Update app.py

app.py

@@ -1,3 +1,4 @@
+import os
 import gradio as gr
 import pandas as pd
 import google.generativeai as genai
@@ -9,17 +10,16 @@ from reportlab.lib.styles import getSampleStyleSheet, ParagraphStyle
 import plotly.express as px
 import plotly.graph_objects as go
 import tempfile
-import os
 from datetime import datetime
-from dotenv import load_dotenv
-
-
-# Load environment variables
-load_dotenv()
+import numpy as np
+from xgboost import XGBRegressor
 
 # Configure Gemini API
 GEMINI_API_KEY = os.getenv("gemini_api")
 
+if not GEMINI_API_KEY:
+    raise ValueError("GEMINI_API_KEY environment variable not found")
+
 genai.configure(api_key=GEMINI_API_KEY)
 generation_config = {
     "temperature": 1,
@@ -28,14 +28,61 @@ generation_config = {
     "max_output_tokens": 8192,
 }
 
-model = genai.GenerativeModel(
-    model_name="gemini-pro",
-    generation_config=generation_config,
-)
-
+model = genai.GenerativeModel("gemini-pro", generation_config=generation_config)
 chat_model = genai.GenerativeModel("gemini-pro")
 
-# Enhanced CSS for better styling
+# Create and save a simple model on first run
+def create_initial_model():
+    n_samples = 1000
+    np.random.seed(42)
+    
+    data = {
+        'weight (kilograms)': np.random.uniform(100, 10000, n_samples),
+        'line item value': np.random.uniform(1000, 1000000, n_samples),
+        'cost per kilogram': np.random.uniform(1, 500, n_samples),
+        'shipment mode_Air Charter_weight': np.zeros(n_samples),
+        'shipment mode_Ocean_weight': np.zeros(n_samples),
+        'shipment mode_Truck_weight': np.zeros(n_samples),
+        'shipment mode_Air Charter_line_item_value': np.zeros(n_samples),
+        'shipment mode_Ocean_line_item_value': np.zeros(n_samples),
+        'shipment mode_Truck_line_item_value': np.zeros(n_samples)
+    }
+    
+    modes = ['Air', 'Ocean', 'Truck']
+    for i in range(n_samples):
+        mode = np.random.choice(modes)
+        if mode == 'Air':
+            data['shipment mode_Air Charter_weight'][i] = data['weight (kilograms)'][i]
+            data['shipment mode_Air Charter_line_item_value'][i] = data['line item value'][i]
+        elif mode == 'Ocean':
+            data['shipment mode_Ocean_weight'][i] = data['weight (kilograms)'][i]
+            data['shipment mode_Ocean_line_item_value'][i] = data['line item value'][i]
+        else:
+            data['shipment mode_Truck_weight'][i] = data['weight (kilograms)'][i]
+            data['shipment mode_Truck_line_item_value'][i] = data['line item value'][i]
+    
+    df = pd.DataFrame(data)
+    base_cost = (df['weight (kilograms)'] * df['cost per kilogram'] * 0.8 + 
+                 df['line item value'] * 0.02)
+    
+    air_multiplier = 1.5
+    ocean_multiplier = 0.8
+    truck_multiplier = 1.0
+
+    freight_cost = (
+        base_cost * (air_multiplier * (df['shipment mode_Air Charter_weight'] > 0) +
+                    ocean_multiplier * (df['shipment mode_Ocean_weight'] > 0) +
+                    truck_multiplier * (df['shipment mode_Truck_weight'] > 0))
+    )
+    
+    freight_cost = freight_cost + np.random.normal(0, freight_cost * 0.1)
+    
+    model = XGBRegressor(n_estimators=100, learning_rate=0.1, max_depth=5, random_state=42)
+    model.fit(df, freight_cost)
+    
+    return model
+
+# Enhanced CSS styling
 CUSTOM_CSS = '''
 .gradio-container {
     max-width: 1200px !important;
@@ -175,45 +222,6 @@ input:focus, select:focus, textarea:focus {
     border-top: 1px solid #404040 !important;
     color: #888888 !important;
 }
-
-.tabs {
-    background: #2d2d2d !important;
-    border-radius: 10px !important;
-    padding: 10px !important;
-}
-
-.tab-selected {
-    background: #3498DB !important;
-    color: white !important;
-}
-
-.gr-box {
-    background: #2d2d2d !important;
-    border: 1px solid #404040 !important;
-}
-
-.gr-text-input {
-    background: #363636 !important;
-    color: #ffffff !important;
-}
-
-.gr-checkbox {
-    border-color: #404040 !important;
-}
-
-.gr-checkbox:checked {
-    background-color: #3498DB !important;
-}
-
-.gr-button-primary {
-    background: #3498DB !important;
-    color: white !important;
-}
-
-.gr-button-secondary {
-    background: #404040 !important;
-    color: white !important;
-}
 '''
 
 class SupplyChainState:
@@ -224,13 +232,11 @@ class SupplyChainState:
         self.chat_history = []
         self.analysis_results = {}
         self.freight_predictions = []
-
-        # Load the XGBoost model
-        self.model_path = "optimized_xgboost_model.pkl"
+        
         try:
-            self.freight_model = joblib.load(self.model_path)
+            self.freight_model = create_initial_model()
         except Exception as e:
-            print(f"Warning: Could not load freight prediction model from {self.model_path}: {e}")
+            print(f"Warning: Could not create freight prediction model: {e}")
             self.freight_model = None
 
 def process_uploaded_data(state, sales_file, supplier_file, text_data):
@@ -267,7 +273,6 @@ def perform_demand_forecasting(state):
         response = model.generate_content(prompt)
         analysis_text = response.text
         
-        # Create visualization
         fig = px.line(state.sales_df, title='Historical Sales Data and Forecast')
         fig.update_layout(
             template='plotly_dark',
@@ -307,7 +312,6 @@ def perform_risk_assessment(state):
         response = model.generate_content(prompt)
         analysis_text = response.text
 
-        # Create risk visualization
         fig = px.scatter(state.supplier_df, title='Supplier Risk Assessment')
         fig.update_layout(
             template='plotly_dark',
@@ -324,10 +328,35 @@ def perform_risk_assessment(state):
     except Exception as e:
         return f"❌ Error in risk assessment: {str(e)}", None, "Assessment failed"
 
+def predict_freight_cost(state, weight, line_item_value, cost_per_kg,
+                       shipment_mode, air_charter_weight, ocean_weight, truck_weight,
+                       air_charter_value, ocean_value, truck_value):
+    """Predict freight cost using the model"""
+    if state.freight_model is None:
+        return "Error: Freight prediction model not loaded"
+        
+    try:
+        features = {
+            'weight (kilograms)': weight,
+            'line item value': line_item_value,
+            'cost per kilogram': cost_per_kg,
+            'shipment mode_Air Charter_weight': air_charter_weight if "Air" in shipment_mode else 0,
+            'shipment mode_Ocean_weight': ocean_weight if "Ocean" in shipment_mode else 0,
+            'shipment mode_Truck_weight': truck_weight if "Truck" in shipment_mode else 0,
+            'shipment mode_Air Charter_line_item_value': air_charter_value if "Air" in shipment_mode else 0,
+            'shipment mode_Ocean_line_item_value': ocean_value if "Ocean" in shipment_mode else 0,
+            'shipment mode_Truck_line_item_value': truck_value if "Truck" in shipment_mode else 0
+        }
+        input_data = pd.DataFrame([features])
+        
+        prediction = state.freight_model.predict(input_data)
+        return round(float(prediction[0]), 2)
+    except Exception as e:
+        return f"Error making prediction: {str(e)}"
+
 def chat_with_navigator(state, message):
-    """Handle chat interactions with the SupplyChainAI Navigator"""
+    """Handle chat interactions"""
     try:
-        # Prepare context from available data
         context = "Available data and analysis:\n"
         if state.sales_df is not None:
             context += f"- Sales data with {len(state.sales_df)} records\n"
@@ -338,7 +367,6 @@ def chat_with_navigator(state, message):
         if state.freight_predictions:
             context += f"- Recent freight predictions: {state.freight_predictions[-5:]}\n"
         
-        # Add analysis results
         if state.analysis_results:
             context += "\nRecent analysis results:\n"
             for analysis_type, results in state.analysis_results.items():
@@ -359,38 +387,12 @@ def chat_with_navigator(state, message):
 
         response = chat_model.generate_content(prompt)
         
-        state.chat_history.append(("user", message))
-        state.chat_history.append(("assistant", response.text))
+        state.chat_history.append({"role": "user", "content": message})
+        state.chat_history.append({"role": "assistant", "content": response.text})
         
         return state.chat_history
     except Exception as e:
-        return [(msg_type, msg) for msg_type, msg in state.chat_history] + [("assistant", f"Error: {str(e)}")]
-
-def predict_freight_cost(state, weight, line_item_value, cost_per_kg,
-                       shipment_mode, air_charter_weight, ocean_weight, truck_weight,
-                       air_charter_value, ocean_value, truck_value):
-    """Predict freight cost using the loaded model"""
-    if state.freight_model is None:
-        return "Error: Freight prediction model not loaded"
-        
-    try:
-        features = {
-            'weight (kilograms)': weight,
-            'line item value': line_item_value,
-            'cost per kilogram': cost_per_kg,
-            'shipment mode_Air Charter_weight': air_charter_weight if "Air" in shipment_mode else 0,
-            'shipment mode_Ocean_weight': ocean_weight if "Ocean" in shipment_mode else 0,
-            'shipment mode_Truck_weight': truck_weight if "Truck" in shipment_mode else 0,
-            'shipment mode_Air Charter_line_item_value': air_charter_value if "Air" in shipment_mode else 0,
-            'shipment mode_Ocean_line_item_value': ocean_value if "Ocean" in shipment_mode else 0,
-            'shipment mode_Truck_line_item_value': truck_value if "Truck" in shipment_mode else 0
-        }
-        input_data = pd.DataFrame([features])
-        
-        prediction = state.freight_model.predict(input_data)
-        return round(float(prediction[0]), 2)
-    except Exception as e:
-        return f"Error making prediction: {str(e)}"
+        return [{"role": "assistant", "content": f"Error: {str(e)}"}]
 
 def generate_pdf_report(state, analysis_options):
     """Generate PDF report with analysis results"""
@@ -402,7 +404,6 @@ def generate_pdf_report(state, analysis_options):
         styles = getSampleStyleSheet()
         story = []
         
-        # Enhanced title style
         title_style = ParagraphStyle(
             'CustomTitle',
             parent=styles['Heading1'],
@@ -411,22 +412,18 @@ def generate_pdf_report(state, analysis_options):
             textColor=colors.HexColor('#2C3E50')
         )
         
-        # Add title
         story.append(Paragraph("SupplyChainAI Navigator Report", title_style))
         story.append(Spacer(1, 12))
         
-        # Add timestamp
         timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
         story.append(Paragraph(f"Generated on: {timestamp}", styles['Normal']))
         story.append(Spacer(1, 20))
         
-        # Add executive summary
         story.append(Paragraph("Executive Summary", styles['Heading2']))
         summary_text = "This report provides a comprehensive analysis of supply chain data, including demand forecasting, risk assessment, and optimization recommendations."
         story.append(Paragraph(summary_text, styles['Normal']))
         story.append(Spacer(1, 20))
         
-        # Add analysis results
         if state.analysis_results:
             for analysis_type, results in state.analysis_results.items():
                 if analysis_type in analysis_options:
@@ -442,7 +439,6 @@ def generate_pdf_report(state, analysis_options):
                     
                     story.append(Spacer(1, 20))
         
-        # Add freight predictions if available
         if state.freight_predictions:
             story.append(Paragraph("Recent Freight Cost Predictions", styles['Heading2']))
             story.append(Spacer(1, 12))
@@ -468,7 +464,6 @@ def generate_pdf_report(state, analysis_options):
             story.append(table)
             story.append(Spacer(1, 20))
         
-        # Build PDF
         doc.build(story)
         return pdf_path
     except Exception as e:
@@ -481,7 +476,6 @@ def run_analyses(state, choices, sales_file, supplier_file, text_data):
     figures = []
     status_messages = []
 
-    # Process data first
     process_status = process_uploaded_data(state, sales_file, supplier_file, text_data)
     if "Error" in process_status:
         return process_status, None, process_status
@@ -505,7 +499,6 @@ def run_analyses(state, choices, sales_file, supplier_file, text_data):
 
     combined_results = "\n\n".join(results)
     combined_status = "\n".join(status_messages)
-    
     final_figure = figures[-1] if figures else None
     
     return combined_results, final_figure, combined_status
@@ -528,7 +521,6 @@ def create_interface():
                 gr.Markdown("# 🚢 SupplyChainAI Navigator", elem_classes="app-title")
                 gr.Markdown("### Intelligent Supply Chain Analysis & Optimization", elem_classes="app-subtitle")
                 gr.Markdown("An AI-powered platform for comprehensive supply chain analytics", elem_classes="app-description")
-                gr.Markdown("Created by Aditya Ratan", elem_classes="creator-info")
         
         # Main Content Tabs
         with gr.Tabs() as tabs:
@@ -686,7 +678,8 @@ def create_interface():
                 chatbot = gr.Chatbot(
                     label="Chat History",
                     elem_classes="chat-container",
-                    height=400
+                    height=400,
+                    type="messages"
                 )
                 with gr.Row():
                     msg = gr.Textbox(
@@ -724,7 +717,6 @@ def create_interface():
                 truck_inputs: gr.update(visible=mode=="🚛 Truck")
             }
         
-        # Connect all components
         upload_button.click(
             fn=lambda *args: process_uploaded_data(state, *args),
             inputs=[sales_data_upload, supplier_data_upload, text_input_area],
@@ -770,11 +762,12 @@ def create_interface():
     
     return demo
 
+
+# Update the launch parameters in __main__:
 if __name__ == "__main__":
     demo = create_interface()
     demo.launch(
-        share=True,
+        server_name="0.0.0.0",  # Add this line
+        server_port=7860,       # Add this line
         debug=True
-    )
-        
-        # Enhanced title
+    )