Metarank Demo

This is a demo outlining how you can use Metarank in real-world scenarios. The code showcases how you can utilize deployed Metarank instance in your web application and covers ranking results, sending feedback and analyzing response.

Prerequisites

The demo project utilizes Node.js for the backend and ReactJS for the frontend, so you must have Node.js installed in order to run the project. We also use Yarn for package management (npm i -g yarn for quick install).

Running using Docker

You can easily run the project with docker-compose:

• run docker-compose build to build the images
• run METARANK_URL=SOME-URL MODEL_NAME=SOME-MODEL docker-compose up to run the images

You can access the frontend application at localhost:3000 when both containers are running.

Running using yarn workspaces

Both frontend and api projects are wrapped in yarn workspace, so you can run both projects using one yarn command.

• run yarn in the project folder to install all dependencies
• run npm run start to start both projects simultaniously

You still need to provide the METARANK_URL environment variable, e.g. METARANK_URL=http://localhost:8080 npm run start

Running projects separately

install the packages in the frontend and server folders

• cd frontend && npm i
• cd server && npm i

run the backend

• use METARANK_URL environment variable to provide the URL of your Metarank installation in the format http://localhost:8080
• use MODEL_NAME environment variable to specify the name of the Metarank model from your configuration file. By default it's xgboost as in the Ranklens Demo
• use PORT environment vvaraible to provide the port on which API will run. By default port 3001 is used
• cd server && npm run start to run application

run the frontend

• cd frontend && npm run start. By default the frontend will run on port 3000