rainymotion

<img src="https://raw.githubusercontent.com/hydrogo/rainymotion/master/docs/notebooks/images/rainymotion_logo.png" alt="rainymotion logo" width="100px"/>

rainymotion: Python library for radar-based precipitation nowcasting based on optical flow techniques

Idea

The main idea of the rainymotion library is to provide an open baseline solution for radar-based precipitation nowcasting.

Development

rainymotion is based only on free and open source software -- we tried to make a clue between the best scientific libraries to make them work together for providing reliable precipitation nowcasts.

<img src="https://raw.githubusercontent.com/hydrogo/rainymotion/master/docs/notebooks/images/rainymotionisbasedonfoss.png" alt="rainymotion logo" width="400px">

Documentation

rainymotion documentation is hosted by Read the Docs.

Example

To obtain precipitation nowcasts using the rainymotion models you need to follow the only three steps:

1. initialize the model (you have 4 variants)
2. load data to the model's placeholder
3. run the model

# import rainymotion model
from rainymotion.models import Dense

# initialize model instance
model = Dense()

# load the data using your custom DataLoader function
model.input_data = DataLoader("/path/to/data")

# run the model
nowcasts = model.run()

Reference

Please cite rainymotion as Ayzel, G., Heistermann, M., and Winterrath, T.: Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1), Geosci. Model Dev., 12, 1387-1402, https://doi.org/10.5194/gmd-12-1387-2019, 2019.