Skip to content

Installation

Quick install

pip install additory

Requirements

  • Python 3.9+ (tested on 3.9, 3.10, 3.11, 3.12, 3.13)
  • polars ≥ 0.19.0 — required, used internally for all operations
  • pyarrow ≥ 10.0.0 — required for DataFrame serialization between Python and Rust

Optional dependencies

pandas support

additory works with both polars and pandas DataFrames. If you want to pass pandas DataFrames directly:

pip install pandas

When you pass a pandas DataFrame, additory converts it to polars internally, performs the operation, and converts the result back to pandas. You can override this with the as_type parameter on any function.

Verify the installation

import additory as add
print(add.__version__)
# 0.1.3a11

Check that the Rust bindings loaded:

print(add.RUST_AVAILABLE)
# True

If RUST_AVAILABLE is False, only Python-only modes (like @knn imputation) will work. See Troubleshooting for help with Rust binding issues.

Building from source

additory's core is written in Rust with Python bindings via PyO3. To build from source you need:

  1. Rust toolchain — install via rustup
  2. maturin — the build tool for PyO3 projects
# Install maturin
pip install maturin

# Clone the repository
git clone https://github.com/sekarkrishna/additory.git
cd additory

# Build and install in development mode
maturin develop --release

For a wheel you can distribute:

maturin build --release
pip install target/wheels/additory-*.whl

Development mode

Use maturin develop (without --release) for faster builds during development. The --release flag enables Rust optimizations and is recommended for benchmarking and production use.