Introduction

Unfortunately, pyStan is not compatible with Windows/Mac by default. However, this repository provides a containerized version that can be used on any environment. Users can add Python scripts in the ./stan/ directory, which will be automatically copied to the containerized Linux environment.

Installation

To use docker-compose to execute a particular script named fululu.py located in the ./stan/ directory, follow these steps:

1. Open a terminal and navigate to the directory containing the docker-compose.yml file and the fululu.py script.

2. Build the Docker image by running the following command:

   docker-compose build
   

Usage

1. Once the build is complete, start the container and run the fululu.py script by running the following command:

   docker-compose run pystan python fululu.py
   

   This command starts the container and runs the fululu.py script using the Python interpreter installed in the container.

2. Alternatively, you can use the following command to start a Jupyter Notebook environment, which allows for creating notebooks for Bayesian data analysis:

   docker-compose up
   

   This command starts a Jupyter Notebook environment accessible at http://127.0.0.1:8888.

3. To execute Python from inside the docker, run the following command:

   docker-compose run pystan python
   

4. Once any of the commands finish running, the container will automatically stop and remove itself.

Editing Scripts

The scripts can be edited using any text editor. At each launch of the script using docker-compose, the scripts will be copied to the Linux environment and executed there.

Note: Certain functions, such as plt.plot(), do not work natively. It is better to use plt.savefig() instead.