# Food Preference Supervised Classifier

This folder includes the implementation of the Food Preference Supervise Classifier project for the course PA026.

## How to run

This project can be run via `docker compose`. In order to run it, make sure you have [Docker](https://www.docker.com/products/docker-desktop/) installed on your local machine. After installing, run the followind command in the root of the project directory to run both frontend and backend services:

```
docker compose up
```

This should start both services. The frontend page is then available locally on http://localhost:5173.

## Training the classifier

The project already contains a pretrained model of the classifier, so the training can be skipped in order to run the project. However, in order to train the model, run the `create_classifier.py` script locally.  
The best way to ensure that the dependencies match is to create a new virtual environment via conda:

```
conda create --name myenv python=3.10
```

Then activate the environment and install the required dependencies:

```
conda activate myenv
pip install -r fastapi/requirements.txt
```

Then run the `create_classifier.py` script to train and save the classifier via autogluon:

```
python create_classifier.py
```