Spaces:

StrategyStack
/

README

Sleeping

README / app.py

Fix indentation error on line 247 and separate concatenated statements on lines 247-249

0b3754d verified about 1 year ago

14.9 kB

import gradio as gr
import sys
import os
import json
from typing import List, Dict, Any, Optional
from pathlib import Path
from linkedin_data import fetch_linkedin_profile, get_cache_stats
from data_enrichment import DataEnrichmentManager

# --- Data Storage Configuration ---
DATA_DIR = Path("/data")
USER_DATA_DIR = DATA_DIR / "users"
USER_DATA_DIR.mkdir(parents=True, exist_ok=True)

def get_user_dir(username: str) -> Path:
    """Get or create user-specific data directory"""
    user_dir = USER_DATA_DIR / username
    user_dir.mkdir(parents=True, exist_ok=True)
    return user_dir

def load_user_data(username: str, data_type: str) -> Any:
    """Load user-specific data from their directory"""
    user_dir = get_user_dir(username)
    data_file = user_dir / f"{data_type}.json"
    if data_file.exists():
        with open(data_file, 'r') as f:
            return json.load(f)
    return []

def save_user_data(username: str, data_type: str, data: Any):
    """Save user-specific data to their directory"""
    user_dir = get_user_dir(username)
    data_file = user_dir / f"{data_type}.json"
    with open(data_file, 'w') as f:
        json.dump(data, f, indent=2)

def get_enrichment_manager(username: str) -> DataEnrichmentManager:
    """Get or create DataEnrichmentManager for a user"""
    return DataEnrichmentManager(username)
# --- Consent/Event Audit Panel ---
def consent_audit_panel(consent_events: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
    """
    UI Panel to show all consent events, opt-outs, and compliance flags per contact and batch.
    Accepts a list of consent event dicts with structure as returned from backend JSON.
    """
    if profile is None:
        return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."
    
    username = profile.username
    # Load user-specific consent events
    user_events = load_user_data(username, "consent_events")
    if not user_events:
        user_events = consent_events  # Use provided sample if no user data
    
    audit_md = f"# Consent/Event Audit\n\n**User:** {username}\n\nPer-contact and batch consent history."
    for event in user_events:
                audit_md += f"\n- Contact: {event.get('contact_id','?')} | Event: {event.get('event_type','?')} | Opt-out: {event.get('opt_out','?')}"
            
    # LinkedIn enrichment section
    audit_md += "\n\n---\n\n## 🔗 LinkedIn Profile Enrichment\n\n"
    
    # Get cache stats
    cache_stats = get_cache_stats(username)
    audit_md += f"**Cached LinkedIn Profiles:** {cache_stats.get('cached_profiles', 0)}\n\n"
    
    # Sample LinkedIn enrichment (demonstrating the feature)
    audit_md += "**To enrich a contact with LinkedIn data:**\n"
    audit_md += "1. Add a 'linkedin_url' field to your consent events\n"
    audit_md += "2. The system will automatically fetch and cache profile data\n"
    audit_md += "3. Cached profiles are stored in your isolated data directory\n\n"
    
    # Example with sample LinkedIn URL (if configured)
    sample_linkedin_url = "https://linkedin.com/in/sample-profile"
    audit_md += f"**Example:** To test LinkedIn enrichment, add RAPIDAPI_KEY to Space Secrets.\n\n"
    
    if user_events and len(user_events) > 0:
        # Check if any event has LinkedIn data
        for event in user_events:
            if 'linkedin_url' in event and event['linkedin_url']:
                linkedin_url = event['linkedin_url']
                audit_md += f"\n### LinkedIn Profile for Contact {event.get('contact_id', '?')}\n\n"
                
                # Fetch LinkedIn profile
                result = fetch_linkedin_profile(linkedin_url, username=username)
                
                if result.get('success'):
                    profile_data = result.get('data', {})
                    from_cache = result.get('from_cache', False)
                    
                    audit_md += f"**Full Name:** {profile_data.get('full_name', 'N/A')}\n\n"
                    audit_md += f"**Headline:** {profile_data.get('headline', 'N/A')}\n\n"
                    audit_md += f"**Location:** {profile_data.get('location', 'N/A')}\n\n"
                    audit_md += f"**Data Source:** {'Cached (30-day TTL)' if from_cache else 'Fresh from API'}\n\n"
                    
                    # Show first experience if available
                    experiences = profile_data.get('experiences', [])
                    if experiences and len(experiences) > 0:
                        exp = experiences[0]
                        audit_md += f"**Current Role:** {exp.get('title', 'N/A')} at {exp.get('company', 'N/A')}\n\n"
                else:
                    error_msg = result.get('message', 'Unknown error')
                    audit_md += f"⚠️ **Error:** {error_msg}\n\n"
                    
        audit_md += f"\n- Contact: {event.get('contact_id','?')} | Event: {event.get('event_type','?')} | Opt-out: {event.get('opt_out','?')}"
    return gr.Markdown(audit_md)

# --- Dynamic Branching Composer for Cadence Editing ---
def branching_composer(branches: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
    """
    Visual sequence builder for editing cadences.
    Structure assumed to be a list of dict branches with JSON mirroring backend.
    """
    if profile is None:
        return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."
    
    username = profile.username
    # Load user-specific branches
    user_branches = load_user_data(username, "branches")
    if not user_branches:
        user_branches = branches  # Use provided sample if no user data
    
    builder_md = f"# Dynamic Cadence Branching\n\n**User:** {username}\n\n## Sequence Tree:"
    for b in user_branches:
        builder_md += f"\n- Step {b.get('step','?')}: {b.get('description','?')} (Next: {b.get('next','?')})"
    return gr.Markdown(builder_md)

# --- Live Diagnostics Panel for Cadence Errors/Health ---
def diagnostics_panel(events: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
    """
    Incident/event log UI widget to show errors and health of cadence system.
    Format is a simple list of event dicts from backend state.
    """
    if profile is None:
        return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."
    
    username = profile.username
    # Load user-specific diagnostic events
    user_events = load_user_data(username, "diagnostics")
    if not user_events:
        user_events = events  # Use provided sample if no user data
    
    diag_md = f"# Live Diagnostics: Cadence Health\n\n**User:** {username}"
    for e in user_events:
        diag_md += f"\n- [{e.get('timestamp','?')}] {e.get('type','?')}: {e.get('message','?')} {'[ERROR]' if e.get('level')=='error' else ''}"
    return gr.Markdown(diag_md)

# --- Audit Trail Export Widget ---
def audit_trail_widget(profile: Optional[gr.OAuthProfile]) -> str:
    """
    CSV export widget for compliance officer.
    Returns download button for current batch consent data.
    """
    if profile is None:
        return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."
    
    username = profile.username
    export_md = f"# Audit Trail Export\n\n**User:** {username}\n\nDownload compliance reports in CSV format.\n\n*Export functionality will be implemented based on your specific requirements.*"
    return gr.Markdown(export_md)

# Sample Data (for demo purposes when no user data exists)
consent_events_sample = [
    {"contact_id": "c123", "event_type": "opt_in", "opt_out": False, "timestamp": "2025-01-10"},
    {"contact_id": "c456", "event_type": "opt_out", "opt_out": True, "timestamp": "2025-01-11"},
    {"contact_id": "c789", "event_type": "consent_updated", "opt_out": False, "timestamp": "2025-01-12"},
]

branches_sample = [
    {"step": 1, "description": "Initial outreach email", "next": 2},
    {"step": 2, "description": "Follow-up call", "next": 3},
    {"step": 3, "description": "Send proposal", "next": 4},
    {"step": 4, "description": "Close or nurture", "next": None},
]

diagnostics_sample = [
    {"timestamp": "2025-01-15 10:30", "type": "rate_limit", "message": "LinkedIn API approaching limit (85%)", "level": "warning"},
    {"timestamp": "2025-01-15 11:00", "type": "success", "message": "Batch processed successfully", "level": "info"},
    {"timestamp": "2025-01-15 11:30", "type": "error", "message": "Failed to send email to c999", "level": "error"},
]

audit_history_sample = [
    {"batch_id": "b001", "contacts": 150, "opt_outs": 5, "timestamp": "2025-01-10"},
    {"batch_id": "b002", "contacts": 200, "opt_outs": 8, "timestamp": "2025-01-12"},
]

# --- Main Application ---
with gr.Blocks(title="StrategyStack UI") as demo:
    gr.Markdown("# 🚀 StrategyStack Recovery, Enrichment & Outreach Platform")
    
    # OAuth Login Button
    gr.LoginButton()
    
    gr.Markdown("""
    ## Welcome to StrategyStack
    
    **Customer-Ready Multi-Tenant Platform**
    
    - ✅ **Secure Authentication**: Sign in with your Hugging Face account
    - ✅ **Data Isolation**: Your data is stored separately and securely
    - ✅ **Privacy First**: We don't store your data centrally - you own it
    - ✅ **Compliance Ready**: Full audit trails and consent management
    
    Please sign in above to access your personalized dashboard.
    """)
    
    # Create tabbed interface
    with gr.Tabs():
        with gr.Tab("Consent/Event Audit"):
            consent_output = gr.Markdown()
            consent_refresh = gr.Button("Refresh Data")
            user_profile_1 = gr.State()
            
        with gr.Tab("Cadence Branching Composer"):
            branching_output = gr.Markdown()
            branching_refresh = gr.Button("Refresh Data")
            user_profile_2 = gr.State()
            
        with gr.Tab("Live Diagnostics"):
            diagnostics_output = gr.Markdown()
            diagnostics_refresh = gr.Button("Refresh Data")
            user_profile_3 = gr.State()
            
        with gr.Tab("Audit Trail Export"):
            audit_output = gr.Markdown()
            audit_refresh = gr.Button("Refresh Data")
            user_profile_4 = gr.State()

    with gr.Tab("Data Enrichment"):
        enrichment_output = gr.Markdown()
        enrichment_refresh = gr.Button("Refresh Data")
        user_profile_5 = gr.State()    
    # Load initial data
    def load_consent_data(profile: Optional[gr.OAuthProfile] = None):
        return consent_audit_panel(consent_events_sample, profile)
    
    def load_branching_data(profile: Optional[gr.OAuthProfile] = None):
        return branching_composer(branches_sample, profile)
    
    def load_diagnostics_data(profile: Optional[gr.OAuthProfile] = None):
        return diagnostics_panel(diagnostics_sample, profile)
    
    def load_audit_data(profile: Optional[gr.OAuthProfile] = None):
        return audit_trail_widget(profile)

def load_enrichment_data(profile: Optional[gr.OAuthProfile] = None):
    if profile is None:
                return "# 👤 Authentication Required\n\nPlease sign in with your Hugging Face account to access data enrichment features."
    username = profile.username
    manager = get_enrichment_manager(username)
        stats = manager.get_enrichment_stats()
        enrichment_md = f"# 📊 Data Enrichment Dashboard\n\n**User:** {username}\n\n"
        enrichment_md += "## Enrichment Statistics\n\n"
        enrichment_md += f"- **Total API Calls:** {stats.get('total_api_calls', 0)}\n"
        enrichment_md += f"- **Cache Hits:** {stats.get('cache_hits', 0)}\n"
        enrichment_md += f"- **Cache Misses:** {stats.get('cache_misses', 0)}\n"
        enrichment_md += f"- **Total Enrichments:** {stats.get('total_enrichments', 0)}\n"
        enrichment_md += f"- **Successful Enrichments:** {stats.get('successful_enrichments', 0)}\n"
        enrichment_md += f"- **Failed Enrichments:** {stats.get('failed_enrichments', 0)}\n\n"
        
        # Calculate success rate
        if stats.get('total_enrichments', 0) > 0:
            success_rate = (stats.get('successful_enrichments', 0) / stats['total_enrichments']) * 100
            enrichment_md += f"- **Success Rate:** {success_rate:.1f}%\n\n"
        
        enrichment_md += "## Available APIs\n\n"
        enrichment_md += "- ✅ LinkedIn Profile Enrichment (4 providers with fallback)\n"
        enrichment_md += "- ✅ Company Profile Enrichment\n"
        enrichment_md += "- ✅ Apollo.io Person Enrichment\n"
        enrichment_md += "- ✅ Twitter User Data\n"
        enrichment_md += "- ✅ Facebook Data\n"
        enrichment_md += "- ✅ Google Search\n"
        enrichment_md += "- ✅ Mediafy Enrichment\n"
        enrichment_md += "- ✅ Local Business Data (LetScrape)\n"
        enrichment_md += "- ✅ Website Contacts Scraper (LetScrape)\n\n"
        
        enrichment_md += "## Features\n\n"
        enrichment_md += "- 🔒 **Per-User Data Isolation**: Your enrichment data is stored separately\n"
        enrichment_md += "- ⚡ **Intelligent Caching**: 30-day TTL to minimize API costs\n"
        enrichment_md += "- 🔄 **Automatic Fallback**: Multiple providers ensure high availability\n"
        enrichment_md += "- 📈 **Real-time Statistics**: Track API usage and success rates\n\n"
        
        enrichment_md += "## Usage\n\n"
        enrichment_md += "To use data enrichment in your workflows:\n"
        enrichment_md += "1. Add contact data with fields like `linkedin_url`, `email`, `domain`\n"
        enrichment_md += "2. The system automatically enriches contacts using all available APIs\n"
        enrichment_md += "3. Enriched data is cached for 30 days to optimize performance\n"
        enrichment_md += "4. View statistics and monitor API usage from this dashboard\n\n"
        
        return gr.Markdown(enrichment_md)
    
    # Set up event handlers
    demo.load(load_consent_data, inputs=None, outputs=consent_output)
    demo.load(load_branching_data, inputs=None, outputs=branching_output)
    demo.load(load_diagnostics_data, inputs=None, outputs=diagnostics_output)
    demo.load(load_audit_data, inputs=None, outputs=audit_output)
        demo.load(load_enrichment_data, inputs=None, outputs=enrichment_output)
    
    consent_refresh.click(load_consent_data, inputs=None, outputs=consent_output)
    branching_refresh.click(load_branching_data, inputs=None, outputs=branching_output)
    diagnostics_refresh.click(load_diagnostics_data, inputs=None, outputs=diagnostics_output)
    audit_refresh.click(load_audit_data, inputs=None, outputs=audit_output)
    enrichment_refresh.click(load_enrichment_data, inputs=None, outputs=enrichment_output)

if __name__ == "__main__":
    demo.launch()

	import gradio as gr
	import sys
	import os
	import json
	from typing import List, Dict, Any, Optional
	from pathlib import Path
	from linkedin_data import fetch_linkedin_profile, get_cache_stats
	from data_enrichment import DataEnrichmentManager

	# --- Data Storage Configuration ---
	DATA_DIR = Path("/data")
	USER_DATA_DIR = DATA_DIR / "users"
	USER_DATA_DIR.mkdir(parents=True, exist_ok=True)

	def get_user_dir(username: str) -> Path:
	"""Get or create user-specific data directory"""
	user_dir = USER_DATA_DIR / username
	user_dir.mkdir(parents=True, exist_ok=True)
	return user_dir

	def load_user_data(username: str, data_type: str) -> Any:
	"""Load user-specific data from their directory"""
	user_dir = get_user_dir(username)
	data_file = user_dir / f"{data_type}.json"
	if data_file.exists():
	with open(data_file, 'r') as f:
	return json.load(f)
	return []

	def save_user_data(username: str, data_type: str, data: Any):
	"""Save user-specific data to their directory"""
	user_dir = get_user_dir(username)
	data_file = user_dir / f"{data_type}.json"
	with open(data_file, 'w') as f:
	json.dump(data, f, indent=2)

	def get_enrichment_manager(username: str) -> DataEnrichmentManager:
	"""Get or create DataEnrichmentManager for a user"""
	return DataEnrichmentManager(username)
	# --- Consent/Event Audit Panel ---
	def consent_audit_panel(consent_events: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
	"""
	UI Panel to show all consent events, opt-outs, and compliance flags per contact and batch.
	Accepts a list of consent event dicts with structure as returned from backend JSON.
	"""
	if profile is None:
	return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."

	username = profile.username
	# Load user-specific consent events
	user_events = load_user_data(username, "consent_events")
	if not user_events:
	user_events = consent_events # Use provided sample if no user data

	audit_md = f"# Consent/Event Audit\n\nUser: {username}\n\nPer-contact and batch consent history."
	for event in user_events:
	audit_md += f"\n- Contact: {event.get('contact_id','?')} \| Event: {event.get('event_type','?')} \| Opt-out: {event.get('opt_out','?')}"

	# LinkedIn enrichment section
	audit_md += "\n\n---\n\n## 🔗 LinkedIn Profile Enrichment\n\n"

	# Get cache stats
	cache_stats = get_cache_stats(username)
	audit_md += f"Cached LinkedIn Profiles: {cache_stats.get('cached_profiles', 0)}\n\n"

	# Sample LinkedIn enrichment (demonstrating the feature)
	audit_md += "To enrich a contact with LinkedIn data:\n"
	audit_md += "1. Add a 'linkedin_url' field to your consent events\n"
	audit_md += "2. The system will automatically fetch and cache profile data\n"
	audit_md += "3. Cached profiles are stored in your isolated data directory\n\n"

	# Example with sample LinkedIn URL (if configured)
	sample_linkedin_url = "https://linkedin.com/in/sample-profile"
	audit_md += f"Example: To test LinkedIn enrichment, add RAPIDAPI_KEY to Space Secrets.\n\n"

	if user_events and len(user_events) > 0:
	# Check if any event has LinkedIn data
	for event in user_events:
	if 'linkedin_url' in event and event['linkedin_url']:
	linkedin_url = event['linkedin_url']
	audit_md += f"\n### LinkedIn Profile for Contact {event.get('contact_id', '?')}\n\n"

	# Fetch LinkedIn profile
	result = fetch_linkedin_profile(linkedin_url, username=username)

	if result.get('success'):
	profile_data = result.get('data', {})
	from_cache = result.get('from_cache', False)

	audit_md += f"Full Name: {profile_data.get('full_name', 'N/A')}\n\n"
	audit_md += f"Headline: {profile_data.get('headline', 'N/A')}\n\n"
	audit_md += f"Location: {profile_data.get('location', 'N/A')}\n\n"
	audit_md += f"Data Source: {'Cached (30-day TTL)' if from_cache else 'Fresh from API'}\n\n"

	# Show first experience if available
	experiences = profile_data.get('experiences', [])
	if experiences and len(experiences) > 0:
	exp = experiences[0]
	audit_md += f"Current Role: {exp.get('title', 'N/A')} at {exp.get('company', 'N/A')}\n\n"
	else:
	error_msg = result.get('message', 'Unknown error')
	audit_md += f"⚠️ Error: {error_msg}\n\n"

	audit_md += f"\n- Contact: {event.get('contact_id','?')} \| Event: {event.get('event_type','?')} \| Opt-out: {event.get('opt_out','?')}"
	return gr.Markdown(audit_md)

	# --- Dynamic Branching Composer for Cadence Editing ---
	def branching_composer(branches: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
	"""
	Visual sequence builder for editing cadences.
	Structure assumed to be a list of dict branches with JSON mirroring backend.
	"""
	if profile is None:
	return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."

	username = profile.username
	# Load user-specific branches
	user_branches = load_user_data(username, "branches")
	if not user_branches:
	user_branches = branches # Use provided sample if no user data

	builder_md = f"# Dynamic Cadence Branching\n\nUser: {username}\n\n## Sequence Tree:"
	for b in user_branches:
	builder_md += f"\n- Step {b.get('step','?')}: {b.get('description','?')} (Next: {b.get('next','?')})"
	return gr.Markdown(builder_md)

	# --- Live Diagnostics Panel for Cadence Errors/Health ---
	def diagnostics_panel(events: List[Dict[str, Any]], profile: Optional[gr.OAuthProfile]) -> str:
	"""
	Incident/event log UI widget to show errors and health of cadence system.
	Format is a simple list of event dicts from backend state.
	"""
	if profile is None:
	return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."

	username = profile.username
	# Load user-specific diagnostic events
	user_events = load_user_data(username, "diagnostics")
	if not user_events:
	user_events = events # Use provided sample if no user data

	diag_md = f"# Live Diagnostics: Cadence Health\n\nUser: {username}"
	for e in user_events:
	diag_md += f"\n- [{e.get('timestamp','?')}] {e.get('type','?')}: {e.get('message','?')} {'[ERROR]' if e.get('level')=='error' else ''}"
	return gr.Markdown(diag_md)

	# --- Audit Trail Export Widget ---
	def audit_trail_widget(profile: Optional[gr.OAuthProfile]) -> str:
	"""
	CSV export widget for compliance officer.
	Returns download button for current batch consent data.
	"""
	if profile is None:
	return "# 🔒 Authentication Required\n\nPlease sign in with your Hugging Face account to access this feature."

	username = profile.username
	export_md = f"# Audit Trail Export\n\nUser: {username}\n\nDownload compliance reports in CSV format.\n\nExport functionality will be implemented based on your specific requirements."
	return gr.Markdown(export_md)

	# Sample Data (for demo purposes when no user data exists)
	consent_events_sample = [
	{"contact_id": "c123", "event_type": "opt_in", "opt_out": False, "timestamp": "2025-01-10"},
	{"contact_id": "c456", "event_type": "opt_out", "opt_out": True, "timestamp": "2025-01-11"},
	{"contact_id": "c789", "event_type": "consent_updated", "opt_out": False, "timestamp": "2025-01-12"},
	]

	branches_sample = [
	{"step": 1, "description": "Initial outreach email", "next": 2},
	{"step": 2, "description": "Follow-up call", "next": 3},
	{"step": 3, "description": "Send proposal", "next": 4},
	{"step": 4, "description": "Close or nurture", "next": None},
	]

	diagnostics_sample = [
	{"timestamp": "2025-01-15 10:30", "type": "rate_limit", "message": "LinkedIn API approaching limit (85%)", "level": "warning"},
	{"timestamp": "2025-01-15 11:00", "type": "success", "message": "Batch processed successfully", "level": "info"},
	{"timestamp": "2025-01-15 11:30", "type": "error", "message": "Failed to send email to c999", "level": "error"},
	]

	audit_history_sample = [
	{"batch_id": "b001", "contacts": 150, "opt_outs": 5, "timestamp": "2025-01-10"},
	{"batch_id": "b002", "contacts": 200, "opt_outs": 8, "timestamp": "2025-01-12"},
	]

	# --- Main Application ---
	with gr.Blocks(title="StrategyStack UI") as demo:
	gr.Markdown("# 🚀 StrategyStack Recovery, Enrichment & Outreach Platform")

	# OAuth Login Button
	gr.LoginButton()

	gr.Markdown("""
	## Welcome to StrategyStack

	Customer-Ready Multi-Tenant Platform

	- ✅ Secure Authentication: Sign in with your Hugging Face account
	- ✅ Data Isolation: Your data is stored separately and securely
	- ✅ Privacy First: We don't store your data centrally - you own it
	- ✅ Compliance Ready: Full audit trails and consent management

	Please sign in above to access your personalized dashboard.
	""")

	# Create tabbed interface
	with gr.Tabs():
	with gr.Tab("Consent/Event Audit"):
	consent_output = gr.Markdown()
	consent_refresh = gr.Button("Refresh Data")
	user_profile_1 = gr.State()

	with gr.Tab("Cadence Branching Composer"):
	branching_output = gr.Markdown()
	branching_refresh = gr.Button("Refresh Data")
	user_profile_2 = gr.State()

	with gr.Tab("Live Diagnostics"):
	diagnostics_output = gr.Markdown()
	diagnostics_refresh = gr.Button("Refresh Data")
	user_profile_3 = gr.State()

	with gr.Tab("Audit Trail Export"):
	audit_output = gr.Markdown()
	audit_refresh = gr.Button("Refresh Data")
	user_profile_4 = gr.State()

	with gr.Tab("Data Enrichment"):
	enrichment_output = gr.Markdown()
	enrichment_refresh = gr.Button("Refresh Data")
	user_profile_5 = gr.State()
	# Load initial data
	def load_consent_data(profile: Optional[gr.OAuthProfile] = None):
	return consent_audit_panel(consent_events_sample, profile)

	def load_branching_data(profile: Optional[gr.OAuthProfile] = None):
	return branching_composer(branches_sample, profile)

	def load_diagnostics_data(profile: Optional[gr.OAuthProfile] = None):
	return diagnostics_panel(diagnostics_sample, profile)

	def load_audit_data(profile: Optional[gr.OAuthProfile] = None):
	return audit_trail_widget(profile)

	def load_enrichment_data(profile: Optional[gr.OAuthProfile] = None):
	if profile is None:
	return "# 👤 Authentication Required\n\nPlease sign in with your Hugging Face account to access data enrichment features."
	username = profile.username
	manager = get_enrichment_manager(username)
	stats = manager.get_enrichment_stats()
	enrichment_md = f"# 📊 Data Enrichment Dashboard\n\nUser: {username}\n\n"
	enrichment_md += "## Enrichment Statistics\n\n"
	enrichment_md += f"- Total API Calls: {stats.get('total_api_calls', 0)}\n"
	enrichment_md += f"- Cache Hits: {stats.get('cache_hits', 0)}\n"
	enrichment_md += f"- Cache Misses: {stats.get('cache_misses', 0)}\n"
	enrichment_md += f"- Total Enrichments: {stats.get('total_enrichments', 0)}\n"
	enrichment_md += f"- Successful Enrichments: {stats.get('successful_enrichments', 0)}\n"
	enrichment_md += f"- Failed Enrichments: {stats.get('failed_enrichments', 0)}\n\n"

	# Calculate success rate
	if stats.get('total_enrichments', 0) > 0:
	success_rate = (stats.get('successful_enrichments', 0) / stats['total_enrichments']) * 100
	enrichment_md += f"- Success Rate: {success_rate:.1f}%\n\n"

	enrichment_md += "## Available APIs\n\n"
	enrichment_md += "- ✅ LinkedIn Profile Enrichment (4 providers with fallback)\n"
	enrichment_md += "- ✅ Company Profile Enrichment\n"
	enrichment_md += "- ✅ Apollo.io Person Enrichment\n"
	enrichment_md += "- ✅ Twitter User Data\n"
	enrichment_md += "- ✅ Facebook Data\n"
	enrichment_md += "- ✅ Google Search\n"
	enrichment_md += "- ✅ Mediafy Enrichment\n"
	enrichment_md += "- ✅ Local Business Data (LetScrape)\n"
	enrichment_md += "- ✅ Website Contacts Scraper (LetScrape)\n\n"

	enrichment_md += "## Features\n\n"
	enrichment_md += "- 🔒 Per-User Data Isolation: Your enrichment data is stored separately\n"
	enrichment_md += "- ⚡ Intelligent Caching: 30-day TTL to minimize API costs\n"
	enrichment_md += "- 🔄 Automatic Fallback: Multiple providers ensure high availability\n"
	enrichment_md += "- 📈 Real-time Statistics: Track API usage and success rates\n\n"

	enrichment_md += "## Usage\n\n"
	enrichment_md += "To use data enrichment in your workflows:\n"
	enrichment_md += "1. Add contact data with fields like `linkedin_url`, `email`, `domain`\n"
	enrichment_md += "2. The system automatically enriches contacts using all available APIs\n"
	enrichment_md += "3. Enriched data is cached for 30 days to optimize performance\n"
	enrichment_md += "4. View statistics and monitor API usage from this dashboard\n\n"

	return gr.Markdown(enrichment_md)

	# Set up event handlers
	demo.load(load_consent_data, inputs=None, outputs=consent_output)
	demo.load(load_branching_data, inputs=None, outputs=branching_output)
	demo.load(load_diagnostics_data, inputs=None, outputs=diagnostics_output)
	demo.load(load_audit_data, inputs=None, outputs=audit_output)
	demo.load(load_enrichment_data, inputs=None, outputs=enrichment_output)

	consent_refresh.click(load_consent_data, inputs=None, outputs=consent_output)
	branching_refresh.click(load_branching_data, inputs=None, outputs=branching_output)
	diagnostics_refresh.click(load_diagnostics_data, inputs=None, outputs=diagnostics_output)
	audit_refresh.click(load_audit_data, inputs=None, outputs=audit_output)
	enrichment_refresh.click(load_enrichment_data, inputs=None, outputs=enrichment_output)

	if __name__ == "__main__":
	demo.launch()