zeppelin-sqlflow

Project Description

Integrate Ant Group open source machine learning tool SQLFlow into the web-based notebook tool Apache Zeppelin to expand the interpreter language supported by Zeppelin.

Motivation

Integrate SQLFlow into Apache Zeppelin:

• Rich SQLFlow usage scenarios. Users can also use SQLFlow in this notebook environment;
• Expanded the interpreter language supported by Zeppelin;
• SQLFlow can be used in conjunction with other languages or data processors on this platform to fully demonstrate its convenience in AI applications.

Quick Overview

Here are examples for training a TensorFlow DNNClassifier model using sample data Iris.train, and running prediction using the trained model in Zeppelin environment.

How to install zeppelin-sqlflow

The installation is divided into the following steps:

• Prepare the interpreter project code jar package zeppelin-sqlflow-0.9.0.jar and the configuration file interpreter-setting.json;

• Switch to the installation directory of Zeppelin and create the interpreter/sqlflow subdirectory:

mkdir ~/zeppelin-0.9.0-SNAPSHOT/interpreter/sqlflow

• Copy the jar package and configuration file in step 1 to the directory created in step 2:

# This is the path on my machine, for reference only.
cp ~/zeppelin-sqlflow-0.9.0.jar ~/zeppelin-0.9.0-SNAPSHOT/interpreter/sqlflow
cp ~/interpreter-setting.json ~/zeppelin-0.9.0-SNAPSHOT/interpreter/sqlflow

• Start zeppelin

cd ~/zeppelin-0.9.0-SNAPSHOT/bin/
./zeppelin-daemon.sh start

• Refer to the manual OperationGuide for specific use steps.

Developer introduction

The project was initiated by the foundation platform development team of the IT department in PCCC. It aims to strengthen the function expansion of the Zeppelin in data exploration and model training，which provides data scientists with more comprehensive and rich data processing tools.