datamat 0.2.0a1

Matrix Operations on Data Arrays

Table of Contents

1. DataMat
   1. Features
   2. Repository Layout
   3. Getting Started
   4. Development Workflow
   5. Documentation
   6. Online Docs
   7. Optional Dependencies
   8. License

DataMat

Matrix Operations on Data Arrays. A pandas-friendly toolkit that wraps rich matrix algebra primitives in DataFrame/Series-aware containers (DataMat, DataVec).

Features

• Drop-in replacements for many numpy matrix operations that preserve pandas metadata.
• Helpers for aligning MultiIndex structures and constructing dummy variables from categorical sources.
• A DataVec~/~DataMat API that mirrors pandas while adding linear-algebra conveniences (projection, leverage, SVD, etc.).

Repository Layout

src/datamat/        Core source code (DataMat, DataVec, utilities)
docs/               Historical literate notebooks and design notes
tests/              Unit tests
scripts/            Development tooling (e.g., pre-push hook)

Getting Started

1. Install Python 3.11 (e.g., via pyenv) and Poetry (https://python-poetry.org/docs/).

2. Create and initialise the virtual environment:

   poetry env use 3.11
   poetry install --with dev
   # optional extras, e.g. lsms, can be added via `poetry add`
   

3. Run the quality gates and test suite:

   poetry run ruff check .
   poetry run black --check .
   poetry run mypy src tests
   poetry run pytest
   # or simply
   make check
   

Development Workflow

• Contribution guidelines: CONTRIBUTING.org
• Agent-specific policies: AGENTS.org
• Pre-push automation: ln -s ../../scripts/pre-push.sh .git/hooks/pre-push

Documentation

The historical literate source from the original Metrics Miscellany project lives in docs/metrics_miscellany.org. It remains the best place to reference derivations and design notes. Current development happens directly in the Python modules under src/datamat/; augment the Org materials when adding new capabilities.

Online Docs

Rendered documentation generated by MkDocs is available at https://ligon.github.io/DataMat/.

Optional Dependencies

• Robust Stata ingestion relies on lsms.tools.from_dta. Install the lsms package if you plan to call datamat.read_stata.

License

The project continues under the original license; see LICENSE.txt.