Climpy helps you to work with climata data to analyse climate related event. The functionalities include

• Finding temporal or spatio-temporal events defined by a preset conditions or defining your conditions
• Selecting events based on some preset selectors or custom selectors
• Doing spatio-temporal cross validation and writing them in machine learning friendly formats like zarr
• Providing en exhaustive set of metrics used with climate data or climate related hazards. ](https://codecov.io/gh/ClimAI/climpy)

The package is divided into three parts

• The package has three modules
  • transform: It has a set of conditions forming a criterion which can be applied to any dataset. A subclass of criterion called hazard helps to create an event. Select class helps in selecting the events based on some rules. Lastly datasets can be linked to events.
  • ml_data: The module helps to create X, Y pairs of machine learning dataset. It focuses on rigorous spatio-temporal cross validation and writing datasets in efficient dataset formats like zarr
  • metrics: The module provides an exhaustive list of metrics for spatio-temporal cross validation

Installation:

• The package requires libgeos-dev. On Debain(linux) it can be installed using the command sudo apt -y install libgeos-dev. On mac one can use brew install geos Cartopy has libgeos-dev as a dependency. # More testing required.
• Then the package can be installed using pip install climpy