Installation

We provide two ways to use pyAMNESIA.

  • If you want a lightweight tool that runs in a command-line interface, you may want to install pyAMNESIA as a standalone tool;
  • If you want to use CICADA to run the tool and have a pretty GUI to guide you, then you may want to install pyAMNESIA as a CICADA module.

We advise that you use conda to create a virtual environment for the tool.

conda create -n pyamnesia python==3.8  # or python==3.7
conda activate pyamnesia

Install pyAMNESIA

You only need to git clone pyAMNESIA:

# clone the repo
git clone https://gitlab.com/cossartlab/pyamnesia.git
cd pyamnesia
# install the dependencies
sh requirements.sh

Note

On Windows, you may not be able to use the sh command. To solve this, you may either download Cygwin or write the commands in requirements.sh yourself:

pip install -r requirements.txt
conda install -c conda-forge skan
conda install -c conda-forge hdbscan

The installation relies on conda to install skan, which is the package we use to skeletonize images. If you cannot use conda, please install skan and hdbscan with pip; we cannot guarantee that it will run smoothly though.

If you want to use pyAMNESIA as a standalone tool, no need for you to read the next section and you can read the how-to page to run the tool.

Install CICADA

In order to have a GUI to guide you during the pyAMNESIA analysis, you can install CICADA using pip, or from source (for more information, please visit CICADA’s documentation):

# using pip
pip install -U pycicada
# from source
git clone https://gitlab.com/cossartlab/cicada.git

Warning

As for now, CICADA still is a work in progress, we advise that you clone the stable version we worked on and pip install some additional packages to make it run:

# install stable version
git clone -b stable-pyamnesia https://gitlab.com/theodumont/cicada.git
cd cicada
# pip install some packages
pip install -r requirements.txt
pip install neo elephant

CICADA and pyAMNESIA are now installed and you can go to the how-to page to run the tool.