app.py

import gradio as gr
import requests
import json
from typing import List, Dict, Any
import pandas as pd
from urllib.parse import urlparse

# Hardcoded API endpoint - Updated to correct endpoint
API_ENDPOINT = "https://jaybene1-testapicontacts.hf.space/contacts"

# Professional color scheme inspired by sale leaseback theme - Fixed theme configuration
theme = gr.themes.Soft(
    primary_hue="blue",
    secondary_hue="slate",
    neutral_hue="slate",
    text_size="md",
    spacing_size="md",
    radius_size="md"
).set(
    # Background colors
    body_background_fill="#1a1a1a",
    body_background_fill_dark="#1a1a1a",
    
    # Text colors
    body_text_color="#e5e5e5",
    body_text_color_dark="#e5e5e5",
    
    # Block colors
    block_background_fill="#2d2d2d",
    block_background_fill_dark="#2d2d2d",
    block_border_color="#404040",
    block_border_color_dark="#404040",
    block_label_text_color="#e5e5e5",
    block_label_text_color_dark="#e5e5e5",
    block_title_text_color="#e5e5e5",
    block_title_text_color_dark="#e5e5e5",
    
    # Input colors (using valid property names)
    input_background_fill="#2d2d2d",
    input_background_fill_dark="#2d2d2d",
    input_border_color="#555555",
    input_border_color_dark="#555555",
    input_placeholder_color="#a0a0a0",
    input_placeholder_color_dark="#a0a0a0",
    
    # Button colors
    button_primary_background_fill="#1e40af",
    button_primary_background_fill_hover="#2563eb",
    button_primary_text_color="#ffffff",
    button_secondary_background_fill="#404040",
    button_secondary_background_fill_hover="#555555",
    button_secondary_text_color="#ffffff"
)

def normalize_url(url: str) -> str:
    """
    Normalize URL for consistent comparison.
    
    Args:
        url: The URL to normalize
        
    Returns:
        Normalized URL string
    """
    if not url:
        return ""
    
    # Remove protocol and www
    normalized = url.lower().replace('https://', '').replace('http://', '').replace('www.', '')
    
    # Remove trailing slash and path
    normalized = normalized.split('/')[0]
    
    # Remove port numbers
    normalized = normalized.split(':')[0]
    
    return normalized

def extract_domain_variations(url: str) -> List[str]:
    """
    Extract various domain variations for matching.
    
    Args:
        url: The URL to extract domains from
        
    Returns:
        List of domain variations
    """
    normalized = normalize_url(url)
    variations = [normalized]
    
    # Add with www
    variations.append(f"www.{normalized}")
    
    # Add subdomains if any
    if '.' in normalized:
        parts = normalized.split('.')
        if len(parts) > 2:
            # Add without subdomain
            variations.append('.'.join(parts[-2:]))
    
    return variations

def filter_contacts_by_url(contacts, search_url):
    """
    Enhanced filter contacts based on the search URL.
    This is a client-side filter in case the API doesn't filter properly.
    """
    if not isinstance(contacts, list):
        return contacts
    
    # Get domain variations for the search URL
    search_domains = extract_domain_variations(search_url)
    
    filtered_contacts = []
    for contact in contacts:
        if isinstance(contact, dict):
            # Check various fields that might contain the domain
            fields_to_check = [
                contact.get('website', ''),
                contact.get('company_website', ''),
                contact.get('domain', ''),
                contact.get('email', ''),
                contact.get('company', ''),
                str(contact.get('url', '')),
                str(contact.get('source', '')),
                str(contact.get('company_url', '')),
                str(contact.get('origin_url', ''))
            ]
            
            # Check if any field contains any of the search domains
            contact_matched = False
            for field in fields_to_check:
                if field:
                    field_normalized = normalize_url(str(field))
                    for search_domain in search_domains:
                        if search_domain in field_normalized or field_normalized in search_domain:
                            filtered_contacts.append(contact)
                            contact_matched = True
                            break
                    if contact_matched:
                        break
            
            # Also check email domains
            if not contact_matched:
                email = contact.get('email', '')
                if email and '@' in email:
                    email_domain = email.split('@')[1].lower()
                    for search_domain in search_domains:
                        if search_domain in email_domain or email_domain in search_domain:
                            filtered_contacts.append(contact)
                            break
    
    return filtered_contacts

def search_contacts(url: str) -> tuple[str, str]:
    """
    Search for contacts associated with a given URL using the Hugging Face API.
    
    Args:
        url: The website URL to search for contacts
        
    Returns:
        Tuple of (formatted_results, raw_json)
    """
    
    if not url.strip():
        return "Please enter a website URL", ""
    
    # Clean and validate URL
    url = url.strip()
    original_url = url
    if not url.startswith(('http://', 'https://')):
        url = 'https://' + url
    
    try:
        # Use GET request with URL parameter
        params = {
            'url': url
        }
        
        # Make the API request to the hardcoded endpoint using GET
        response = requests.get(
            API_ENDPOINT,
            params=params,
            timeout=30
        )
        
        if response.status_code == 200:
            try:
                result = response.json()
                
                # Hugging Face Spaces API returns data in a different format
                if isinstance(result, dict) and 'data' in result:
                    actual_result = result['data'][0] if result['data'] else {}
                else:
                    actual_result = result
                
                # Apply client-side filtering
                if isinstance(actual_result, list):
                    original_count = len(actual_result)
                    filtered_result = filter_contacts_by_url(actual_result, url)
                    
                    # If we filtered out everything or got very few results compared to original,
                    # it's likely the API returned all contacts instead of URL-specific ones
                    if len(filtered_result) == 0:
                        return f"No contacts found for: {original_url}\n\nThis URL may not be in our database.", json.dumps(result, indent=2)
                    elif original_count > 50 and len(filtered_result) < 5:
                        # Likely got all contacts, filtered to very few - show warning
                        formatted_output = f"**Warning:** API returned {original_count} total contacts. After filtering for '{original_url}', found {len(filtered_result)} matches.\n\n"
                        formatted_output += "These results may not be accurate. The URL might not be in our database.\n\n"
                        formatted_output += format_contact_results(filtered_result, url)
                    else:
                        formatted_output = format_contact_results(filtered_result, url)
                    
                    actual_result = filtered_result
                    
                elif isinstance(actual_result, dict) and 'contacts' in actual_result:
                    original_count = len(actual_result['contacts'])
                    filtered_contacts = filter_contacts_by_url(actual_result['contacts'], url)
                    
                    if len(filtered_contacts) == 0:
                        return f"No contacts found for: {original_url}\n\nThis URL may not be in our database.", json.dumps(result, indent=2)
                    elif original_count > 50 and len(filtered_contacts) < 5:
                        formatted_output = f"**Warning:** API returned {original_count} total contacts. After filtering for '{original_url}', found {len(filtered_contacts)} matches.\n\n"
                        formatted_output += "These results may not be accurate. The URL might not be in our database.\n\n"
                        formatted_output += format_contact_results({'contacts': filtered_contacts}, url)
                    else:
                        formatted_output = format_contact_results({'contacts': filtered_contacts}, url)
                    
                    actual_result['contacts'] = filtered_contacts
                else:
                    # Single contact or unknown format
                    formatted_output = format_contact_results(actual_result, url)
                
                # Add debugging info
                debug_info = f"\n\n**Debug Info:**\n"
                debug_info += f"- Request URL: {API_ENDPOINT}\n"
                debug_info += f"- Search URL: {url}\n"
                debug_info += f"- Original response size: {len(result.get('data', [result])) if isinstance(result, dict) else len(result) if isinstance(result, list) else 'N/A'}\n"
                debug_info += f"- Filtered results: {len(actual_result) if isinstance(actual_result, list) else len(actual_result.get('contacts', [])) if isinstance(actual_result, dict) else 'N/A'}\n"
                debug_info += f"- Client-side filtering applied: Yes\n"
                
                if 'formatted_output' not in locals():
                    formatted_output = format_contact_results(actual_result, url)
                
                formatted_output += debug_info
                raw_json = json.dumps(result, indent=2)
                
                return formatted_output, raw_json
                
            except json.JSONDecodeError:
                return f"Error: Invalid JSON response from API", response.text
                
        else:
            return f"API Error ({response.status_code}): {response.text}", ""
            
    except requests.exceptions.Timeout:
        return "Request timeout. Please try again.", ""
    except requests.exceptions.ConnectionError:
        return "Connection error. Please check your API endpoint.", ""
    except Exception as e:
        return f"Error: {str(e)}", ""

def format_contact_results(results: Dict[Any, Any], url: str) -> str:
    """
    Format the API results into a readable format.
    
    Args:
        results: The JSON response from the API
        url: The searched URL
        
    Returns:
        Formatted string with contact information
    """
    
    output = f"# Contact Search Results for: {url}\n\n"
    
    # Handle different possible result structures
    if isinstance(results, dict):
        if 'contacts' in results:
            contacts = results['contacts']
        elif 'results' in results:
            contacts = results['results']
        elif 'data' in results:
            contacts = results['data']
        else:
            contacts = results
            
        if isinstance(contacts, list) and len(contacts) > 0:
            output += f"**Found {len(contacts)} contact(s):**\n\n"
            
            for i, contact in enumerate(contacts, 1):
                output += f"## Contact {i}\n"
                
                # Handle different contact field names
                name = contact.get('name') or contact.get('full_name') or contact.get('contact_name') or "N/A"
                email = contact.get('email') or contact.get('email_address') or "N/A"
                title = contact.get('title') or contact.get('job_title') or contact.get('position') or "N/A"
                company = contact.get('company') or contact.get('organization') or "N/A"
                phone = contact.get('phone') or contact.get('phone_number') or "N/A"
                linkedin = contact.get('linkedin') or contact.get('linkedin_url') or "N/A"
                
                output += f"- **Name:** {name}\n"
                output += f"- **Email:** {email}\n"
                output += f"- **Title:** {title}\n"
                output += f"- **Company:** {company}\n"
                output += f"- **Phone:** {phone}\n"
                output += f"- **LinkedIn:** {linkedin}\n\n"
                
                # Add any additional fields
                additional_fields = {k: v for k, v in contact.items() 
                                   if k not in ['name', 'full_name', 'contact_name', 'email', 'email_address', 
                                              'title', 'job_title', 'position', 'company', 'organization', 
                                              'phone', 'phone_number', 'linkedin', 'linkedin_url']}
                
                if additional_fields:
                    output += "**Additional Information:**\n"
                    for key, value in additional_fields.items():
                        output += f"- **{key.replace('_', ' ').title()}:** {value}\n"
                    output += "\n"
                    
                output += "---\n\n"
                
        else:
            output += "No contacts found for this URL.\n\n"
            
    elif isinstance(results, list):
        if len(results) > 0:
            output += f"**Found {len(results)} contact(s):**\n\n"
            for i, contact in enumerate(results, 1):
                output += f"## Contact {i}\n"
                output += f"{contact}\n\n"
        else:
            output += "No contacts found for this URL.\n\n"
    else:
        output += f"**Result:** {results}\n\n"
    
    output += f"*Search completed at: {pd.Timestamp.now().strftime('%Y-%m-%d %H:%M:%S')}*"
    
    return output

# Create the Gradio interface
with gr.Blocks(theme=theme, title="Kwekel Companies Contact Search Tool", css="""
    /* Root and body styles */
    :root {
        --body-text-color: #e5e5e5 !important;
        --neutral-50: #1a1a1a !important;
        --neutral-100: #2d2d2d !important;
        --neutral-200: #404040 !important;
        --neutral-300: #555555 !important;
        --neutral-400: #6b7280 !important;
        --neutral-500: #9ca3af !important;
        --neutral-600: #d1d5db !important;
        --neutral-700: #e5e7eb !important;
        --neutral-800: #f3f4f6 !important;
        --neutral-900: #ffffff !important;
        --neutral-950: #ffffff !important;
    }
    
    .gradio-container {
        max-width: 1200px !important;
        margin: 0 auto !important;
        background-color: #1a1a1a !important;
        color: #e5e5e5 !important;
    }
    
    /* Header styles with blue gradient */
    .header {
        background: linear-gradient(135deg, #1e3a8a 0%, #3b82f6 50%, #1e40af 100%);
        color: white;
        padding: 2rem;
        border-radius: 0.5rem;
        margin-bottom: 2rem;
        text-align: center;
        border: 1px solid #404040;
        box-shadow: 0 4px 15px rgba(30, 58, 138, 0.3);
    }
    .header h1 {
        margin: 0;
        font-size: 2.5rem;
        font-weight: 700;
        margin-bottom: 0.5rem;
        color: #ffffff !important;
        text-shadow: 2px 2px 4px rgba(0,0,0,0.3);
    }
    
    /* Info box styles */
    .info-box {
        background: #2d2d2d;
        border: 1px solid #1e40af;
        border-radius: 0.5rem;
        padding: 1rem;
        margin: 1rem 0;
        color: #ffffff;
    }
    .info-box h3 {
        color: #ffffff !important;
        margin-top: 0;
    }
    .info-box p, .info-box ol, .info-box li {
        color: #e5e5e5 !important;
    }
    
    /* Footer styles */
    .footer {
        text-align: center;
        margin-top: 2rem;
        padding-top: 2rem;
        border-top: 1px solid #404040;
        color: #cccccc !important;
    }
    
    /* Global text color overrides */
    body, .gradio-container, .gradio-container * {
        color: #e5e5e5 !important;
    }
    
    /* Input field styles */
    .gr-textbox input, 
    .gr-textbox textarea,
    input[type="text"], 
    input[type="email"], 
    input[type="url"], 
    textarea {
        color: #e5e5e5 !important;
        background-color: #2d2d2d !important;
        border: 1px solid #555555 !important;
    }
    
    /* Input placeholder styles */
    .gr-textbox input::placeholder,
    .gr-textbox textarea::placeholder,
    input::placeholder,
    textarea::placeholder {
        color: #a0a0a0 !important;
    }
    
    /* Label styles */
    label, .gr-label {
        color: #e5e5e5 !important;
    }
    
    /* Specific styling for Website URL input label */
    .gr-textbox:has(input[placeholder*="example.com"]) label {
        color: #3b3a3a !important;
    }
    
    /* Markdown and code styles */
    .gr-markdown, .gr-markdown * {
        color: #e5e5e5 !important;
    }
    
    .gr-code, .gr-code * {
        color: #3b3a3a !important;
        background-color: #2d2d2d !important;
    }
    
    /* Code block specific styling */
    .gr-code pre, .gr-code pre * {
        color: #3b3a3a !important;
    }
    
    /* Form and block styles */
    .gr-form, .gr-form * {
        color: #e5e5e5 !important;
    }
    
    .gr-block, .gr-block * {
        color: #e5e5e5 !important;
    }
    
    /* Panel and container styles */
    .gr-panel {
        background-color: #2d2d2d !important;
        border: 1px solid #404040 !important;
    }
    
    .gr-container {
        background-color: #1a1a1a !important;
    }
    
    /* Button text visibility */
    .gr-button {
        color: #ffffff !important;
    }
    
    /* Primary button with blue gradient */
    .gr-button.primary {
        background: linear-gradient(135deg, #1e3a8a 0%, #3b82f6 50%, #1e40af 100%) !important;
        border: none !important;
        box-shadow: 0 4px 15px rgba(30, 58, 138, 0.3) !important;
        transition: all 0.3s ease !important;
    }
    
    .gr-button.primary:hover {
        background: linear-gradient(135deg, #1e40af 0%, #3b82f6 50%, #2563eb 100%) !important;
        box-shadow: 0 6px 20px rgba(30, 58, 138, 0.4) !important;
        transform: translateY(-2px) !important;
    }
    
    /* Ensure info text is visible */
    .gr-info {
        color: #a0a0a0 !important;
    }
    
    /* Additional specific overrides */
    .gr-textbox-label,
    .gr-markdown-label,
    .gr-code-label {
        color: #e5e5e5 !important;
    }
    
    /* Override for Website URL label specifically */
    .gr-textbox:has(input[placeholder*="example.com"]) .gr-textbox-label {
        color: #3b3a3a !important;
    }
    
    /* Override any remaining dark text */
    p, span, div, h1, h2, h3, h4, h5, h6, li, td, th {
        color: #e5e5e5 !important;
    }
    
    /* Ensure all text inputs have proper contrast */
    input:focus, textarea:focus {
        color: #e5e5e5 !important;
        background-color: #2d2d2d !important;
        border-color: #1e40af !important;
    }
    
    /* Specific targeting for the Website URL input text */
    .gr-textbox input[placeholder*="example.com"] {
        color: #3b3a3a !important;
    }
    
    /* JSON code block text color */
    .gr-code[data-testid*="json"] pre,
    .gr-code[data-testid*="json"] code,
    .gr-code[language="json"] pre,
    .gr-code[language="json"] code {
        color: #3b3a3a !important;
    }
""") as demo:
    
    # Header
    gr.HTML("""
    <div class="header">
        <h1>Kwekel Companies Contact Search Tool</h1>
    </div>
    """)
    
    # Info box
    gr.HTML(f"""
    <div class="info-box">
        <h3>Instructions:</h3>
        <ol>
            <li>Enter the website URL you want to search for contacts</li>
            <li>Click "Search Contacts" to get results</li>
            <li>View formatted results and raw JSON response</li>
        </ol>
        <p><strong>API Endpoint:</strong> {API_ENDPOINT}</p>
        <p><strong>Note:</strong> The system will filter results to match your specific URL and warn you if the URL might not be in the database.</p>
    </div>
    """)
    
    with gr.Row():
        with gr.Column(scale=2):
            url_input = gr.Textbox(
                label="Website URL",
                placeholder="example.com or https://example.com",
                info="Enter the website URL to search for contacts",
                lines=1
            )
            
            with gr.Row():
                search_btn = gr.Button("Search Contacts", variant="primary", scale=2)
                clear_btn = gr.Button("Clear", variant="secondary", scale=1)
    
    with gr.Row():
        with gr.Column(scale=1):
            results_output = gr.Markdown(
                label="Contact Results",
                value="Enter a URL and click 'Search Contacts' to see results here."
            )
        
        with gr.Column(scale=1):
            json_output = gr.Code(
                label="Raw JSON Response",
                language="json",
                value=""
            )
    
    # Event handlers
    search_btn.click(
        fn=search_contacts,
        inputs=[url_input],
        outputs=[results_output, json_output]
    )
    
    clear_btn.click(
        fn=lambda: ("", ""),
        outputs=[url_input, results_output]
    )
    
    # Footer
    gr.HTML("""
    <div class="footer">
        <p>© 2025 Kwekel Companies Contact Search Tool</p>
    </div>
    """)

# Launch the app
if __name__ == "__main__":
    demo.launch(
        share=True,
        server_name="0.0.0.0",
        server_port=7860,
        show_error=True,
        debug=True
    )