import pandas as pd
import pickle
import skops.io as sio


# Load Model
model = sio.load("cpu_price_model.skops")


# Load Feature Names
with open("feature_names.pkl", "rb") as f:
    feature_names = pickle.load(f)


# Prediction Function
def predict_cpu_price(features: dict) -> float:
    """
    Predict CPU price from input features.

    Args:
        features (dict): Dictionary containing all required features.

    Returns:
        float: Predicted CPU price.
    """

    df = pd.DataFrame([features])

    missing_features = set(feature_names) - set(df.columns)

   if missing_features:
      raise ValueError(
           f"Missing features: {missing_features}"
       )

    df = df[feature_names]
    # Make prediction
    prediction = model.predict(df)

    return float(prediction[0])


# Example Usage
if __name__ == "__main__":

    sample_input = {
        "tdp": 125,
        "cores": 8,
        "logicals": 16,
        "cpuCount": 1,
        "rank": 2500,
        "samples": 120,
        "extracted_ghz": 3.6,
        "speed_ghz": 3.6,
        "turbo_ghz": 5.0,
        "cost_per_rank_point": 0.12,
        "cost_per_core": 45.5,
        "brand_encoded": 0,
        "category_final_encoded": 1,
        "socket_final_encoded": 3
    }

    predicted_price = predict_cpu_price(sample_input)

    print(f"Predicted CPU Price: ${predicted_price:.2f}")