databricks/spark-sklearn

(Deprecated) Scikit-learn integration package for Apache Spark

Deprecation

This project is deprecated. We now recommend using scikit-learn and Joblib Apache Spark Backend to distribute scikit-learn hyperparameter tuning tasks on a Spark cluster:

You need pyspark>=2.4.4 and scikit-learn>=0.21 to use Joblib Apache Spark Backend, which can be installed using pip:

pip install joblibspark

The following example shows how to distributed GridSearchCV on a Spark cluster using joblibspark. Same applies to RandomizedSearchCV.

from sklearn import svm, datasets
from sklearn.model_selection import GridSearchCV
from joblibspark import register_spark
from sklearn.utils import parallel_backend

register_spark() # register spark backend

iris = datasets.load_iris()
parameters = {'kernel':('linear', 'rbf'), 'C':[1, 10]}
svr = svm.SVC(gamma='auto')

clf = GridSearchCV(svr, parameters, cv=5)

with parallel_backend('spark', n_jobs=3):
    clf.fit(iris.data, iris.target)

Scikit-learn integration package for Apache Spark

This package contains some tools to integrate the Spark computing framework with the popular scikit-learn machine library. Among other things, it can:

• train and evaluate multiple scikit-learn models in parallel. It is a distributed analog to the multicore implementation included by default in scikit-learn
• convert Spark's Dataframes seamlessly into numpy ndarray or sparse matrices
• (experimental) distribute Scipy's sparse matrices as a dataset of sparse vectors

It focuses on problems that have a small amount of data and that can be run in parallel. For small datasets, it distributes the search for estimator parameters (GridSearchCV in scikit-learn), using Spark. For datasets that do not fit in memory, we recommend using the distributed implementation in `Spark MLlib.

This package distributes simple tasks like grid-search cross-validation. It does not distribute individual learning algorithms (unlike Spark MLlib).

Installation

This package is available on PYPI:

pip install spark-sklearn

This project is also available as Spark package.

The developer version has the following requirements:

• scikit-learn 0.18 or 0.19. Later versions may work, but tests currently are incompatible with 0.20.
• Spark >= 2.1.1. Spark may be downloaded from the Spark website. In order to use this package, you need to use the pyspark interpreter or another Spark-compliant python interpreter. See the Spark guide for more details.
• nose (testing dependency only)
• pandas, if using the pandas integration or testing. pandas==0.18 has been tested.

If you want to use a developer version, you just need to make sure the python/ subdirectory is in the PYTHONPATH when launching the pyspark interpreter:

PYTHONPATH=$PYTHONPATH:./python:$SPARK_HOME/bin/pyspark

You can directly run tests:

cd python && ./run-tests.sh

This requires the environment variable SPARK_HOME to point to your local copy of Spark.

Example

Here is a simple example that runs a grid search with Spark. See the Installation section on how to install the package.

from sklearn import svm, datasets
from spark_sklearn import GridSearchCV
iris = datasets.load_iris()
parameters = {'kernel':('linear', 'rbf'), 'C':[1, 10]}
svr = svm.SVC(gamma='auto')
clf = GridSearchCV(sc, svr, parameters)
clf.fit(iris.data, iris.target)

This classifier can be used as a drop-in replacement for any scikit-learn classifier, with the same API.

Documentation

API documentation is currently hosted on Github pages. To build the docs yourself, see the instructions in docs/.