src/predict.py

# src/predict.py

import os
import pickle
import pandas as pd
from model_training import train_model  # Your training function

# Path to the model file
MODEL_PATH = os.path.join(os.path.dirname(__file__), '../models/crop_model.pkl')

def load_model():
    """
    Loads the crop prediction model from disk.
    If the model file does not exist, trains a new model and saves it.
    """
    if os.path.exists(MODEL_PATH):
        # Load existing model
        with open(MODEL_PATH, 'rb') as f:
            model = pickle.load(f)
    else:
        # Train new model and save it
        print("Model not found. Training a new model...")
        model = train_model()  # Make sure your model_training.py has train_model() returning a fitted model
        os.makedirs(os.path.dirname(MODEL_PATH), exist_ok=True)
        with open(MODEL_PATH, 'wb') as f:
            pickle.dump(model, f)
        print("Model trained and saved.")
    return model

# Load model once when the app starts
model = load_model()

def predict_crop(input_features: pd.DataFrame):
    """
    Predicts the crop based on input features.
    Args:
        input_features: pd.DataFrame with columns ['N', 'P', 'K', 'temperature', 'humidity', 'ph', 'rainfall']
    Returns:
        List of predicted crop names
    """
    predictions = model.predict(input_features)
    return predictions.tolist()