Energy Efficiency Investments Derisking Service backend(Pilot 8)
The Database.sql script creates the necessary tables in a PostgreSQL/PostGIS database located at NTUA's premises.
This folder includes scripts that split unstructured Excel files into CSV files, except for read_gpkg.py, which reads the State Land Dataset from a .gpkg file provided by REA, including geometrical information of housing in Riga. Other scripts like read_riga_dhs.py, read_envelope.py, read_energy_efficiency.py, and read_audits.py handle operations on various datasets and create tables in the corresponding database.
This folder contains a Dockerfile for the API and all necessary files for the backend of the application to be functional.
This folder contains all datasets used for the service. Some were provided by the partner REA, while others were created using data-manipulation scripts.
This folder contains files related to machine learning. pytorch-kfold creates a physics-informed neural network, ml_algorithm prepares final datasets for machine learning algorithms, and ml-for-backend creates shallow regression models to predict envelope segments and energy consumption of buildings with less accuracy than the physics-informed algorithms.
This folder contains a preprint of the paper derived from the machine learning algorithm developed for this service.
To run the API:
- Build the Docker image using the provided Dockerfile.
docker build -t rea-fastapi:2.0 .- Run the image to create the container of the api
docker run -d -p 8000:8000 -v ~/rea-uc/apisrc:/app/src --name rea-api-final-2 rea-fastapi:2.0To run the PostGIS:
docker compose up -d --build