Update app.py

app.py

@@ -1,42 +1,63 @@
 import gradio as gr
-import numpy as np
 import pandas as pd
-from tensorflow.keras.models import load_model
-import tensorflow.keras.losses
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
 from sklearn.preprocessing import MinMaxScaler
+from tensorflow.keras.models import load_model
 
+# Load dataset
+df = pd.read_csv("TEMP_ANNUAL_SEASONAL_MEAN.csv")
 
-# Explicitly define the loss function
-custom_objects = {"mse": tensorflow.keras.losses.MeanSquaredError()}
-
-# Load the trained LSTM model with custom_objects
-model = load_model("climate_model.h5", custom_objects=custom_objects)
-  # Ensure this file is uploaded
+# Convert 'YEAR' to datetime and set as index
+df.columns = df.columns.str.upper()
+df['YEAR'] = pd.to_datetime(df['YEAR'], format='%Y')
+df.set_index('YEAR', inplace=True)
 
-# Load dataset to get scaler reference
-df = pd.read_csv("TEMP_ANNUAL_SEASONAL_MEAN.csv")
-df['ANNUAL'] = pd.to_numeric(df['ANNUAL'], errors='coerce')  # Ensure correct type
-data = df[['ANNUAL']].dropna()
+# Select relevant temperature column
+temp_col = 'ANNUAL'
+df[temp_col] = pd.to_numeric(df[temp_col], errors='coerce')  # Handle errors
+data = df[[temp_col]].dropna()
 
-# Normalize using MinMaxScaler
+# Normalize the data
 scaler = MinMaxScaler()
 data_scaled = scaler.fit_transform(data)
 
-# Function to predict future temperature
+# Load trained model
+model = load_model("climate_model.h5")
+
+# Prediction function
 def predict_temperature(year):
-    last_10_years = data_scaled[-10:].reshape(1, 10, 1)  # Last 10 years as input
-    prediction = model.predict(last_10_years)
-    predicted_temp = scaler.inverse_transform(prediction)[0][0]  # Convert back to original scale
-    return f"Predicted Temperature for {year}: {predicted_temp:.2f}°C"
+    year = int(year)
+    
+    # Generate input for LSTM
+    sequence_length = 10
+    last_seq = data_scaled[-sequence_length:].reshape(1, sequence_length, 1)
+
+    # Predict
+    prediction_scaled = model.predict(last_seq)
+    predicted_temp = scaler.inverse_transform(prediction_scaled)[0][0]
+
+    # Generate graph
+    fig, ax = plt.subplots(figsize=(10, 5))
+    ax.plot(df.index[-20:], df[temp_col][-20:], label="Actual", color='blue')
+    ax.axvline(pd.Timestamp(year, 1, 1), color='red', linestyle='dashed', label="Predicted Year")
+    ax.scatter(pd.Timestamp(year, 1, 1), predicted_temp, color='red', marker='o', label=f"Predicted: {predicted_temp:.2f}°C")
+    ax.set_xlabel("Year")
+    ax.set_ylabel("Temperature (°C)")
+    ax.set_title("Actual vs Predicted Annual Mean Temperature in India")
+    ax.legend()
+
+    return predicted_temp, fig  # Return numeric output and graph
 
 # Gradio Interface
-iface = gr.Interface(
-    fn=predict_temperature,
-    inputs=gr.Number(label="Enter Year"),
-    outputs=gr.Textbox(label="Predicted Temperature"),
-    title="Climate Prediction",
-    description="Enter a year to predict the average annual temperature."
-)
-
-# Launch Gradio App
-iface.launch()
+with gr.Blocks() as demo:
+    gr.Markdown("## Climate Change Temperature Prediction")
+    year_input = gr.Number(label="Enter Year (e.g., 2027)", value=2027)
+    predict_button = gr.Button("Predict")
+    output_temp = gr.Number(label="Predicted Temperature (°C)")
+    output_plot = gr.Plot()
+
+    predict_button.click(predict_temperature, inputs=[year_input], outputs=[output_temp, output_plot])
+
+demo.launch()