Formose Reservoir Computation

Repository containing all data and code/notebooks relevant to the Formose Reservoir Computation publication.

Repository structure

• data contains the extracted ion signals, input flow profiles, as well as the processed data to perform the various reservoir computation tasks
• preprocess contains notebooks converting the raw ion signals together with the input flow profiles into processed data files
• analysis contains notebooks used to perform the various reservoir computation tasks and for creation of all data-related figures in the publication and supporting information.
• plots and plots_extended contain all data-related figures for the main manuscript and supporting information respectively (these are generated from the notebooks in the analysis directory)
• tables contains tables in csv and latex format with scores and accuracies of training tasks

Installation

The code in this repository requires Python 3.11 together with a few standard packages (numpy, pandas, scikit-learn, etc...). All packages can be installed by creating a Conda environment from environment.yaml.

A typical installation should take less than 1 hour.

Important: Installing AMICI requirements

For calculation of the in silico carbon metabolism model, we use AMICI. To correctly install AMICI, extra steps are required: Installation of the AMICI Python package has the following prerequisites:

• Python>=3.9
• SWIG>=3.0
• CBLAS compatible BLAS library (e.g., OpenBLAS, CBLAS, Atlas, Accelerate, Intel MKL)
• a C++17 compatible C++ compiler and a C compiler (e.g., g++, clang, Intel C++ compiler, mingw)

Ubuntu 22.04

On Ubuntu, the requirements can be installed via apt:

sudo apt install libatlas-base-dev swig

# optionally for HDF5 support (recommended):
sudo apt install libhdf5-serial-dev

OSX

On OSX, the requirements can be installed via homebrew:

brew install swig

# optionally for HDF5 support:
brew install hdf5

# optionally for parallel simulations:
brew install libomp

Windows

Installation on Windows directly is not recommended. We advice using the Windows Subsystem for Linux (WSL) and follow the instructions for installation on linux.

Create a Conda environment

After installing the AMICI requirements, the environment with all Python packages can be installed by running:

conda env create -f environment.yaml

after which the environment can be activated by running:

conda activate formose_rc

Figure fonts

Figures created in the Jupyter notebooks use Arial as a font. This font is not installed on Ubuntu by default, but can be installed by running:

sudo apt install ttf-mscorefonts-installer
sudo fc-cache -f

and subsequently deleting the matplotlib font cache:

rm ~/.cache/matplotlib -rf