additory

additory

Elegant data operations for DataFrames

A Rust-powered Python library for intuitive data transformations, lookups, synthetic data generation, and DataFrame comparison — with Polars and Pandas.

Installation

pip install additory

Requirements: Python 3.9+ · Polars (required) · pyarrow (required) · Pandas (optional)

Quick Start

import additory as add
import polars as pl

Bring data from another DataFrame

orders = pl.DataFrame({'order_id': [1, 2], 'customer_id': [101, 102], 'amount': [250, 180]})
customers = pl.DataFrame({'customer_id': [101, 102], 'name': ['Alice', 'Bob']})

result = add.to(orders, customers, 'name', 'customer_id')

Transform data

df = pl.DataFrame({'price': [100, 200, 300], 'quantity': [2, 3, 1]})

result = add.transform('@calc', df, strategy={'total': 'price * quantity', 'tax': 'total * 0.1'})
result = add.transform('@filter', df, where='price > 150')
result = add.transform('@sort', df, by='price', strategy='desc')
result = add.transform('@aggregate', df, by='category', strategy={'price': 'sum'})

Generate synthetic data

result = add.synthetic('@new', n=1000, strategy={
    'age': 'normal(40, 10)',
    'salary': 'normal(75000, 15000)',
})

# Augment existing data
result = add.synthetic(df, n=100)

Compare DataFrames

diff = add.scan('@diff', old=df_jan, new=df_feb)

Use dynamic expressions

patients = pl.DataFrame({'weight': [70, 85], 'height': [1.70, 1.80]})
result = add.bmi(patients)

Pipe compatibility

result = (
    orders
    .pipe(add.to, customers, 'name', 'customer_id')
    .pipe(add.to, products, 'price', 'product_id')
)

Core Functions

Transform Modes

@calc · @filter · @sort · @aggregate · @harmonize · @round · @transpose · @extract · @onehot · @label · @deduce · @split

Features

• 🔗 Intuitive Lookups — bring columns from external sources with natural syntax
• ⚡ 12 Transform Modes — calculate, filter, sort, aggregate, encode, impute
• 🔍 DataFrame Diff — compare snapshots with add.scan('@diff')
• 🧮 Dynamic Expressions — 74 built-in formulas across 7 categories, extensible via .add files
• 🎲 Synthetic Data — generate realistic test data or augment existing datasets
• 📊 Lineage Tracking — trace data provenance with lineage=True
• 🔄 Pipe Compatibility — fluent chaining with df.pipe()
• 🚀 Rust Performance — ~95% Rust core via PyO3
• 🐼 Polars & Pandas — works with both DataFrame libraries

Documentation

📚 sekarkrishna.github.io/additory

Version

Current version: 0.1.3a11

License

MIT License — see LICENSE for details.

Author: Krishnamoorthy Sankaran