Deploy an applicaton package into a CRIS#
Example: Duck Demo App#
Duck is a demo web-application for the CLINT project.
Smartduck is a demo web-application of AI-enhanced Climate Science. It is based on the Phoenix web-application from the Birdhouse collection and makes use of the PyWPS Python package, which is an implementation of the Web Processing Service standard from the Open Geospatial Consortium.
Smartduck uses CRAI, a state-of-the-art deep learning based inpainting technology to infill missing values in climate datasets[1].
The current demo gives the possibility to infill near-surface air temperature anomalies in the HadCRUT4 and HadCRUT5 datasets. The input and output netCDF files are handled through an intuitive user interface.
The current demo gives the possibility to infill near-surface air temperature anomalies in the HadCRUT4 and HadCRUT5 datasets.
Install duck with conda#
We need mamba to install the requirements.
Use your existing mamba or install it from here:
conda-forge/miniforge
Get the source:
git clone https://github.com/climateintelligence/duck.git
cd duck
Create the conda environment for duck:
mamba env create
Activate the duck environment:
conda activate duck
Install duck in this environment:
pip install -e .
Start your duck as a web service:
duck start -d
See it is responding by sending a WPS GetCapabilities request:
http://localhost:5000?service=WPS&request=GetCapabilities
You can stop the service with:
duck stop