Training for behavioral analysis

Data

The data from Mearns et al. (2020) is available on Mendeley here (total uncompressed size is ~25 GB).

After extracting the data, the directory tree should be organized as follows:

|-- data_directory {call this whatever you want}
    |-- kinematics
        |-- 2017081001
            |-- 2017081001171128.csv
            |-- 2017081001171228.csv
            ...
        |-- 2017081002
        ...
    |-- transitions
        |-- T.npy
        |-- transition_matrices.npy
        |-- USVa.npy
        |-- USVs.npy
        |-- WTW.npy
    |-- exemplar_distance_matrix.npy
    |-- exemplars.csv
    |-- isomap.npy
    |-- mapped_bouts.csv

Environment

To install the environment run:

conda env create -f environment.yml 

To activate the environment run:

conda activate behavior_analysis_training 

To install the ethomap package run:

pip install ethomap

Jupyter

Install JupyterLab in your base anaconda environment:

conda install -c conda-forge jupyterlab

To make your environment accessible in JupyterLab, first activate your environment (see above) and run:

conda install -c anaconda ipykernel

python -m ipykernel install --user --name=behavior_analysis_training

To launch JupyterLab:

jupyter-lab

If this does not work and you get a 404 error try running the following in your base conda environment:

jupyter serverextension enable --py jupyterlab --user

conda install -c conda-forge nodejs

Tutorial

To run the tutorial, open the tutorial.ipynb in JupyterLab and set the environment to behavior_analysis_training.