This repository contains both the app-frontend and app-service for the URL phishing detection application. In the app the user can enter a URL after which the user is asked whether he/she thinks the URL is phishing or not. Afterwards, the app will return the result of the model that is trained to detect phishing.
All instructions assume that you are currently in the working app directory. To run the app, simply build the container using:
docker build -t <tag> .
docker run -p 5000:5000 <tag>
You can now access the app through your browser. Please note that the app alone does not do anything, as the required backbone with a model-service is not running.
After accessing the app, you should see the following screen (the version follows from lib-version):
After entering a URL in the "Enter URL..." field, press the "Submit URL" to get the model prediction (please note that this will return "Unknown" if the model-service is not running). Before the model prediction is presented, the user is asked on their own prediction whether the URL is phishing or not. This screen looks as follows:
