import joblib
import pandas as pd
import numpy as np
import gradio as gr
from tensorflow.keras.models import load_model
from sklearn.preprocessing import MinMaxScaler
import tensorflow as tf

mm = joblib.load("/home/user/app/mm_X_scaler.pkl")

mm_y = joblib.load("/home/user/app/mm_y_scaler.pkl")

# Load the trained model
model = load_model('/home/user/app/model_superior_Superior.h5')  # Update with your model path

def preprocess_input(csv_file):
    """
    Preprocess the input CSV file containing vibrational data
    """
    try:
        # Read the CSV file
        df = pd.read_csv(csv_file.name,index_col=0)
        

        
        # Preprocess the data
        data = df.values
        data_scaled = mm.fit_transform(data)
        data_reshaped = data_scaled.reshape(data_scaled.shape[0], 1, data_scaled.shape[1])
        
        # Make predictions
        predictions = model.predict(data_reshaped)
        predictions = mm_y.inverse_transform(predictions)
        
        # Create results DataFrame
        results_df = pd.DataFrame({
            'Timestamp': df['AssignedDate'] if 'AssignedDate' in df.columns else range(len(predictions)),
            'Prediction': predictions.flatten(),
            'Maintenance Type': ['Predictive' if p >= 0.5 else 'Preventive' for p in predictions.flatten()]
        })
        
        # Save results to CSV
        results_path = "predictions.csv"
        results_df.to_csv(results_path, index=True)
        
        # Calculate summary statistics
        total_samples = len(predictions)
        predictive_count = sum(predictions.flatten() >= 0.5)
        preventive_count = total_samples - predictive_count
        
        summary = f"""
        Analysis Results:
        ----------------
        Total Samples: {total_samples}
        Predictive Maintenance Required: {predictive_count} ({predictive_count/total_samples*100:.1f}%)
        Preventive Maintenance Required: {preventive_count} ({preventive_count/total_samples*100:.1f}%)
        
        Detailed results have been saved to 'predictions.csv'
        """
        
        return summary

    except Exception as e:
        return f"Error processing file: {str(e)}"

# Create Gradio interface
iface = gr.Interface(
    fn=preprocess_input,
    inputs=[
        gr.File(
            label="Upload Vibrational Data CSV",
            type="filepath",
            file_types=[".csv"]
        )
    ],
    outputs=gr.Textbox(
        label="Prediction Results",
        lines=10
    ),
    title="CNC Machine Maintenance Predictor",
    description="""
    Upload a CSV file containing CNC machine vibrational data to predict maintenance types.
    The CSV should contain the following columns: PC1, PC2, PC3, PC4, PC5, PC6, PC7, PC8, PC9, PC441, PC442.
    The model will classify each record as either Predictive or Preventive maintenance.
    """,
    examples=[
        ["/home/user/app/X.csv"]  # Add example data file if available
    ]
)

# Launch the interface
if __name__ == "__main__":
    iface.launch(share=True)