Metacognitive Learning Tool box

https://re.is.mpg.de

What Is This?

This repository contains two modules used to analyse metacognitive learning in human. src/computationa_microscope contains the code for the computational microscope src/mcrl_modelling contains the code to fit the metacognitive reinforcement learning models (MCRL) to the data.

How To Use This

Installation

To install as a package in python 3.8+:

git clone https://github.com/RationalityEnhancement/mcl_toolbox.git
pip install -e .
cd mcl_toolbox

Import data

Assuming you are working with the Mouselab-MDP repository and with a postgres database:

1. Navigate to src/import_data
2. Put your dataclip (csv file) in the folder src/import_data/data
3. Run src/import_data/reformat_csv.py to create the required mouselab-mdp.csv and participants.csv for each condition as well as an overall file

Note: you might have to use your own import code depending on your requirements.

Analysis modules

1. Navigate to mcl_toolbox/
2. Run python mcl_toolbox/infer_participant_sequences.py to analyse the click sequence of each participant
3. Run python mcl_toolbox/infer_sequences.py to analyse the click sequence average over conditions
4. Run python mcl_toolbox/fit_mcrl_models.py to fit the MCRL models

Note: see each folder or each file for detailed instructions.

Testing

There are very simple integration tests in tests/ to run analysis modules quickly to check whether analysis modules will run. To run these, run:

chmod +x test_analysis.sh
./test_analysis.sh

#TODO unit tests

Development

Please fork your own feature branch and merge in the dev branch.