This folder contains all the code for running the front end in isolation, but not for training machine learning models.
If you only want to run the tool, but not develop it further, then everything you need is in this folder.
In order to run the tool you need to also have installed Apache Tika (https://tika.apache.org/) and run it (java -jar tika[VERSION].jar.
You need to install the requirements in requirements.txt.
pip install -r requirements.txt
Then you can run the tool using Python from the command line:
python application.py
It is recommended to open this folder in an IDE such as Pycharm, since there are many Python files that interact and import one another.
The tool is architected as two separate Docker containers: the front end, which does most of the logic, and a Tika server which is Java-based and is for PDF extraction.
They can be run as Docker Compose using the docker-compose.yml, or individually.
If you run both containers individually, it is recommended to run the script in build_deploy.sh, or you can run more simply the following commands:
docker build -t protocols-front .
docker run protocols-front
To run the Tika container, cd into tika and run
docker build -t protocols-tika .
docker run protocols-tika
You will need to set environment variables on the front end container COMMIT_ID and TIKA_SERVER_ENDPOINT so that the front end knows where to find Tika.
If you have any problems, check the ports that both servers are running on and check that the ports are all correctly exposed and mapped.