Main repository for the Noise Extraction and User Profile Linking project (NOEL).
This repository contains the code used for a web application based on a research project [1] regarding residual noise extraction and smartphone linking, achieved by computing the correlations between each pair of picture noises, which in turn allows for clustering of the given pictures.
Docker CE and docker-compose are both required for this setup. A makefile is provided to control docker-compose services.
To ensure a correct startup of all the provided services, one needs to configure the following files.
Key and Certificate files are required for SSL setup. These can be generated through OpenSSL (openssl req -newkey rsa:2048 -nodes -keyout key.pem -x509 -days 365 -out certificate.pem) or provided through a more sofisticated setup by integrating with certbot. Certbot integration is not currently supported, but should be easy to implement through certbot/certbot docker image and edits to the nginx configuration.
cp docker/common.env.example docker/common.envFill in the required values (refer to the Standalone setup for reference).
Additionally, generate/type a random string as a value for DJANGO_SECRET, since it will be used as a base for Django encryption and hashing.
Django debug mode can be turned off by setting DJANGO_DEBUG to False.
A basic nginx config file is provided. Additional setup may be required depending on the context.
To simply launch the web app by itself, copy the provided .env.example file to .env, and fill in the required data.
# Skips certain steps during session processing; useful for debug
SESSION_SKIP=
# API key for linked Facebook App
FB_API_KEY=
# Redis backend url
REDIS_URL=redis://localhost
# Postgres config
POSTGRES_URL=localhost
POSTGRES_DB=postgres
POSTGRES_USER=postgres
POSTGRES_PASS=postgressNote: in order for the application to function properly, it needs to be served over SSL, since Facebook API only accepts https requests.
To make the provided docker-compose services executable, NOEL's Docker image must be built. This can be achieved by running
docker build . -f docker/Dockerfile -t <desired-tag>
# or
make docker-buildNote: to obtain the BM3D executable (required for both standalone setup AND Docker build, follow the instructions provided in the submodule's repository.