Scrounger is an open source content search application powered by Boon AI, a GUI-driven machine learning integration platform. The app is designed to let its users experience what it’s like to load their media assets and run an ML-powered search within a custom app, in under an hour. The purpose of the Scrounger application is to:
- Guide developers on the best practices of integrating ML-generated metadata into a proprietary or third party app using the Boon AI Python library.
- Enable business end-users to evaluate the accelerated value Boon AI's search capability can deliver to their organization.
- Serve as a starting point for a media storage and search app that can be extended into a proprietary app to meet unique business needs.
Scrounger consists of a single Docker container that can be configured using environment variables. It can be deployed anywhere a container can be run but we have included easy instructions for deploying in the two most common scenarios.
- Boon AI API Key with "Assets Read" permission - API Key Instructions
We suggest using Digital Ocean to get up and running with the least amount of friction. The deployment process tasks about 10 minutes and costs less than $20/month. Full instructions are at the link below.
Digital Ocean Deployment Instructions
The following instructions are intended for an on-prem deployment used for testing and evaluation purposes. This is not intended to be a production-grade installation; the server will only accept HTTP connections on port 80 and users are tracked in a sqlite database.
To start up scrounger run the following commands on a server that has docker installed and is exposed on your network.
docker pull boonai/scrounger
docker run -d -p 80:8080 -v /var/lib/scrounger:/applications/scrounger/django_backend/scrounger/sqlite -e BOONAI_API_KEY='<BOONAI_API_KEY>' boonai/scrounger
Scrounger will now be running at http://<hostname_or_ip_address>
Scrounger is open-sourced under the MIT license and we encourage anyone to fork this repo and use this as a base for building a proprietary web application that leverages Boon AI for improving search capabilities of unknown and unstructured data. Below is a description of the applications architecture and how to get started extending it.
+------------------------+
| |
| NGINX Server |
| |
+------------------------+
| |
| |
| |
v v
+-------------------------+ +---------------------------+
| | | |
| React Node Application | | Django uWSGI Application |
| (Frontend) | | (Backend) |
| | | |
+-------------------------+ +---------------------------+
|
v
+---------------------+
| |
| Relational Database |
| |
+---------------------+
Scrounger consists of 3 major components.
Nginx server exposed to the outside world. It is responsible for routing
requests to either the frontend or backend application based on the path. Any request beginning
with /api
or /admin
is routed to the Django backend and all other requests are routed to
the React frontend.
Frontend application responsible for rendering all pages that are sent to the client. It is a React app served by a Next.js server running on port 3000.
Backend application responsible for providing the API endpoints the frontend uses for fetching data and handling authentication and user administration. It is a Django/uWSGI application served by a gunicorn server on port 8000.
To get started extending any of these components view the README.md found in each of their respective folders for more information.