Dataset Preprocesser Emission Reduction

Features

Changelog

• Added more metrics as goal metrics
• Added images datasets
• Add method for reducing dataset (balanced vs uniform)
• Added notebook containing analysis of the tool on actual data
• Uploaded to github and dockerized

To do

• Add more TS datasets
• Clean UI
• Add a section that shows user what are the environmental impacts when using the tool (carbon emission reduction, etc)

Run the tool

Docker and Docker-Compose (Recommended)

• Install Docker
• From the root of the project run the following

First install everything with

docker-compose --build

Now to run

docker-compose up -d

The first time the database needs to be set, run in order the following commands

curl -X 'GET' 'http://localhost:5000/api/load_experiments_on_db?reset_db=true' -H 'accept: application/json'

curl -X 'GET' 'http://localhost:5000/api/load_completeness_curves?load_only_missing=false' -H 'accept: application/json'

curl -X 'GET' 'http://localhost:5000/api/train_regressors' -H 'accept: application/json'

You can now visit the interface in https://localhost:8080

To finish the process after

docker-compose down

Installing everything manually

• Install the following requirements
  • For the backend
    • python 3.10
    • From the backend/ directory, run pip install -r requirements.txt
  • For the frontend
    • install node 16
    • From the 'frontend/' directory, run npm install and npm run build

Now to run the backend

uvicorn main:app --port 3000 

And the frontend

npm run serve

The first time the database needs to be set, run in order the following commands

curl -X 'GET' 'http://localhost:5000/api/load_experiments_on_db?reset_db=true' -H 'accept: application/json'

curl -X 'GET' 'http://localhost:5000/api/load_completeness_curves?load_only_missing=false' -H 'accept: application/json'

curl -X 'GET' 'http://localhost:5000/api/train_regressors' -H 'accept: application/json'

You can now visit the interface in https://localhost:8080