Update app.py

app.py

@@ -2,6 +2,8 @@ import gradio as gr
 import requests
 import re
 import json
+import csv
+import io
 from urllib.parse import urlparse, urljoin
 import time
 import random
@@ -213,6 +215,8 @@ CONTACTS_DB = [
 def extract_domain(url):
     """Extract domain from URL"""
     try:
+        if not url.startswith(('http://', 'https://')):
+            url = 'https://' + url
         parsed = urlparse(url)
         domain = parsed.netloc.lower()
         # Remove www. if present
@@ -245,15 +249,104 @@ def find_contacts_by_website(website_url):
 def simulate_website_scraping(url):
     """Simulate scraping a website and finding contact information"""
     # Add some delay to simulate real scraping
-    time.sleep(random.uniform(1, 2))
+    time.sleep(random.uniform(0.5, 1))
     
     # Find matching contacts from our database
     contacts = find_contacts_by_website(url)
     
-    # Only return contacts if we found exact matches
-    # Don't return random contacts if no match found
     return contacts
 
+def parse_csv_file(file_content):
+    """Parse CSV file and extract website URLs"""
+    websites = []
+    try:
+        # Decode file content
+        content = file_content.decode('utf-8')
+        
+        # Parse CSV
+        csv_reader = csv.DictReader(io.StringIO(content))
+        
+        # Look for common website column names
+        website_columns = ['website', 'url', 'domain', 'site', 'web', 'homepage']
+        
+        for row in csv_reader:
+            # Find website column
+            website_url = None
+            for col_name in row.keys():
+                if col_name.lower() in website_columns:
+                    website_url = row[col_name]
+                    break
+            
+            if website_url and website_url.strip():
+                websites.append(website_url.strip())
+        
+        return websites
+    except Exception as e:
+        print(f"Error parsing CSV: {e}")
+        return []
+
+def search_csv_websites(csv_file, max_results=10):
+    """Search for contacts from websites listed in CSV file"""
+    if csv_file is None:
+        return "Please upload a CSV file", ""
+    
+    try:
+        # Parse CSV file
+        websites = parse_csv_file(csv_file)
+        
+        if not websites:
+            return "No websites found in CSV file. Please ensure your CSV has a column named 'website', 'url', or 'domain'.", ""
+        
+        all_contacts = []
+        processed_websites = []
+        
+        # Search each website
+        for website in websites[:20]:  # Limit to first 20 websites
+            contacts = simulate_website_scraping(website)
+            if contacts:
+                all_contacts.extend(contacts)
+                processed_websites.append(website)
+        
+        # Remove duplicates based on email
+        unique_contacts = []
+        seen_emails = set()
+        for contact in all_contacts:
+            if contact['email'] not in seen_emails:
+                unique_contacts.append(contact)
+                seen_emails.add(contact['email'])
+        
+        # Limit results
+        unique_contacts = unique_contacts[:max_results]
+        
+        if not unique_contacts:
+            return f"No contacts found for the {len(websites)} websites in the CSV file.", ""
+        
+        # Format results
+        results_text = f"CONTACT DISCOVERY REPORT\n"
+        results_text += f"Websites Processed: {len(processed_websites)}\n"
+        results_text += f"Total Websites in CSV: {len(websites)}\n"
+        results_text += f"Unique Contacts Found: {len(unique_contacts)}\n"
+        results_text += f"{'='*60}\n\n"
+        
+        for i, contact in enumerate(unique_contacts, 1):
+            results_text += f"CONTACT #{i}\n"
+            results_text += f"Name: {contact['first_name']} {contact['last_name']}\n"
+            results_text += f"Position: {contact['job_title']}\n"
+            results_text += f"Email: {contact['email']}\n"
+            results_text += f"Phone: {contact['phone']}\n"
+            results_text += f"Company: {contact['company']}\n"
+            results_text += f"Website: {contact['website']}\n\n"
+        
+        # Create CSV output
+        csv_output = "First Name,Last Name,Job Title,Email,Phone,Company,Website\n"
+        for contact in unique_contacts:
+            csv_output += f"{contact['first_name']},{contact['last_name']},{contact['job_title']},{contact['email']},{contact['phone']},{contact['company']},{contact['website']}\n"
+        
+        return results_text, csv_output
+        
+    except Exception as e:
+        return f"Error processing CSV file: {str(e)}", ""
+
 def search_website_contacts(website_url, max_results=10):
     """Main function to search for contacts on a website"""
     if not website_url:
@@ -274,30 +367,19 @@ def search_website_contacts(website_url, max_results=10):
         contacts = contacts[:max_results]
         
         # Format results
-        results_text = f"📊 **CONTACT INTELLIGENCE REPORT**\n"
-        results_text += f"🌐 **Website:** {website_url}\n"
-        results_text += f"📈 **Contacts Found:** {len(contacts)}\n"
+        results_text = f"CONTACT INTELLIGENCE REPORT\n"
+        results_text += f"Website: {website_url}\n"
+        results_text += f"Contacts Found: {len(contacts)}\n"
         results_text += f"{'='*60}\n\n"
         
-        contact_data = []
         for i, contact in enumerate(contacts, 1):
-            results_text += f"🏢 **CONTACT #{i}**\n"
-            results_text += f"├─ **First Name:** {contact['first_name']}\n"
-            results_text += f"├─ **Last Name:** {contact['last_name']}\n"
-            results_text += f"├─ **Position:** {contact['job_title']}\n"
-            results_text += f"├─ **Email:** {contact['email']}\n"
-            results_text += f"├─ **Phone:** {contact['phone']}\n"
-            results_text += f"└─ **Company:** {contact['company']}\n\n"
-            
-            # Prepare data for CSV/table format
-            contact_data.append([
-                contact['first_name'],
-                contact['last_name'],
-                contact['job_title'],
-                contact['email'],
-                contact['phone'],
-                contact['company']
-            ])
+            results_text += f"CONTACT #{i}\n"
+            results_text += f"First Name: {contact['first_name']}\n"
+            results_text += f"Last Name: {contact['last_name']}\n"
+            results_text += f"Position: {contact['job_title']}\n"
+            results_text += f"Email: {contact['email']}\n"
+            results_text += f"Phone: {contact['phone']}\n"
+            results_text += f"Company: {contact['company']}\n\n"
         
         # Create a simple table format for the second output
         table_text = "First Name,Last Name,Job Title,Email,Phone,Company\n"
@@ -314,9 +396,8 @@ def get_all_available_websites():
     websites = list(set([contact['website'] for contact in CONTACTS_DB]))
     return "\n".join(sorted(websites))
 
-# Custom CSS matching Kwekel Companies professional theme
+# Custom CSS
 custom_css = """
-/* Kwekel-inspired Corporate Theme */
 .gradio-container {
     background: linear-gradient(135deg, #f8f9fa 0%, #e9ecef 100%);
     font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
@@ -369,14 +450,6 @@ custom_css = """
     font-weight: 600;
 }
 
-.quick-search-grid {
-    display: grid;
-    grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
-    gap: 10px;
-    margin: 20px 0;
-}
-
-/* Button Styling */
 .primary-btn {
     background: linear-gradient(135deg, #1e40af 0%, #3b82f6 100%);
     color: white;
@@ -410,7 +483,6 @@ custom_css = """
     background: #f8fafc;
 }
 
-/* Input Fields */
 .custom-input {
     border: 2px solid #d1d5db;
     border-radius: 8px;
@@ -424,7 +496,6 @@ custom_css = """
     box-shadow: 0 0 0 3px rgba(59, 130, 246, 0.1);
 }
 
-/* Results Styling */
 .results-container {
     background: white;
     border: 1px solid #e5e7eb;
@@ -444,84 +515,128 @@ custom_css = """
 }
 """
 
-# Create Gradio interface with custom theme
-with gr.Blocks(css=custom_css, title="ZoomInfo Contact Discovery", theme=gr.themes.Base()) as app:
+# Create Gradio interface
+with gr.Blocks(css=custom_css, title="Contact Discovery Platform", theme=gr.themes.Base()) as app:
     gr.HTML("""
     <div class="main-header">
-        <h1>ZoomInfo Contact Discovery</h1>
-        <p>Professional Contact Discovery & Lead Generation Platform</p>
-        <p style="font-size: 0.95em; opacity: 0.8;">Advanced website analysis for corporate intelligence gathering</p>
+        <h1>Contact Discovery Platform</h1>
+        <p>Professional Contact Discovery & Lead Generation Tool</p>
+        <p style="font-size: 0.95em; opacity: 0.8;">Advanced website analysis for contact intelligence gathering</p>
     </div>
     """)
     
-    with gr.Row():
-        with gr.Column(scale=3):
-            gr.HTML('<div class="section-header"> SEARCH PARAMETERS</div>')
+    with gr.Tabs():
+        # Single Website Search Tab
+        with gr.TabItem("Single Website Search"):
+            with gr.Row():
+                with gr.Column(scale=2):
+                    gr.HTML('<div class="section-header">Search Parameters</div>')
+                    
+                    website_input = gr.Textbox(
+                        label="Target Website URL",
+                        placeholder="Enter company website (e.g., techflowsolutions.com)",
+                        value="",
+                        elem_classes=["custom-input"]
+                    )
+                    
+                    with gr.Row():
+                        max_results = gr.Slider(
+                            minimum=1,
+                            maximum=20,
+                            value=8,
+                            step=1,
+                            label="Maximum Results",
+                            elem_classes=["custom-input"]
+                        )
+                        
+                        search_btn = gr.Button(
+                            "Execute Search", 
+                            variant="primary", 
+                            size="lg",
+                            elem_classes=["primary-btn"]
+                        )
             
-            website_input = gr.Textbox(
-                label="Target Website URL",
-                placeholder="Enter company website (e.g., techflowsolutions.com)",
-                value="",
-                elem_classes=["custom-input"]
-            )
+            gr.HTML('<div class="section-header">Search Results</div>')
             
             with gr.Row():
-                max_results = gr.Slider(
-                    minimum=1,
-                    maximum=20,
-                    value=8,
-                    step=1,
-                    label="📊 Maximum Results",
-                    elem_classes=["custom-input"]
+                results_display = gr.Textbox(
+                    label="Contact Intelligence Report",
+                    lines=18,
+                    max_lines=35,
+                    show_copy_button=True,
+                    elem_classes=["results-container"]
                 )
                 
-                search_btn = gr.Button(
-                    "🔍 EXECUTE SEARCH", 
-                    variant="primary", 
-                    size="lg",
-                    elem_classes=["primary-btn"]
+                csv_output = gr.Textbox(
+                    label="Export Data (CSV Format)",
+                    lines=18,
+                    max_lines=35,
+                    show_copy_button=True,
+                    elem_classes=["results-container"]
                 )
-            
-        with gr.Column(scale=2):
-            gr.HTML("""
-            <div class="tips-section">
-                <h3>💼 CORPORATE INTELLIGENCE</h3>
-                <ul style="margin: 10px 0; padding-left: 20px;">
-                    <li><strong>Real-Time Analysis:</strong> Live website scanning</li>
-                    <li><strong>Executive Contacts:</strong> C-level and key personnel</li>
-                    <li><strong>Direct Communications:</strong> Email & phone verification</li>
-                    <li><strong>Export Ready:</strong> CSV format for CRM integration</li>
-                </ul>
-                <div style="background: #1e40af; color: white; padding: 10px; border-radius: 5px; margin-top: 15px; text-align: center; font-weight: 600;">
-                    ENTERPRISE-GRADE CONTACT DISCOVERY
-                </div>
-            </div>
-            """)
-    
-    gr.HTML('<div class="section-header">📈 INTELLIGENCE RESULTS</div>')
-    
-    with gr.Row():
-        results_display = gr.Textbox(
-            label="📋 Contact Intelligence Report",
-            lines=18,
-            max_lines=35,
-            show_copy_button=True,
-            elem_classes=["results-container"]
-        )
         
-        csv_output = gr.Textbox(
-            label="📊 Export Data (CSV Format)",
-            lines=18,
-            max_lines=35,
-            show_copy_button=True,
-            elem_classes=["results-container"]
-        )
+        # CSV Upload Tab
+        with gr.TabItem("CSV Bulk Search"):
+            with gr.Row():
+                with gr.Column(scale=2):
+                    gr.HTML('<div class="section-header">CSV Upload</div>')
+                    
+                    csv_file = gr.File(
+                        label="Upload CSV File",
+                        file_types=[".csv"],
+                        elem_classes=["custom-input"]
+                    )
+                    
+                    gr.HTML("""
+                    <div style="background: #f8fafc; padding: 15px; border-radius: 8px; border-left: 4px solid #1e40af; margin: 10px 0;">
+                        <strong>CSV Format Requirements:</strong><br>
+                        • Include a column named 'website', 'url', or 'domain'<br>
+                        • One website per row<br>
+                        • Example: techflowsolutions.com, greenleafconsult.com
+                    </div>
+                    """)
+                    
+                    with gr.Row():
+                        csv_max_results = gr.Slider(
+                            minimum=1,
+                            maximum=50,
+                            value=20,
+                            step=1,
+                            label="Maximum Results",
+                            elem_classes=["custom-input"]
+                        )
+                        
+                        csv_search_btn = gr.Button(
+                            "Process CSV", 
+                            variant="primary", 
+                            size="lg",
+                            elem_classes=["primary-btn"]
+                        )
+            
+            gr.HTML('<div class="section-header">CSV Results</div>')
+            
+            with gr.Row():
+                csv_results_display = gr.Textbox(
+                    label="CSV Processing Report",
+                    lines=18,
+                    max_lines=35,
+                    show_copy_button=True,
+                    elem_classes=["results-container"]
+                )
+                
+                csv_export_output = gr.Textbox(
+                    label="Export Data (CSV Format)",
+                    lines=18,
+                    max_lines=35,
+                    show_copy_button=True,
+                    elem_classes=["results-container"]
+                )
     
     # Sample websites section
-    with gr.Accordion("🏢 CORPORATE DATABASE SAMPLES", open=False):
+    with gr.Accordion("Sample Websites Database", open=False):
         gr.HTML('<div style="background: #f8fafc; padding: 15px; border-radius: 8px; border-left: 4px solid #1e40af;">')
         sample_websites = gr.Textbox(
-            label="Available Corporate Websites in Database",
+            label="Available Websites in Database",
             value=get_all_available_websites(),
             lines=8,
             interactive=False,
@@ -530,19 +645,19 @@ with gr.Blocks(css=custom_css, title="ZoomInfo Contact Discovery", theme=gr.them
         gr.HTML('</div>')
     
     # Quick search buttons
-    gr.HTML('<div class="section-header">⚡ QUICK ACCESS CORPORATE TARGETS</div>')
+    gr.HTML('<div class="section-header">Quick Access Sample Websites</div>')
     
     with gr.Row():
-        quick_btn1 = gr.Button("🔷 TechFlow Solutions", size="sm", elem_classes=["secondary-btn"])
-        quick_btn2 = gr.Button("🌿 GreenLeaf Consulting", size="sm", elem_classes=["secondary-btn"])
-        quick_btn3 = gr.Button("🎯 BlueSky Marketing", size="sm", elem_classes=["secondary-btn"])
-        quick_btn4 = gr.Button("⚛️ Quantum Dynamics", size="sm", elem_classes=["secondary-btn"])
+        quick_btn1 = gr.Button("TechFlow Solutions", size="sm", elem_classes=["secondary-btn"])
+        quick_btn2 = gr.Button("GreenLeaf Consulting", size="sm", elem_classes=["secondary-btn"])
+        quick_btn3 = gr.Button("BlueSky Marketing", size="sm", elem_classes=["secondary-btn"])
+        quick_btn4 = gr.Button("Quantum Dynamics", size="sm", elem_classes=["secondary-btn"])
     
     with gr.Row():
-        quick_btn5 = gr.Button("📦 Stellar Logistics", size="sm", elem_classes=["secondary-btn"])
-        quick_btn6 = gr.Button("💰 Nexus Financial", size="sm", elem_classes=["secondary-btn"])
-        quick_btn7 = gr.Button("🏥 Horizon Health", size="sm", elem_classes=["secondary-btn"])
-        quick_btn8 = gr.Button("🏭 Phoenix Manufacturing", size="sm", elem_classes=["secondary-btn"])
+        quick_btn5 = gr.Button("Stellar Logistics", size="sm", elem_classes=["secondary-btn"])
+        quick_btn6 = gr.Button("Nexus Financial", size="sm", elem_classes=["secondary-btn"])
+        quick_btn7 = gr.Button("Horizon Health", size="sm", elem_classes=["secondary-btn"])
+        quick_btn8 = gr.Button("Phoenix Manufacturing", size="sm", elem_classes=["secondary-btn"])
     
     # Event handlers
     search_btn.click(
@@ -551,6 +666,12 @@ with gr.Blocks(css=custom_css, title="ZoomInfo Contact Discovery", theme=gr.them
         outputs=[results_display, csv_output]
     )
     
+    csv_search_btn.click(
+        fn=search_csv_websites,
+        inputs=[csv_file, csv_max_results],
+        outputs=[csv_results_display, csv_export_output]
+    )
+    
     # Quick search button handlers
     quick_btn1.click(lambda: "techflowsolutions.com", outputs=website_input)
     quick_btn2.click(lambda: "greenleafconsult.com", outputs=website_input)
@@ -570,14 +691,14 @@ with gr.Blocks(css=custom_css, title="ZoomInfo Contact Discovery", theme=gr.them
             ["quantumdynamics.org", 6]
         ],
         inputs=[website_input, max_results],
-        label="🎯 Sample Corporate Searches"
+        label="Sample Searches"
     )
     
     # Footer
     gr.HTML("""
     <div style="text-align: center; padding: 30px 20px; background: linear-gradient(135deg, #64748b 0%, #475569 100%); color: white; border-radius: 15px; margin-top: 30px;">
-        <h3 style="margin: 0 0 10px 0;">KWEKEL CONTACT INTELLIGENCE PLATFORM</h3>
-        <p style="margin: 0; opacity: 0.9;">Professional-grade corporate contact discovery and lead generation technology</p>
+        <h3 style="margin: 0 0 10px 0;">Contact Intelligence Platform</h3>
+        <p style="margin: 0; opacity: 0.9;">Professional-grade contact discovery and lead generation technology</p>
         <p style="margin: 10px 0 0 0; font-size: 0.9em; opacity: 0.7;">Powered by advanced web intelligence algorithms</p>
     </div>
     """)