Skip to content

Contributing

Contributions are welcome — bug fixes, data updates, new features, and documentation improvements. This page covers the development workflow.

Set up the development environment

# Clone the repository
git clone https://github.com/Jaypatel1511/nmtc-application-builder.git
cd nmtc-application-builder

# Create and activate a virtual environment
python -m venv .venv
source .venv/bin/activate     # Windows: .venv\Scripts\activate

# Install in editable mode with all dev dependencies
pip install -e ".[dev]"

The [dev] extra installs pytest, pytest-cov, jupyter, and all output/viz libraries (python-docx, openpyxl, reportlab, matplotlib).

Running tests

# Run the full test suite
PYTHONPATH=. pytest tests/ -v

# Run a specific test file
PYTHONPATH=. pytest tests/intelligence/test_win_probability.py -v

# Run with coverage
PYTHONPATH=. pytest tests/ -v --cov=nmtcapp --cov-report=term-missing

Tests are organized to mirror the source tree:

tests/
  core/             # Application, CDEProfile, Pipeline, PipelineProject
  data/             # Historical awards, schema constants
  intelligence/     # Distress, geographic, sector, impact, win probability, recommendations
  optimizer/        # PipelineOptimizer, OptimizationConstraints
  renderers/        # Word, Excel, PDF, Markdown builders
  sections/         # Section A–E generators
  tables/           # Table builders (distress, geographic, impact, etc.)
  validation/       # Completeness, consistency, eligibility, readiness score
  visualization/    # Plot functions
  integrations/     # Adapter stubs

All tests should pass before submitting a pull request.

Code style

  • Formatting: The codebase follows standard Python formatting conventions. Functions have type annotations for parameters and return types where practical.
  • Docstrings: Public classes and methods have docstrings following the existing style — one-line summary, extended description if needed, and an Example:: block showing runnable code.
  • Module headers: Each module has a top-level docstring explaining its purpose, the data sources it draws from (if applicable), and any important methodology notes.
  • Methodology disclosures: Modules that produce alignment scores or optimization results carry a _METHODOLOGY string constant that is always exposed on result objects. Never remove or weaken these disclosures — they are a core design requirement.
  • No implicit imports: All public-facing imports go through __init__.py. Internal cross-module imports use full absolute paths.

Submitting a pull request

  1. Fork the repository and create a feature branch from main.

  2. Write tests for any new functionality. The existing test suite uses pytest fixtures defined in tests/conftest.py. Use CDEProfile.sample() and Pipeline.sample() in tests to avoid external API calls.

  3. Verify all tests pass: 

    PYTHONPATH=. pytest tests/ -v

  4. Update the relevant documentation page in docs/ if your change affects user-facing behavior, API signatures, or methodology.

  5. Open a pull request against the main branch with a clear description of:

  6. What the change does
  7. Why it is needed
  8. Any methodology changes (if applicable)

Reporting issues

Use the GitHub issue tracker to report bugs, request features, or ask questions.

When reporting a bug, please include: - Python version and OS - Installed library version (import nmtcapp; print(nmtcapp.__version__)) - Minimal reproducible example - Full traceback if applicable

Contributing data updates

When the CDFI Fund publishes new award announcement data or annual reports, the historical_awards.py constants should be updated. If you have access to new data:

  1. Update the relevant constants in nmtcapp/data/historical_awards.py
  2. Add the new round entry to NMTC_AWARD_ROUNDS with a source comment
  3. Update APPLICATION_VOLUME_TRENDS lists
  4. Update WINNER_DISTRESS_PATTERNS, WINNER_GEOGRAPHIC_PATTERNS, WINNER_IMPACT_BENCHMARKS, or WINNER_SECTOR_PATTERNS if new annual report data changes the aggregate statistics
  5. Update the module docstring to reflect the new data coverage period
  6. Run the full test suite — some tests assert on specific constant values and may need updating
  7. Submit a pull request with the data source URL in the commit message

Data contributions are particularly valuable as they directly improve the accuracy of alignment scoring for all users.

Running the example notebooks

The examples/ directory contains three Jupyter notebooks that demonstrate the full platform workflow:

cd examples
jupyter notebook
  • 01_quickstart.ipynb — 5-minute end-to-end workflow
  • 02_full_application_walkthrough.ipynb — Complete application preparation with all output formats
  • 03_intelligence_and_optimization.ipynb — Deep dive into scoring, recommendations, and optimizer

To re-execute all notebooks from scratch (useful before a release):

PYTHONPATH=. jupyter nbconvert --to notebook --execute examples/01_quickstart.ipynb
PYTHONPATH=. jupyter nbconvert --to notebook --execute examples/02_full_application_walkthrough.ipynb
PYTHONPATH=. jupyter nbconvert --to notebook --execute examples/03_intelligence_and_optimization.ipynb