most recent tagged version:

pip install uclales-utils

or directly from github to get the development version:

pip install git+https://github.com/leifdenby/uclales-utils#egg=uclales-utils

Usage

Extracting 2D cross-sections and 3D fields from UCLALES output

Because UCLALES creates a netCDF for each individual core (when running multi-core simulations using MPI) these files must be aggregated together to extract the full 3D field (or a 2D cross-section) for a variable. uclales-utils has functionality implemented to extract the 3D field for a single timestep or 2D cross-section of a specific variable. To make the extraction faster, and to break the extraction down into individual steps that can be checked, this is implemented using the luigi pipeline package. Executing the pipeline may either be done using a single worker, or if your computer has multiple CPUs you may speed up the extraction process by using multiple workers.

For serial executing of the extraction run

python -m luigi --module uclales.output Extract --kind <3d or 2d> --file-prefix <file-prefix> --var-name <variable> [--tn <timestep>] [--orientation <cross-section-orientation>] --local-scheduler

For example, to extract the 3D vertical velocity (w) field at the 5th timestep (counting the initial time at 0) from a collecting of output files prefixed by rico in the filename (i.e. the 3D files are called rico.########.nc)

python -m luigi --module uclales.output Extract --kind 3d --file-prefix rico --tn 5 --var-name w --local-scheduler

Or to extract say the 2D field liquid-water path (lwp) you would run

python -m luigi --module uclales.output Extract --kind 2d --file-prefix rico --var-name lwp --orientation xy --local-scheduler

You can optionally provide the arguments --source-path and --dest-path to set which paths to search for input from and where the output will be stored (the default is the current working path by default). Intermediate files will be stored in partials.

To run the extraction across multiple workers in parallel you must start luigid in a separate process, and then run the above command replacing --local-scheduler with --workers <number-of-workers>

For example if you have 8 cores on your machine you might run

python -m luigi --module uclales.output Extract --kind 3d --file-prefix rico --tn 5 --var-name w --workers 8

While luigid is running you can check the progress on the extraction process by using luigi's web-interface and opening the URL http://localhost:8082/ in your browser.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

uclales_utils-0.1.4-py3-none-any.whl (16.9 kB view details)

Uploaded Python 3

File details

Details for the file uclales_utils-0.1.4-py3-none-any.whl.

File metadata

File hashes

Hashes for uclales_utils-0.1.4-py3-none-any.whl
Algorithm Hash digest
SHA256 72af2f583f73d962a07dca726f1b3673ba3dff3781db919ddcb407c8e185c10f
MD5 416116135493a85dd1430b8987187884
BLAKE2b-256 08d87aa3b42cba1781c7f16b804ed20af483993bc9fef9a6b1370451f58ed04f

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page