A toolbox of ensemble based data-assimilation algorithms for use with Firedrake.
As a prerequisite, it is required to have Firedrake installed alongside it's dependencies. Instructions on how to this can be found at: http://firedrakeproject.org/download.html. Then, it is required to have the Firedrake virtualenv activated; instructions for this can be found by again following the aforementioned link.
Finally, for some demos to work it is required to have Firedrake-MLMC installed (a side-package of http://firedrakeproject.org). Instructions on how to do this can be found at: https://github.com/firedrakeproject/firedrake-mlmc.
Then, to install firedrake_da, type the following commands into the terminal:
git clone https://github.com/firedrake_da
pip install -e ./firedrake_da
To build the documentation, type
make html
whilst in the docs
repository, or alternatively go to
For any enquiries, please contact: a.gregory14@imperial.ac.uk
Mr. Alastair Gregory, Research Postgraduate, Imperial College London
The C code for the Earth Mover's Distance computations in this package, used to create kernels for Firedrake is from the following reference: Y. Rubner, C. Tomasi, and L. J. Guibas. The earth mover’s distance as a metric for image retrieval. IJCV, 2000.
I would like to thank Dr. Colin Cotter and the Firedrake team at Imperial College London for help in designing and developing this open source package.