3 Simple Steps to Build Classifai Data Annotation Tool from Source

Classifai is an OS-agnostic data annotation tool open sourced in 2020. It is currently actively used by AI practitioners to build customized use cases for different industries.

Motivation

With the active development roadmap for bug fixes and new features, the build of Classifai from source reflect the latest changes in a timely manner.

Alternatively, stable builds are downloadable from here.

Gist of Content

This walk-through demonstrates the installation and running of Classifai from source code.

While the OS demonstrated in this walk-through is via Windows, Classifai supports MacOS and other Linux-based OSes such as CentOS and Ubuntu.

1. Installation of Java

The only external dependency to install for the very first time is Java of version 14. Here, AdoptOpenJDK 14 is installed from the official download page after choosing the installation package for the desired OS.

Check if Java is correctly installed by typing

java -version

in the command prompt.

There’s also a need to set up JAVA_HOME environment path after installation is completed. JAVA_HOME path should points to the root folder of the Java folder.

Check the folder path of the former installed Java using the command

where java

Then, set the environment path for JAVA_HOME pointing to the root directory of the Java folder.

Check if JAVA_HOME is correctly set by using the command

echo %JAVA_HOME%

2. Download source code

There are 3 options to download source code.

• Download as zip
• Clone with https
• Clone with ssh

For the first option, download the zipped file from Classifai repository in Github. Proceed with unzipping the file in the desired path.

Alternatively, simply copy the classifai repository url and git clone the copied link.

git clone https://github.com/CertifaiAI/classifai.git 
# or 
git clone git@github.com:CertifaiAI/classifai.git

3. Build Classifai

Now, the setup is ready to build Classifai tool.

Open up the command prompt and point the current directory towards Classifai folder.

cd {path}/classifai

For Windows, run the building script startClassifai.bat in the command prompt.

Alternatively, for other Linux-based OS or MacOS, run startClassifai.sh instead.

./startClassifai.sh

During the process, if ran into error where JAVA_HOME is not found, refer to the first step 1. Installation of Java on how to setup JAVA_HOME environment path.

Note: Running the first build downloads dependencies required by the tool. It takes 3–10 minutes to finish the build.

Towards the end of the build, Classifai will be initiated on the local machine and you can start your data annotation journey.

Read more about Classifai from the references below

• Classifai Readme: https://github.com/CertifaiAI/classifai
• Documentation: https://docs.classifai.ai/

Chia Wei (codenamewei) is a data science lead with 8 years experiences working actively on Big Data, Machine Learning, and Deep Learning.

She is also an Open Source Contributor towards data science tools such as Classifai.

Connect on LinkedIn or Github.