Better Code, Better Science

An open-source book on building better code for science using AI.

Author: Russ Poldrack

Contributors: @giorgioarcara, @yarikoptic

Content will be posted on Substack and added here as it is released.

A rendered version of the complete book is available at https://poldrack.github.io/BetterCodeBetterScience/.

Table of contents (see here for a more detailed roadmap)

• Introduction
  • Part 1
  • Part 2
• Essential Tools and Techniques
  • Part 1: Version control
  • Part 2: Large Language Models
  • Part 3: Managing environments
• Principles of Software Engineering
  • Part 1: Intro
  • Part 2: Clean Coding
  • Part 3: Modularity and the "God object"
  • Part 4: Global variables
  • Part 5: Defensive coding
  • Part 6: Coding portably
  • Part 7: Code formatting tools
  • Part 8: Managing technical debt
• Software Testing
  • Part 1: Introduction
  • Part 2: The structure of a good software test
  • Part 3: The challenges of AI-generated tests
  • Part 4: Test-driven development and AI-assisted coding
  • Part 5: Text fixtures and mocking
  • Part 6: Parameterized testing
  • Part 7: Property-based testing
  • Part 8: Automated testing with GitHub Actions
  • Part 9: Optimizing the testing workflow
• Coding with AI
  • Using AI coding assistants
  • AI-assisted coding approaches
• Project Structure and Management
• Validating Scientific Analyses
• Optimizing Performance
• Data Organization and Management
• Workflow Management
• High-Performance Computing
• Sharing Research Objects

Setup to run code

Database examples

The examples in Chapter 7 require several database systems to be installed.

Mongodb

Install: https://www.mongodb.com/docs/v7.0/administration/install-community/

Neo4j

Install: https://neo4j.com/docs/operations-manual/current/installation/