Development

Everything you need to contribute to ModernDiD and understand its internals. For usage documentation see the User Guide.

• Contributing to ModernDiD
  • Development process
  • Guidelines
  • Stylistic guidelines
  • Code quality tools
  • Test coverage
  • R validation tests
  • Building documentation
  • Continuous integration
  • Registering new public API
  • Dependency management
• Architecture and API Design
  • Overview
  • The Preprocessing Pipeline
  • Consistent Argument Naming
  • Result Object Design
  • Implementation Standards
  • The Formatting System
  • Creating a New Estimator
  • Plotting Architecture
  • Integration with Sensitivity Analysis
• Creating a New Estimator
  • How estimators plug into the pipeline
  • Module Layout
  • Step 1: Define the Configuration
  • Step 2: Integrate with the Preprocessing Pipeline
  • Step 3: Define the Result Object
  • Step 4: Implement the Estimator
  • Step 5: Add Aggregation Support (Multi-Period Estimators)
  • Step 6: Add Formatted Output
  • Step 7: Add Plotting Support
  • Step 8: Export the Public API
  • Step 9: Add Maketables Support
  • Step 10: Add Distributed Support (optional)
  • Step 11: Write Tests
• Testing ModernDiD
  • How to run the test suite
  • How to write tests
• Git Workflow and Conventions
  • Commit message format
  • Branch naming
  • Working with branches
  • Pull request workflow
• Reviewing and Maintainer Guidelines
  • For reviewers
  • For maintainers
• Debugging Guide
  • General strategies
  • Numerical issues
  • Debugging Numba-compiled code
  • Debugging CuPy and GPU code
  • Debugging distributed execution
  • Test failure patterns
  • Using a debugger
  • Profiling
• Benchmarking
  • Running benchmarks
  • Interpreting results
  • Benchmark structure
  • Adding a new benchmark
• Distributed Computing
  • Scope and design rule
  • Execution modes
  • Cell-level decomposition
  • Distributed nuisance estimation
  • Tree-reduce aggregation
  • Wide-pivot optimization
  • Cohort-level parallelism
  • Influence function streaming
  • End-to-end data flow
  • Distributed bootstrap
  • Memory management
  • Spark backend mechanics
• Release Process
  • Overview
  • Pre-release checklist
  • Preparing the release
  • Automated publishing
  • Post-release
  • Hotfix releases
  • Troubleshooting