app.py

# Import the libraries
import os
import uuid
import joblib
import json
import time
import gradio as gr
import pandas as pd

from huggingface_hub import HfApi
from huggingface_hub import CommitScheduler
from pathlib import Path
from sklearn.model_selection import train_test_split
from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder
from datetime import datetime

# Load the Hugging Face API token
hf_token = os.getenv("HF_TOKEN")

# Initialize Hugging Face API
api = HfApi()
# Run the training script placed in the same directory as app.py
# The training script will train and persist a linear regression
# model with the filename 'model.joblib'
os.system("python train.py")
# Load the freshly trained model from disk 
model = joblib.load('model.joblib')
# Prepare the logging functionality
log_file = Path("logs/") / f"data_{uuid.uuid4()}.json"
log_folder = log_file.parent

scheduler = CommitScheduler(
    repo_id="insurance-charge-mlops-logs",  # provide a name "insurance-charge-mlops-logs" for the repo_id
    repo_type="dataset",
    folder_path=log_folder,
    path_in_repo="data",
    every=2,
    token=hf_token
)

# Define the predict function which will take features, convert to dataframe and make predictions using the saved model
# the functions runs when 'Submit' is clicked or when a API request is made
def insurance_charge_predictor(age, bmi, children, sex, smoker, region):
    input_data = pd.DataFrame({
        'age': [age],
        'bmi': [bmi],
        'children': [children],
        'sex': [sex],
        'smoker': [smoker],
        'region': [region]
    })

    prediction = model.predict(input_data)
    # While the prediction is made, log both the inputs and outputs to a  log file
    # While writing to the log file, ensure that the commit scheduler is locked to avoid parallel
    # access
    with scheduler.lock:
        with log_file.open("a") as f:
            f.write(json.dumps(
                {
                    'age': age,
                    'bmi': bmi,
                    'children': children,
                    'sex': sex,
                    'smoker': smoker,
                    'region': region,
                    'prediction': prediction[0]
                }
            ))
            f.write("\n")

    return prediction[0]

# Set up UI components for input and output
# Create the gradio interface, make title "HealthyLife Insurance Charge Prediction"
with gr.Blocks() as demo:
    gr.Markdown("HealthyLife Insurance Charge Prediction")
    status = gr.State(["none"])
    with gr.Row():
        with gr.Column():
            age = gr.Slider(minimum=18, maximum=100, value=25, label="Age", step=1)
            bmi = gr.Slider(minimum=10, maximum=50, value=25, label="BMI")
            children = gr.Slider(minimum=0, maximum=10, value=0, label="Children", step=1)
            sex = gr.Radio(choices=["male", "female"], label="Sex")
            smoker = gr.Radio(choices=["yes", "no"], label="Smoker")
            region = gr.Dropdown(choices=["northeast", "northwest", "southeast", "southwest"], label="Region")

        with gr.Column():
            output = gr.Number(label="Predicted Insurance Charge")
            predict_btn = gr.Button("Submit")

            predict_btn.click(insurance_charge_predictor, inputs=[age, bmi, children, sex, smoker, region], outputs=output)

# Launch with a load balancer
demo.queue()
demo.launch(share=True, show_error=True)