MedsynBackend

This Repository contains the backend Server for the generative AI extension of the OHIF viewer. See https://github.com/TomWartm/Viewers for help.

Model Weights

For model weights see in google drive: Generative AI in Radiology > MedSyn Model Parameter > MedSyn or in https://github.com/batmanlab/MedSyn

Run

1. Clone this repository (on a machine with large GPU RAM)
   • git clone https://github.com/TomWartm/MedsynBackend
2. Navigate to the cloned project's directory
3. Install required python packages conda env create --file environment.yml
4. Actiave environment conda activate medsyn-3-8
5. Navigate to src folder
6. Run flask server python app.py

Network configurations

• This is deplloyed using nginx and flask.
• To see the status of the nginx, sudo systemctl status nginx and to see the status of the flask app sudo systemctl status flask_app
• To modify the flask app configurations go to /etc/systemd/system/flask_app.service
• To run the gunicorn setup not through flask and nginx, you should first stop the fllask_app sudo systemctl stop flask_app then go to ~/MedsynBackend/src/, activate the medsyn-3-8 conda environment if it's not activated then run this command /home/exouser/miniconda3/envs/medsyn-3-8/bin/gunicorn --workers 1 --bind 0.0.0.0:5000 app:app