This document explains how to deploy AutoClassWeb as a local web application.

Due to security reasons, its usage should be strictly restricted to a local use only (lab or institute should be fine). This app shouldn't be accessible to anyone from the internet.

AutoClassWeb relies on Docker.

Depending of your operating system, follow these instructions to install:

• Docker for Windows
• Docker for Mac
• Docker for Linux/Ubuntu

Have also a look to Bret Fisher's video if you need more details and step-by-step instructions to install Docker.

For now on, be sure to have docker running in the background.

Download the autoclassweb-app-master.zip file, extract the content of this file and copy the obtained folder (autoclassweb-app-master) somewhere on your computer.

Open the autoclassweb-app-master folder. According to your operating system, run the proper installation script.

For Windows, double-click on the install_for_windows.bat script. Wait a few minutes.

For Mac or Linux/Ubuntu, open a terminal, run the following command and wait a few minutes:

$ bash install_for_mac_linux.sh

Remark: app parameters are in the file autoclassweb.cfg in the config directory. Defaults should be OK.

For Windows, double-click on the file autoclassweb-app.bat. Accept security warnings if any.

For Mac or Linux/Ubuntu, open a terminal and run the following command:

$ bash autoclassweb-app.sh

Remark: don't worry if you encounter this error message Error response from daemon: No such container: autoclassweb. This means the app in not running yet. This is fine.

Once the app is started, open a web browser to the following address:

http://127.0.0.1:5000

The directory demo in autoclassweb-app-master contains some files for a quick demo:

• Upload the file demo_real_location.tsv as Real Location.
• Upload the file demo_real_scalar.tsv as Real Scalar.
• Keep all other parameters as defaults.

The classification should take a couple of seconds to perform. Refresh the page often to get the results.

You should expect 3 classes with 100 members each, and with the means of variables x and y being approximately {3, 4}, {4, 0} and {7, 5} (not necessarily in this order).

While idle, the app will use very little resources. If you want to stop the app anyway, run the following command in a terminal:

$ docker stop autoclassweb

If you move the install folder autoclassweb-app-master somewhere else, you have to re-run the installation script.

The installation procedure is inspired from START-R developed by Thomas Denecker.