Skip to content

Latest commit

 

History

History
59 lines (40 loc) · 2.84 KB

README.rst

File metadata and controls

59 lines (40 loc) · 2.84 KB

pyCliTS

What is pyCliTS?

Python Climate Time Series package is open-source python package for easy manipulation with climatic geo-spatial time series such as the reanalysis or CMIP5 outputs, which are usually distributed as netCDF4 files. The package includes functions for:

  • manipulating the data [temporal and spatial slicing, interpolating, subtracting the climatological cycle = anomalising, normalising, filtering, subsampling, etc.]
  • computing continuous complex wavelet transform [CCWT]
  • constructing spatio-temporal surrogate data using Monte-Carlo approach [Fourier transform surrogates, amplitude adjusted FT, iterative amplitude adjusted FT, autoregressive surrogates using the VAR(p) model, multifractal surrogates]
  • computing Singular Spectrum Analysis
  • computing mutual information and conditional mutual information [using equidistant, equiquantal binning and k-nearest neighbour algorithms]
  • constructing an empirical model from spatio-temporal data based on idea of LIMs [linear inverse modelling].

Uses fast numpy, scipy and scikit-learn libraries and offers multi-thread computations when possible [e.g. computing wavelet transform per grid point].

Documentation

Instead of proper documentation [I plan to add it later though!], I created a few examples on how to work with basic class DataField, SurrogateField, SSA and other functions. These can be found in examples folder. The folder also contain climate data to work on, all of them are publicly available and inside example_data folder, you can find disclaimer with links to the datasets.

Dependencies

pyclits relies on the following open source packages

Required:

Recommended:

(All of them are installed via pip automatically when installing this package, except basemap, since it is not on PyPI. Basemap still can be installed via pip using pip install git+https://github.com/matplotlib/basemap.git)

Contributing

All contributions are welcome! Just drop me an email or pull request.

Versions

  • 0.1: initial version
  • 0.2: various minor bug fixes, added VARIMAX rotation as optional parameter in pca_components() method of DataField

License information

pyclits is MIT-licensed, for more information see the file LICENSE.txt