mleap

<a href="https://combust.github.io/mleap-docs/"><img src="logo.png" alt="MLeap Logo" width="176" height="70" /></a>

[![Build Status](https://github.com/combust/mleap/actions/workflows/test.yml/badge.svg?branch=master)](https://github.com/combust/mleap/actions/workflows/test.yml)
[![Maven Central](https://img.shields.io/maven-central/v/ml.combust.mleap/mleap-base_2.13)](https://central.sonatype.com/artifact/ml.combust.mleap/mleap-base_2.13)
[![PyPI](https://img.shields.io/pypi/v/mleap)](https://pypi.org/project/mleap/)

Deploying machine learning data pipelines and algorithms should not be a time-consuming or difficult task. MLeap allows data scientists and engineers to deploy machine learning pipelines from Spark and Scikit-learn to a portable format and execution engine.

## Documentation

Documentation is available at [https://combust.github.io/mleap-docs/](https://combust.github.io/mleap-docs/).

Read [Serializing a Spark ML Pipeline and Scoring with MLeap](https://github.com/combust/mleap/wiki/Serializing-a-Spark-ML-Pipeline-and-Scoring-with-MLeap) to gain a full sense of what is possible.

## Introduction

MLeap provides a performant, portable execution engine and serialization format (Bundle.ML) for machine learning pipelines. Train a pipeline with Spark or Scikit-learn, export it once, and execute it anywhere on the JVM without a dependency on Spark, sklearn, numpy, or pandas.

Key features:

1. Core execution engine implemented in Scala, with [Spark](http://spark.apache.org/), PySpark, and Scikit-learn support.
2. Export a Spark or Scikit-learn pipeline and run it on the lightweight MLeap Runtime.
3. Two portable serialization formats: JSON and Protobuf.
4. Custom data types and transformers for MLeap data frames and pipelines.
5. Full parity tests between Spark and MLeap pipelines.
6. Optional Spark transformer extension that adds to Spark's built-in transformers.

## Dependency Compatibility Matrix

Other versions besides those listed below may also work (especially more recent Java versions for the JRE), 
but these are the configurations which are tested by mleap.

| MLeap Version | Spark Version | Scala Version    | Java Version | Python Version | XGBoost Version | Tensorflow Version |
|---------------|---------------|------------------|--------------|----------------|-----------------|--------------------|
| 0.26.0        | 4.1.2         | 2.13.18          | 17           | 3.10 - 3.13    | 3.4.0           | 2.16.2             |
| 0.25.3        | 4.1.2         | 2.13.18          | 17           | 3.10 - 3.13    | 3.4.0           | 2.16.2             |
| 0.25.2        | 4.1.2         | 2.13.18          | 17           | 3.10 - 3.13    | 2.0.3           | 2.16.2             |
| 0.25.1        | 4.1.1         | 2.13.17          | 17           | 3.10 - 3.13    | 2.0.3           | 2.16.2             |
| 0.25.0        | 4.1.0         | 2.13.17          | 17           | 3.10 - 3.13    | 2.0.3           | 2.10.1             |
| 0.24.2        |