A ChRIS statsmodels plugin app

pl-statsmodels is a ChRIS ds plugin which takes in CSV data as input files and creates result summary of Ordinary Least Square fit as output files.

Abstract

An application that takes CSV files as input and fits an OLS model on it. The result summary of the OLS fit is stored in output folder.

Installation

pl-statsmodels is a ChRIS plugin, meaning it can run from either within ChRIS or the command-line.

Local Usage

1. Building container image using podman/docker:

docker build --tag pl-statsmodels -f ./Dockerfile

2. Add the CSV file on which the OLS models has to be fitted. Reference
3. Executing container image

docker run --rm -u $(id -u) -v $(pwd)/test_input:/incoming -v $(pwd)/test_output:/outgoing local/pl-statsmodels statsmodels_tool --columns "Lottery ~ Literacy + Wealth + Region" /incoming /outgoing

NOTE: --columns is a required (string) parameter for the statsmodels_tool utility

Testing

docker run --rm pl-statsmodels nosetests

Release

Steps for release can be automated by Github Actions. This section is about how to do those steps manually.

Increase Version Number

Increase the version number in setup.py and commit this file.

Push Container Image

Build and push an image tagged by the version. For example, for version 1.2.3:

docker build -t docker.io/fnndsc/pl-statsmodels:1.2.3 .
docker push docker.io/fnndsc/pl-statsmodels:1.2.3

Get JSON Representation

Run chris_plugin_info to produce a JSON description of this plugin, which can be uploaded to a ChRIS Store.

docker run --rm localhost/fnndsc/pl-statsmodels:dev chris_plugin_info > chris_plugin_info.json