Prepared by: Vikas Neriyanuru

For this lab submission, I built a FastAPI application that predicts the wine cultivar from 13 chemical measurements using the Wine Recognition dataset from scikit-learn.

• Replaced the original dataset with the Wine Recognition dataset
• Replaced the original model with a StandardScaler + LogisticRegression pipeline
• Added a /model-info endpoint to show the saved model metadata and evaluation metrics
• Added a /sample-payloads endpoint to make testing easier
• Updated the prediction response to include both the predicted class name and class probabilities
• Added automated API tests using pytest
• Wrote a new README with my own setup, rerun steps, and sample API call

FastAPI_Labs/
├── model/
├── src/
│   ├── __init__.py
│   ├── data.py
│   ├── main.py
│   ├── predict.py
│   └── train.py
├── tests/
│   └── test_api.py
├── README.md
└── requirements.txt

1. Move into the lab folder:

cd Labs/API_Labs/FastAPI_Labs

2. Create and activate a virtual environment:

python3 -m venv .venv
source .venv/bin/activate

3. Install the required packages:

pip install -r requirements.txt

4. Train the model and save the artifacts:

python3 -m src.train

5. Start the FastAPI application:

uvicorn src.main:app --reload

6. Open the Swagger UI in your browser:

http://127.0.0.1:8000/docs

• GET / returns a basic status message and quick links
• GET /health checks if the API is running and whether the model artifact exists
• GET /model-info returns model metadata, feature names, class labels, and evaluation metrics
• GET /sample-payloads returns sample JSON payloads for each wine class
• POST /predict predicts the wine class from the request body

curl -X POST "http://127.0.0.1:8000/predict" \
  -H "Content-Type: application/json" \
  -d '{
    "alcohol": 14.23,
    "malic_acid": 1.71,
    "ash": 2.43,
    "alcalinity_of_ash": 15.6,
    "magnesium": 127.0,
    "total_phenols": 2.8,
    "flavanoids": 3.06,
    "nonflavanoid_phenols": 0.28,
    "proanthocyanins": 2.29,
    "color_intensity": 5.64,
    "hue": 1.04,
    "od280_od315_of_diluted_wines": 3.92,
    "proline": 1065.0
  }'

{
  "predicted_class_id": 0,
  "predicted_class_name": "class_0",
  "class_probabilities": {
    "class_0": 0.999591,
    "class_1": 0.000377,
    "class_2": 0.000031
  }
}

After running python3 -m src.train, the saved metadata reported:

• Test accuracy: 1.0
• Test macro F1 score: 1.0
• Training rows: 133
• Test rows: 45

python3 -m pytest tests/test_api.py -v

• This lab is from the API_Labs folder, which is different from the folders I used in earlier submissions
• The dataset and model are different from the original repo version
• The README is fully customized for my version of the lab