app.py

import gradio as gr
import json
import zipfile
import io
import os
from datetime import datetime
from dotenv import load_dotenv
import requests
from bs4 import BeautifulSoup
import tempfile
from pathlib import Path
from support_docs import create_support_docs, export_conversation_to_markdown

# Simple URL content fetching using requests and BeautifulSoup
def get_grounding_context_simple(urls):
    """Fetch grounding context using simple HTTP requests"""
    if not urls:
        return ""
    
    context_parts = []
    for i, url in enumerate(urls, 1):
        if url and url.strip():
            content = fetch_url_content(url.strip())
            context_parts.append(f"Context from URL {i} ({url}):\n{content}")
    
    if context_parts:
        return "\n\n" + "\n\n".join(context_parts) + "\n\n"
    return ""

# Import RAG components
try:
    from rag_tool import RAGTool
    HAS_RAG = True
except ImportError:
    HAS_RAG = False
    RAGTool = None

# Load environment variables from .env file
load_dotenv()

# Utility functions
import re

def extract_urls_from_text(text):
    """Extract URLs from text using regex"""
    url_pattern = r'https?://[^\s<>"{}|\\^`\[\]"]+'
    return re.findall(url_pattern, text)

# Template for generated space app (based on mvp_simple.py)
SPACE_TEMPLATE = '''import gradio as gr
import os
import requests
import json
from bs4 import BeautifulSoup
from datetime import datetime
import tempfile

# Configuration
SPACE_NAME = "{name}"
SPACE_DESCRIPTION = "{description}"
SYSTEM_PROMPT = """{system_prompt}"""
MODEL = "{model}"
GROUNDING_URLS = {grounding_urls}
# Get access code from environment variable for security
ACCESS_CODE = os.environ.get("SPACE_ACCESS_CODE", "{access_code}")
ENABLE_DYNAMIC_URLS = {enable_dynamic_urls}
ENABLE_VECTOR_RAG = {enable_vector_rag}
RAG_DATA = {rag_data_json}

# Get API key from environment - customizable variable name
API_KEY = os.environ.get("{api_key_var}")

def fetch_url_content(url):
    """Fetch and extract text content from a URL using requests and BeautifulSoup"""
    try:
        response = requests.get(url, timeout=10, headers={{'User-Agent': 'Mozilla/5.0'}})
        response.raise_for_status()
        soup = BeautifulSoup(response.content, 'html.parser')
        
        # Remove script and style elements
        for script in soup(["script", "style", "nav", "header", "footer"]):
            script.decompose()
        
        # Get text content
        text = soup.get_text()
        
        # Clean up whitespace
        lines = (line.strip() for line in text.splitlines())
        chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
        text = ' '.join(chunk for chunk in chunks if chunk)
        
        # Truncate to ~4000 characters
        if len(text) > 4000:
            text = text[:4000] + "..."
        
        return text
    except Exception as e:
        return f"Error fetching {{url}}: {{str(e)}}"

# Global cache for URL content to avoid re-crawling in generated spaces
_url_content_cache = {{}}

def get_grounding_context():
    """Fetch context from grounding URLs with caching"""
    if not GROUNDING_URLS:
        return ""
    
    # Create cache key from URLs
    cache_key = tuple(sorted([url for url in GROUNDING_URLS if url and url.strip()]))
    
    # Check cache first
    if cache_key in _url_content_cache:
        return _url_content_cache[cache_key]
    
    context_parts = []
    for i, url in enumerate(GROUNDING_URLS, 1):
        if url.strip():
            content = fetch_url_content(url.strip())
            context_parts.append(f"Context from URL {{i}} ({{url}}):\\n{{content}}")
    
    if context_parts:
        result = "\\n\\n" + "\\n\\n".join(context_parts) + "\\n\\n"
    else:
        result = ""
    
    # Cache the result
    _url_content_cache[cache_key] = result
    return result

def export_conversation_to_markdown(conversation_history):
    """Export conversation history to markdown format"""
    if not conversation_history:
        return "No conversation to export."
    
    markdown_content = f"""# Conversation Export
Generated on: {{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}}

---

"""
    
    for i, message in enumerate(conversation_history):
        if isinstance(message, dict):
            role = message.get('role', 'unknown')
            content = message.get('content', '')
            
            if role == 'user':
                markdown_content += f"## User Message {{(i//2) + 1}}\\n\\n{{content}}\\n\\n"
            elif role == 'assistant':
                markdown_content += f"## Assistant Response {{(i//2) + 1}}\\n\\n{{content}}\\n\\n---\\n\\n"
    
    return markdown_content

# Initialize RAG context if enabled
if ENABLE_VECTOR_RAG and RAG_DATA:
    try:
        import faiss
        import numpy as np
        import base64
        
        class SimpleRAGContext:
            def __init__(self, rag_data):
                # Deserialize FAISS index
                index_bytes = base64.b64decode(rag_data['index_base64'])
                self.index = faiss.deserialize_index(index_bytes)
                
                # Restore chunks and mappings
                self.chunks = rag_data['chunks']
                self.chunk_ids = rag_data['chunk_ids']
            
            def get_context(self, query, max_chunks=3):
                """Get relevant context - simplified version"""
                # In production, you'd compute query embedding here
                # For now, return a simple message
                return "\\n\\n[RAG context would be retrieved here based on similarity search]\\n\\n"
        
        rag_context_provider = SimpleRAGContext(RAG_DATA)
    except Exception as e:
        print(f"Failed to initialize RAG: {{e}}")
        rag_context_provider = None
else:
    rag_context_provider = None

def generate_response(message, history):
    """Generate response using OpenRouter API"""
    
    if not API_KEY:
        return "Please set your {api_key_var} in the Space settings."
    
    # Get grounding context
    grounding_context = get_grounding_context()
    
    # Add RAG context if available
    if ENABLE_VECTOR_RAG and rag_context_provider:
        rag_context = rag_context_provider.get_context(message)
        if rag_context:
            grounding_context += rag_context
    
    # If dynamic URLs are enabled, check message for URLs to fetch
    if ENABLE_DYNAMIC_URLS:
        urls_in_message = extract_urls_from_text(message)
        if urls_in_message:
            # Fetch content from URLs mentioned in the message
            dynamic_context_parts = []
            for url in urls_in_message[:3]:  # Limit to 3 URLs per message
                content = fetch_url_content(url)
                dynamic_context_parts.append(f"\\n\\nDynamic context from {{url}}:\\n{{content}}")
            if dynamic_context_parts:
                grounding_context += "\\n".join(dynamic_context_parts)
    
    # Build enhanced system prompt with grounding context
    enhanced_system_prompt = SYSTEM_PROMPT + grounding_context
    
    # Build messages array for the API
    messages = [{{"role": "system", "content": enhanced_system_prompt}}]
    
    # Add conversation history - compatible with Gradio 5.x format
    for chat in history:
        if isinstance(chat, dict):
            # New format: {{"role": "user", "content": "..."}} or {{"role": "assistant", "content": "..."}}
            messages.append(chat)
        else:
            # Legacy format: ("user msg", "bot msg")
            user_msg, bot_msg = chat
            messages.append({{"role": "user", "content": user_msg}})
            if bot_msg:
                messages.append({{"role": "assistant", "content": bot_msg}})
    
    # Add current message
    messages.append({{"role": "user", "content": message}})
    
    # Make API request
    try:
        response = requests.post(
            url="https://openrouter.ai/api/v1/chat/completions",
            headers={{
                "Authorization": f"Bearer {{API_KEY}}",
                "Content-Type": "application/json"
            }},
            json={{
                "model": MODEL,
                "messages": messages,
                "temperature": {temperature},
                "max_tokens": {max_tokens}
            }}
        )
        
        if response.status_code == 200:
            return response.json()['choices'][0]['message']['content']
        else:
            return f"Error: {{response.status_code}} - {{response.text}}"
            
    except Exception as e:
        return f"Error: {{str(e)}}"

# Access code verification
access_granted = gr.State(False)
_access_granted_global = False  # Global fallback

def verify_access_code(code):
    \"\"\"Verify the access code\"\"\"
    global _access_granted_global
    if not ACCESS_CODE:
        _access_granted_global = True
        return gr.update(visible=False), gr.update(visible=True), gr.update(value=True)
    
    if code == ACCESS_CODE:
        _access_granted_global = True
        return gr.update(visible=False), gr.update(visible=True), gr.update(value=True)
    else:
        _access_granted_global = False
        return gr.update(visible=True, value="❌ Incorrect access code. Please try again."), gr.update(visible=False), gr.update(value=False)

def protected_generate_response(message, history):
    \"\"\"Protected response function that checks access\"\"\"
    # Check if access is granted via the global variable
    if ACCESS_CODE and not _access_granted_global:
        return "Please enter the access code to continue."
    return generate_response(message, history)

def export_conversation(history):
    \"\"\"Export conversation to markdown file\"\"\"
    if not history:
        return gr.update(visible=False)
    
    markdown_content = export_conversation_to_markdown(history)
    
    # Save to temporary file
    with tempfile.NamedTemporaryFile(mode='w', suffix='.md', delete=False) as f:
        f.write(markdown_content)
        temp_file = f.name
    
    return gr.update(value=temp_file, visible=True)

# Create interface with access code protection
with gr.Blocks(title=SPACE_NAME) as demo:
    gr.Markdown(f"# {{SPACE_NAME}}")
    gr.Markdown(SPACE_DESCRIPTION)
    
    # Access code section (shown only if ACCESS_CODE is set)
    with gr.Column(visible=bool(ACCESS_CODE)) as access_section:
        gr.Markdown("### 🔐 Access Required")
        gr.Markdown("Please enter the access code provided by your instructor:")
        
        access_input = gr.Textbox(
            label="Access Code",
            placeholder="Enter access code...",
            type="password"
        )
        access_btn = gr.Button("Submit", variant="primary")
        access_error = gr.Markdown(visible=False)
    
    # Main chat interface (hidden until access granted)
    with gr.Column(visible=not bool(ACCESS_CODE)) as chat_section:
        chat_interface = gr.ChatInterface(
            fn=protected_generate_response,
            title="",  # Title already shown above
            description="",  # Description already shown above
            examples=None
        )
        
        # Export functionality
        with gr.Row():
            export_btn = gr.Button("Export Conversation", variant="secondary", size="sm")
            export_file = gr.File(label="Download Conversation", visible=False)
        
        # Connect export functionality
        export_btn.click(
            export_conversation,
            inputs=[chat_interface],
            outputs=[export_file]
        )
    
    # Connect access verification
    if ACCESS_CODE:
        access_btn.click(
            verify_access_code,
            inputs=[access_input],
            outputs=[access_error, chat_section, access_granted]
        )
        access_input.submit(
            verify_access_code,
            inputs=[access_input],
            outputs=[access_error, chat_section, access_granted]
        )

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

# Available models
MODELS = [
    "google/gemma-3-27b-it",          # Open-source, sustainable option
    "google/gemini-2.0-flash-001",    # Fast, reliable, general tasks
    "mistralai/mistral-medium",       # Good for technical topics
    "openai/gpt-4o-nano",            # Balanced performance and cost
    "anthropic/claude-3.5-haiku"     # Complex reasoning and analysis
]

def fetch_url_content(url):
    """Fetch and extract text content from a URL"""
    try:
        response = requests.get(url, timeout=10)
        response.raise_for_status()
        soup = BeautifulSoup(response.content, 'html.parser')
        
        # Remove script and style elements
        for script in soup(["script", "style"]):
            script.decompose()
        
        # Get text content
        text = soup.get_text()
        
        # Clean up whitespace
        lines = (line.strip() for line in text.splitlines())
        chunks = (phrase.strip() for line in lines for phrase in line.split("  "))
        text = ' '.join(chunk for chunk in chunks if chunk)
        
        # Truncate to ~4000 characters
        if len(text) > 4000:
            text = text[:4000] + "..."
        
        return text
    except Exception as e:
        return f"Error fetching {url}: {str(e)}"

def get_grounding_context(urls):
    """Fetch context from grounding URLs"""
    if not urls:
        return ""
    
    context_parts = []
    for i, url in enumerate(urls, 1):
        if url and url.strip():
            content = fetch_url_content(url.strip())
            context_parts.append(f"Context from URL {i} ({url}):\n{content}")
    
    if context_parts:
        return "\n\n" + "\n\n".join(context_parts) + "\n\n"
    return ""

def create_readme(config):
    """Generate README with deployment instructions"""
    return f"""---
title: {config['name']}
emoji: 🤖
colorFrom: blue
colorTo: red
sdk: gradio
sdk_version: 5.35.0
app_file: app.py
pinned: false
---

# {config['name']}

{config['description']}

## Quick Deploy to HuggingFace Spaces

### Step 1: Create the Space
1. Go to https://huggingface.co/spaces
2. Click "Create new Space"
3. Choose a name for your Space
4. Select **Gradio** as the SDK
5. Set visibility (Public/Private)
6. Click "Create Space"

### Step 2: Upload Files
1. In your new Space, click "Files" tab
2. Upload these files from the zip:
   - `app.py`
   - `requirements.txt`
3. Wait for "Building" to complete

### Step 3: Add API Key
1. Go to Settings (gear icon) 
2. Click "Variables and secrets"
3. Click "New secret"
4. Name: `{config['api_key_var']}`
5. Value: Your OpenRouter API key
6. Click "Add"

{f'''### Step 4: Configure Access Control
Your Space is configured with access code protection. Students will need to enter the access code to use the chatbot.

1. Go to Settings (gear icon) 
2. Click "Variables and secrets"
3. Click "New secret"
4. Name: `SPACE_ACCESS_CODE`
5. Value: `{config['access_code']}`
6. Click "Add"

**Important**: The access code is now stored securely as an environment variable and is not visible in your app code.

To disable access protection:
1. Go to Settings → Variables and secrets
2. Delete the `SPACE_ACCESS_CODE` secret
3. The Space will rebuild automatically with no access protection

''' if config['access_code'] else ''}

### Step {4 if not config['access_code'] else 5}: Get Your API Key
1. Go to https://openrouter.ai/keys
2. Sign up/login if needed
3. Click "Create Key"
4. Copy the key (starts with `sk-or-`)

### Step {5 if not config['access_code'] else 6}: Test Your Space
- Go back to "App" tab
- Your Space should be running!
- Try the example prompts or ask a question

## Configuration

- **Model**: {config['model']}
- **Temperature**: {config['temperature']}
- **Max Tokens**: {config['max_tokens']}
- **API Key Variable**: {config['api_key_var']}"""
    
    # Add optional configuration items
    if config['access_code']:
        readme_content += f"""
- **Access Code**: {config['access_code']} (Students need this to access the chatbot)"""
    
    if config.get('enable_dynamic_urls'):
        readme_content += """
- **Dynamic URL Fetching**: Enabled (Assistant can fetch URLs mentioned in conversations)"""
    
    readme_content += f"""

## Customization

To modify your Space:
1. Edit `app.py` in your Space
2. Update configuration variables at the top
3. Changes deploy automatically

## Troubleshooting

- **"Please set your {config['api_key_var']}"**: Add the secret in Space settings
- **Error 401**: Invalid API key or no credits
- **Error 429**: Rate limit - wait and try again
- **Build failed**: Check requirements.txt formatting

## More Help

- HuggingFace Spaces: https://huggingface.co/docs/hub/spaces
- OpenRouter Docs: https://openrouter.ai/docs
- Gradio Docs: https://gradio.app/docs

---

Generated on {datetime.now().strftime('%Y-%m-%d %H:%M:%S')} with Chat U/I Helper
"""
    
    return readme_content

def create_requirements(enable_vector_rag=False):
    """Generate requirements.txt"""
    base_requirements = "gradio>=5.35.0\nrequests>=2.32.3\nbeautifulsoup4>=4.12.3"
    
    if enable_vector_rag:
        base_requirements += "\nfaiss-cpu==1.7.4\nnumpy==1.24.3"
    
    return base_requirements

def generate_zip(name, description, system_prompt, model, api_key_var, temperature, max_tokens, examples_text, access_code="", enable_dynamic_urls=False, url1="", url2="", url3="", url4="", enable_vector_rag=False, rag_data=None):
    """Generate deployable zip file"""
    
    # Process examples
    if examples_text and examples_text.strip():
        examples_list = [ex.strip() for ex in examples_text.split('\n') if ex.strip()]
        examples_json = json.dumps(examples_list)
    else:
        examples_json = json.dumps([
            "Hello! How can you help me?",
            "Tell me something interesting",
            "What can you do?"
        ])
    
    # Process grounding URLs
    grounding_urls = []
    for url in [url1, url2, url3, url4]:
        if url and url.strip():
            grounding_urls.append(url.strip())
    
    # Use the provided system prompt directly
    
    # Create config
    config = {
        'name': name,
        'description': description,
        'system_prompt': system_prompt,
        'model': model,
        'api_key_var': api_key_var,
        'temperature': temperature,
        'max_tokens': int(max_tokens),
        'examples': examples_json,
        'grounding_urls': json.dumps(grounding_urls),
        'access_code': "",  # Access code stored in environment variable for security
        'enable_dynamic_urls': enable_dynamic_urls,
        'enable_vector_rag': enable_vector_rag,
        'rag_data_json': json.dumps(rag_data) if rag_data else 'None'
    }
    
    # Generate files
    app_content = SPACE_TEMPLATE.format(**config)
    # Pass original access_code to README for documentation
    readme_config = config.copy()
    readme_config['access_code'] = access_code or ""
    readme_content = create_readme(readme_config)
    requirements_content = create_requirements(enable_vector_rag)
    
    # Create zip file with clean naming
    filename = f"{name.lower().replace(' ', '_').replace('-', '_')}.zip"
    
    # Create zip in memory and save to disk
    zip_buffer = io.BytesIO()
    with zipfile.ZipFile(zip_buffer, 'w', zipfile.ZIP_DEFLATED) as zip_file:
        zip_file.writestr('app.py', app_content)
        zip_file.writestr('requirements.txt', requirements_content)
        zip_file.writestr('README.md', readme_content)
        zip_file.writestr('config.json', json.dumps(config, indent=2))
    
    # Write zip to file
    zip_buffer.seek(0)
    with open(filename, 'wb') as f:
        f.write(zip_buffer.getvalue())
    
    return filename

# Define callback functions outside the interface
def toggle_rag_section(enable_rag):
    """Toggle visibility of RAG section"""
    return gr.update(visible=enable_rag)

def process_documents(files, current_rag_tool):
    """Process uploaded documents"""
    if not files:
        return "Please upload files first", current_rag_tool
    
    if not HAS_RAG:
        return "RAG functionality not available. Please install required dependencies.", current_rag_tool
    
    try:
        # Initialize RAG tool if not exists
        if not current_rag_tool and RAGTool is not None:
            current_rag_tool = RAGTool()
        
        # Process files
        result = current_rag_tool.process_uploaded_files(files)
        
        if result['success']:
            # Create status message
            status_parts = [f"✅ {result['message']}"]
            
            # Add file summary
            if result['summary']['files_processed']:
                status_parts.append("\n**Processed files:**")
                for file_info in result['summary']['files_processed']:
                    status_parts.append(f"- {file_info['name']} ({file_info['chunks']} chunks)")
            
            # Add errors if any
            if result.get('errors'):
                status_parts.append("\n**Errors:**")
                for error in result['errors']:
                    status_parts.append(f"- {error['file']}: {error['error']}")
            
            # Add index stats
            if result.get('index_stats'):
                stats = result['index_stats']
                status_parts.append(f"\n**Index stats:** {stats['total_chunks']} chunks, {stats['dimension']}D embeddings")
            
            return "\n".join(status_parts), current_rag_tool
        else:
            return f"❌ {result['message']}", current_rag_tool
            
    except Exception as e:
        return f"❌ Error processing documents: {str(e)}", current_rag_tool

def update_sandbox_preview(config_data):
    """Update the sandbox preview with generated content"""
    if not config_data:
        return "Generate a space configuration to see preview here.", "<div style='text-align: center; padding: 50px; color: #666;'>No preview available</div>"
    
    # Create preview info
    preview_text = f"""**Space Configuration:**
- **Name:** {config_data.get('name', 'N/A')}
- **Model:** {config_data.get('model', 'N/A')}
- **Temperature:** {config_data.get('temperature', 'N/A')}
- **Max Tokens:** {config_data.get('max_tokens', 'N/A')}
- **Dynamic URLs:** {'✅ Enabled' if config_data.get('enable_dynamic_urls') else '❌ Disabled'}
- **Vector RAG:** {'✅ Enabled' if config_data.get('enable_vector_rag') else '❌ Disabled'}

**System Prompt Preview:**
```
{config_data.get('system_prompt', 'No system prompt configured')[:500]}{'...' if len(config_data.get('system_prompt', '')) > 500 else ''}
```

**Deployment Package:** `{config_data.get('filename', 'Not generated')}`"""
    
    # Create a basic HTML preview of the chat interface
    preview_html = f"""
    <div style="border: 1px solid #ddd; border-radius: 8px; padding: 20px; background: #f9f9f9;">
        <h3 style="margin-top: 0; color: #333;">{config_data.get('name', 'Chat Interface')}</h3>
        <p style="color: #666; margin-bottom: 20px;">{config_data.get('description', 'A customizable AI chat interface')}</p>
        
        <div style="border: 1px solid #ccc; border-radius: 4px; background: white; min-height: 200px; padding: 15px; margin-bottom: 15px;">
            <div style="color: #888; text-align: center; padding: 50px 0;">Chat Interface Preview</div>
            <div style="background: #f0f8ff; padding: 10px; border-radius: 4px; margin-bottom: 10px; border-left: 3px solid #0066cc;">
                <strong>Assistant:</strong> Hello! I'm ready to help you. How can I assist you today?
            </div>
        </div>
        
        <div style="border: 1px solid #ccc; border-radius: 4px; padding: 10px; background: white;">
            <input type="text" placeholder="Type your message here..." style="width: 70%; padding: 8px; border: 1px solid #ddd; border-radius: 4px; margin-right: 10px;">
            <button style="padding: 8px 15px; background: #0066cc; color: white; border: none; border-radius: 4px; cursor: pointer;">Send</button>
        </div>
        
        <div style="margin-top: 15px; padding: 10px; background: #f0f0f0; border-radius: 4px; font-size: 12px; color: #666;">
            <strong>Configuration:</strong> Model: {config_data.get('model', 'N/A')} | Temperature: {config_data.get('temperature', 'N/A')} | Max Tokens: {config_data.get('max_tokens', 'N/A')}
        </div>
    </div>
    """
    
    return preview_text, preview_html

def on_preview_combined(name, description, system_prompt, enable_research_assistant, role_purpose, intended_audience, key_tasks, additional_context, model, temperature, max_tokens, examples_text, enable_dynamic_urls, enable_vector_rag):
    """Generate configuration and return preview updates"""
    if not name or not name.strip():
        return (
            {},
            gr.update(value="**Error:** Please provide a Space Title to preview", visible=True),
            gr.update(visible=False),
            gr.update(value="Configuration will appear here after preview generation.")
        )
    
    try:
        # Combine system prompt components if research assistant is enabled
        if enable_research_assistant:
            if not role_purpose or not role_purpose.strip():
                return (
                    {},
                    gr.update(value="**Error:** Please provide a Role and Purpose for the research assistant", visible=True),
                    gr.update(visible=False),
                    gr.update(value="Configuration will appear here after preview generation.")
                )
            system_prompt_parts = []
            if role_purpose and role_purpose.strip():
                system_prompt_parts.append(role_purpose.strip())
            if intended_audience and intended_audience.strip():
                system_prompt_parts.append(intended_audience.strip())
            if key_tasks and key_tasks.strip():
                system_prompt_parts.append(key_tasks.strip())
            if additional_context and additional_context.strip():
                system_prompt_parts.append(additional_context.strip())
            
            final_system_prompt = " ".join(system_prompt_parts)
        else:
            if not system_prompt or not system_prompt.strip():
                return (
                    {},
                    gr.update(value="**Error:** Please provide a System Prompt for the assistant", visible=True),
                    gr.update(visible=False),
                    gr.update(value="Configuration will appear here after preview generation.")
                )
            final_system_prompt = system_prompt.strip()
        
        # Create configuration for preview
        config_data = {
            'name': name,
            'description': description,
            'system_prompt': final_system_prompt,
            'model': model,
            'temperature': temperature,
            'max_tokens': max_tokens,
            'enable_dynamic_urls': enable_dynamic_urls,
            'enable_vector_rag': enable_vector_rag,
            'examples_text': examples_text,
            'preview_ready': True
        }
        
        # Generate preview displays
        preview_text = f"""🎉 **Preview Successfully Rendered!** 

Your assistant "{name}" is now configured and ready to test in the Sandbox Preview tab.

**Configuration:**
- **Model:** {model}
- **Temperature:** {temperature}
- **Max Tokens:** {max_tokens}
- **Dynamic URLs:** {'✅ Enabled' if enable_dynamic_urls else '❌ Disabled'}
- **Vector RAG:** {'✅ Enabled' if enable_vector_rag else '❌ Disabled'}

**System Prompt:** 
{final_system_prompt[:200]}{'...' if len(final_system_prompt) > 200 else ''}

✨ **Next Steps:** Switch to the "Sandbox Preview" tab to test your assistant with real conversations before generating the deployment package."""
        
        config_display = f"""### Current Configuration

**Space Details:**
- **Name:** {name}
- **Description:** {description or 'No description provided'}

**Model Settings:**
- **Model:** {model}
- **Temperature:** {temperature}
- **Max Response Tokens:** {max_tokens}

**Features:**
- **Dynamic URL Fetching:** {'✅ Enabled' if enable_dynamic_urls else '❌ Disabled'}
- **Document RAG:** {'✅ Enabled' if enable_vector_rag else '❌ Disabled'}

**System Prompt:**
```
{final_system_prompt}
```

**Example Prompts:**
{examples_text if examples_text and examples_text.strip() else 'No example prompts configured'}
"""
        
        return (
            config_data,
            gr.update(value=preview_text, visible=True),
            gr.update(visible=True),
            gr.update(value=config_display)
        )
        
    except Exception as e:
        return (
            {},
            gr.update(value=f"**Error:** {str(e)}", visible=True),
            gr.update(visible=False),
            gr.update(value="Configuration will appear here after preview generation.")
        )

def update_preview_display(config_data):
    """Update preview display based on config data"""
    if not config_data or not config_data.get('preview_ready'):
        return (
            gr.update(value="**Status:** Configure your space in the Configuration tab and click 'Preview Deployment Package' to see your assistant here.", visible=True),
            gr.update(visible=False),
            gr.update(value="Configuration will appear here after preview generation.")
        )
    
    preview_text = f"""**Preview Ready!** 
    
Your assistant "{config_data['name']}" is configured and ready to test.

**Configuration:**
- **Model:** {config_data['model']}
- **Temperature:** {config_data['temperature']}
- **Max Tokens:** {config_data['max_tokens']}
- **Dynamic URLs:** {'✅ Enabled' if config_data['enable_dynamic_urls'] else '❌ Disabled'}
- **Vector RAG:** {'✅ Enabled' if config_data['enable_vector_rag'] else '❌ Disabled'}

**System Prompt:** 
{config_data['system_prompt'][:200]}{'...' if len(config_data['system_prompt']) > 200 else ''}

Use the chat interface below to test your assistant before generating the deployment package."""
    
    config_display = f"""### Current Configuration

**Space Details:**
- **Name:** {config_data['name']}
- **Description:** {config_data.get('description', 'No description provided')}

**Model Settings:**
- **Model:** {config_data['model']}
- **Temperature:** {config_data['temperature']}
- **Max Response Tokens:** {config_data['max_tokens']}

**Features:**
- **Dynamic URL Fetching:** {'✅ Enabled' if config_data['enable_dynamic_urls'] else '❌ Disabled'}
- **Document RAG:** {'✅ Enabled' if config_data['enable_vector_rag'] else '❌ Disabled'}

**System Prompt:**
```
{config_data['system_prompt']}
```

**Example Prompts:**
{config_data.get('examples_text', 'No example prompts configured') if config_data.get('examples_text', '').strip() else 'No example prompts configured'}
"""
    
    return (
        gr.update(value=preview_text, visible=True),
        gr.update(visible=True),
        gr.update(value=config_display)
    )

def preview_chat_response(message, history, config_data, url1="", url2="", url3="", url4=""):
    """Generate response for preview chat using actual OpenRouter API"""
    if not config_data or not message:
        return "", history
    
    # Get API key from environment
    api_key = os.environ.get("OPENROUTER_API_KEY")
    
    if not api_key:
        response = f"[Preview Mode - No API Key] I'm {config_data.get('name', 'your assistant')} running on {config_data.get('model', 'unknown model')}. To test with real API responses, set your OPENROUTER_API_KEY in the environment. This preview would use your system prompt: {config_data.get('system_prompt', '')[:100]}..."
        history.append([message, response])
        return "", history
    
    try:
        # Get grounding context from URLs if configured
        grounding_urls = [url1, url2, url3, url4]
        grounding_context = get_cached_grounding_context([url for url in grounding_urls if url and url.strip()])
        
        # Add RAG context if available (simplified for preview)
        rag_context = ""
        if config_data.get('enable_vector_rag'):
            rag_context = "\n\n[RAG context would be retrieved here based on similarity search]\n\n"
        
        # If dynamic URLs are enabled, check message for URLs to fetch
        dynamic_context = ""
        if config_data.get('enable_dynamic_urls'):
            urls_in_message = extract_urls_from_text(message)
            if urls_in_message:
                dynamic_context_parts = []
                for url in urls_in_message[:2]:  # Limit to 2 URLs in preview
                    content = fetch_url_content(url)
                    dynamic_context_parts.append(f"\n\nDynamic context from {url}:\n{content}")
                if dynamic_context_parts:
                    dynamic_context = "\n".join(dynamic_context_parts)
        
        # Build enhanced system prompt with all contexts
        enhanced_system_prompt = config_data.get('system_prompt', '') + grounding_context + rag_context + dynamic_context
        
        # Build messages array for the API
        messages = [{"role": "system", "content": enhanced_system_prompt}]
        
        # Add conversation history - handle both formats
        for chat in history:
            if isinstance(chat, list) and len(chat) >= 2:
                # Legacy format: [user_msg, assistant_msg]
                user_msg, assistant_msg = chat[0], chat[1]
                if user_msg:
                    messages.append({"role": "user", "content": user_msg})
                if assistant_msg:
                    messages.append({"role": "assistant", "content": assistant_msg})
        
        # Add current message
        messages.append({"role": "user", "content": message})
        
        # Debug: Log the request being sent
        request_payload = {
            "model": config_data.get('model', 'google/gemini-2.0-flash-001'),
            "messages": messages,
            "temperature": config_data.get('temperature', 0.7),
            "max_tokens": config_data.get('max_tokens', 500)
        }
        
        # Make API request to OpenRouter
        response = requests.post(
            url="https://openrouter.ai/api/v1/chat/completions",
            headers={
                "Authorization": f"Bearer {api_key}",
                "Content-Type": "application/json"
            },
            json=request_payload,
            timeout=30
        )
        
        if response.status_code == 200:
            try:
                response_data = response.json()
                
                # Check if response has expected structure
                if 'choices' not in response_data or not response_data['choices']:
                    assistant_response = f"[Preview Debug] No choices in API response. Response: {response_data}"
                elif 'message' not in response_data['choices'][0]:
                    assistant_response = f"[Preview Debug] No message in first choice. Response: {response_data}"
                elif 'content' not in response_data['choices'][0]['message']:
                    assistant_response = f"[Preview Debug] No content in message. Response: {response_data}"
                else:
                    assistant_content = response_data['choices'][0]['message']['content']
                    
                    # Debug: Check if content is empty
                    if not assistant_content or assistant_content.strip() == "":
                        assistant_response = f"[Preview Debug] Empty content from API. Messages sent: {len(messages)} messages, last user message: '{message}', model: {request_payload['model']}"
                    else:
                        # Add preview indicator
                        assistant_response = f"[Preview Mode] {assistant_content}"
                        
            except (KeyError, IndexError, json.JSONDecodeError) as e:
                assistant_response = f"[Preview Error] Failed to parse API response: {str(e)}. Raw response: {response.text[:500]}"
        else:
            assistant_response = f"[Preview Error] API Error: {response.status_code} - {response.text[:500]}"
            
    except Exception as e:
        assistant_response = f"[Preview Error] {str(e)}"
    
    # Return in the legacy tuple format that Gradio ChatInterface expects
    history.append([message, assistant_response])
    return "", history

def clear_preview_chat():
    """Clear preview chat"""
    return "", []

def export_preview_conversation(history):
    """Export preview conversation to markdown"""
    if not history:
        return gr.update(visible=False)
    
    markdown_content = export_conversation_to_markdown(history)
    
    # Save to temporary file
    import tempfile
    with tempfile.NamedTemporaryFile(mode='w', suffix='.md', delete=False) as f:
        f.write(markdown_content)
        temp_file = f.name
    
    return gr.update(value=temp_file, visible=True)

def on_generate(name, description, system_prompt, enable_research_assistant, model, api_key_var, temperature, max_tokens, examples_text, access_code, enable_dynamic_urls, url1, url2, url3, url4, enable_vector_rag, rag_tool_state):
    if not name or not name.strip():
        return gr.update(value="Error: Please provide a Space Title", visible=True), gr.update(visible=False)
    
    
    try:
        # Get RAG data if enabled
        rag_data = None
        if enable_vector_rag and rag_tool_state:
            rag_data = rag_tool_state.get_serialized_data()
        
        # Combine system prompt components if research assistant is enabled
        if enable_research_assistant:
            # Use the research assistant fields if enabled
            if not role_purpose or not role_purpose.strip():
                return gr.update(value="Error: Please provide a Role and Purpose for the research assistant", visible=True), gr.update(visible=False)
            system_prompt_parts = []
            if role_purpose and role_purpose.strip():
                system_prompt_parts.append(role_purpose.strip())
            if intended_audience and intended_audience.strip():
                system_prompt_parts.append(intended_audience.strip())
            if key_tasks and key_tasks.strip():
                system_prompt_parts.append(key_tasks.strip())
            if additional_context and additional_context.strip():
                system_prompt_parts.append(additional_context.strip())
            
            final_system_prompt = " ".join(system_prompt_parts)
        else:
            # Use the direct system prompt field
            if not system_prompt or not system_prompt.strip():
                return gr.update(value="Error: Please provide a System Prompt for the assistant", visible=True), gr.update(visible=False)
            final_system_prompt = system_prompt.strip()
        
        filename = generate_zip(name, description, final_system_prompt, model, api_key_var, temperature, max_tokens, examples_text, access_code, enable_dynamic_urls, url1, url2, url3, url4, enable_vector_rag, rag_data)
        
        success_msg = f"""**Deployment package ready!**

**File**: `{filename}`

**What's included:**
- `app.py` - Ready-to-deploy chat interface
- `requirements.txt` - Dependencies  
- `README.md` - Step-by-step deployment guide
- `config.json` - Configuration backup

**Next steps:**
1. Download the zip file below
2. Follow the README instructions to deploy on HuggingFace Spaces
3. Set your `{api_key_var}` secret in Space settings

**Your Space will be live in minutes!**"""
        
        # Update sandbox preview
        config_data = {
            'name': name,
            'description': description,
            'system_prompt': final_system_prompt,
            'model': model,
            'temperature': temperature,
            'max_tokens': max_tokens,
            'enable_dynamic_urls': enable_dynamic_urls,
            'enable_vector_rag': enable_vector_rag,
            'filename': filename
        }
        
        return gr.update(value=success_msg, visible=True), gr.update(value=filename, visible=True), config_data
        
    except Exception as e:
        return gr.update(value=f"Error: {str(e)}", visible=True), gr.update(visible=False)

# Global cache for URL content to avoid re-crawling
url_content_cache = {}

def get_cached_grounding_context(urls):
    """Get grounding context with caching to avoid re-crawling same URLs"""
    if not urls:
        return ""
    
    # Filter valid URLs
    valid_urls = [url for url in urls if url and url.strip()]
    if not valid_urls:
        return ""
    
    # Create cache key from sorted URLs
    cache_key = tuple(sorted(valid_urls))
    
    # Check if we already have this content cached
    if cache_key in url_content_cache:
        return url_content_cache[cache_key]
    
    # If not cached, fetch using simple HTTP requests
    grounding_context = get_grounding_context_simple(valid_urls)
    
    # Cache the result
    url_content_cache[cache_key] = grounding_context
    
    return grounding_context

def respond_with_cache_update(message, chat_history, url1="", url2="", url3="", url4=""):
    """Wrapper that updates cache status after responding"""
    msg, history = respond(message, chat_history, url1, url2, url3, url4)
    cache_status = get_cache_status()
    return msg, history, cache_status

def respond(message, chat_history, url1="", url2="", url3="", url4=""):
    # Make actual API request to OpenRouter
    import os
    import requests
    
    # Get API key from environment
    api_key = os.environ.get("OPENROUTER_API_KEY")
    
    if not api_key:
        response = "Please set your OPENROUTER_API_KEY in the Space settings to use the chat support."
        chat_history.append({"role": "user", "content": message})
        chat_history.append({"role": "assistant", "content": response})
        return "", chat_history
    
    # Get grounding context from URLs using cached approach
    grounding_urls = [url1, url2, url3, url4]
    grounding_context = get_cached_grounding_context(grounding_urls)
    
    # Build enhanced system prompt with grounding context
    base_system_prompt = """You are an expert assistant specializing in Gradio configurations for HuggingFace Spaces. You have deep knowledge of:
- Gradio interface components and layouts
- HuggingFace Spaces configuration (YAML frontmatter, secrets, environment variables)
- Deployment best practices for Gradio apps on HuggingFace
- Space settings, SDK versions, and hardware requirements
- Troubleshooting common Gradio and HuggingFace Spaces issues
- Integration with various APIs and models through Gradio interfaces

Provide specific, technical guidance focused on Gradio implementation details and HuggingFace Spaces deployment. Include code examples when relevant. Keep responses concise and actionable."""
    
    enhanced_system_prompt = base_system_prompt + grounding_context
    
    # Build conversation history for API
    messages = [{
        "role": "system", 
        "content": enhanced_system_prompt
    }]
    
    # Add conversation history - Support both new messages format and legacy tuple format
    for chat in chat_history:
        if isinstance(chat, dict):
            # New format: {"role": "user", "content": "..."}
            messages.append(chat)
        elif isinstance(chat, (list, tuple)) and len(chat) >= 2:
            # Legacy format: ("user msg", "bot msg")
            user_msg, assistant_msg = chat[0], chat[1]
            if user_msg:
                messages.append({"role": "user", "content": user_msg})
            if assistant_msg:
                messages.append({"role": "assistant", "content": assistant_msg})
    
    # Add current message
    messages.append({"role": "user", "content": message})
    
    try:
        # Make API request to OpenRouter
        response = requests.post(
            url="https://openrouter.ai/api/v1/chat/completions",
            headers={
                "Authorization": f"Bearer {api_key}",
                "Content-Type": "application/json"
            },
            json={
                "model": "google/gemini-2.0-flash-001",
                "messages": messages,
                "temperature": 0.7,
                "max_tokens": 500
            }
        )
        
        if response.status_code == 200:
            assistant_response = response.json()['choices'][0]['message']['content']
        else:
            assistant_response = f"Error: {response.status_code} - {response.text}"
            
    except Exception as e:
        assistant_response = f"Error: {str(e)}"
    
    chat_history.append({"role": "user", "content": message})
    chat_history.append({"role": "assistant", "content": assistant_response})
    return "", chat_history

def clear_chat():
    return "", []

def clear_url_cache():
    """Clear the URL content cache"""
    global url_content_cache
    url_content_cache.clear()
    return "✅ URL cache cleared. Next request will re-fetch content."

def get_cache_status():
    """Get current cache status"""
    if not url_content_cache:
        return "🔄 No URLs cached"
    return f"💾 {len(url_content_cache)} URL set(s) cached"

def add_urls(count):
    """Show additional URL fields"""
    if count == 2:
        return (gr.update(visible=True), gr.update(visible=False), 
                gr.update(value="+ Add URLs"), gr.update(visible=True), 3)
    elif count == 3:
        return (gr.update(visible=True), gr.update(visible=True), 
                gr.update(value="Max URLs", interactive=False), gr.update(visible=True), 4)
    else:
        return (gr.update(), gr.update(), gr.update(), gr.update(), count)

def remove_urls(count):
    """Hide URL fields"""
    if count == 4:
        return (gr.update(visible=True), gr.update(visible=False, value=""), 
                gr.update(value="+ Add URLs", interactive=True), gr.update(visible=True), 3)
    elif count == 3:
        return (gr.update(visible=False, value=""), gr.update(visible=False, value=""), 
                gr.update(value="+ Add URLs", interactive=True), gr.update(visible=False), 2)
    else:
        return (gr.update(), gr.update(), gr.update(), gr.update(), count)

def add_chat_urls(count):
    """Show additional chat URL fields"""
    if count == 2:
        return (gr.update(visible=True), gr.update(visible=False), 
                gr.update(value="+ Add URLs"), gr.update(visible=True), 3)
    elif count == 3:
        return (gr.update(visible=True), gr.update(visible=True), 
                gr.update(value="Max URLs", interactive=False), gr.update(visible=True), 4)
    else:
        return (gr.update(), gr.update(), gr.update(), gr.update(), count)

def remove_chat_urls(count):
    """Hide chat URL fields"""
    if count == 4:
        return (gr.update(visible=True), gr.update(visible=False, value=""), 
                gr.update(value="+ Add URLs", interactive=True), gr.update(visible=True), 3)
    elif count == 3:
        return (gr.update(visible=False, value=""), gr.update(visible=False, value=""), 
                gr.update(value="+ Add URLs", interactive=True), gr.update(visible=False), 2)
    else:
        return (gr.update(), gr.update(), gr.update(), gr.update(), count)


def toggle_research_assistant(enable_research):
    """Toggle research assistant system prompt"""
    if enable_research:
        combined_prompt = "You are a search tool that provides link-grounded information through web fetching, limiting source criteria to DOI-verified articles from academic databases and official sources. Use https://libkey.io/ to cross-reference and validate article DOIs for inclusion. This tool is designed for students and researchers conducting academic inquiry. Additional responsibilities include analyzing academic sources, fact-checking claims with evidence, providing properly cited research summaries, and helping users navigate scholarly information. Ground all responses in provided URL contexts and any additional URLs you're instructed to fetch. Never rely on memory for factual claims."
        return (
            gr.update(value=combined_prompt),  # Update main system prompt
            gr.update(value=True)  # Enable dynamic URL fetching for research template
        )
    else:
        return (
            gr.update(value=""),  # Clear main system prompt when disabling
            gr.update(value=False)  # Disable dynamic URL setting
        )


# Create Gradio interface with proper tab structure
with gr.Blocks(title="Chat U/I Helper") as demo:
    # Global state for cross-tab functionality
    sandbox_state = gr.State({})
    preview_config_state = gr.State({})
    
    # Global status components that will be defined later
    preview_status = None
    preview_chat_section = None 
    config_display = None
    
    with gr.Tabs():
        with gr.Tab("Configuration"):
            gr.Markdown("# Spaces Configuration")
            gr.Markdown("Convert custom assistants from HuggingChat into chat interfaces with HuggingFace Spaces. Configure and download everything needed to deploy a simple HF space using Gradio.")
    
            with gr.Column():
                name = gr.Textbox(
                    label="Space Title",
                    placeholder="My Course Helper",
                    value="My Custom Space"
                )
                
                description = gr.Textbox(
                    label="Space Description", 
                    placeholder="A customizable AI chat interface for...",
                    lines=2,
                    value=""
                )
                
                model = gr.Dropdown(
                    label="Model",
                    choices=MODELS,
                    value=MODELS[0],
                    info="Choose based on the context and purposes of your space"
                )
                
                api_key_var = gr.Textbox(
                    label="API Key Variable Name",
                    value="OPENROUTER_API_KEY",
                    info="Name for the secret in HuggingFace Space settings"
                )
                
                access_code = gr.Textbox(
                    label="Access Code (Optional)",
                    placeholder="Leave empty for public access, or enter code for student access",
                    info="If set, students must enter this code to access the chatbot",
                    type="password"
                )
                
                with gr.Accordion("Assistant Configuration", open=True):
                    gr.Markdown("### Configure your assistant's behavior and capabilities")
                    gr.Markdown("Define the system prompt and assistant settings. You can use pre-configured templates or custom fields.")

                    # Main system prompt field - always visible
                    system_prompt = gr.Textbox(
                        label="System Prompt",
                        placeholder="You are a helpful assistant that...",
                        lines=4,
                        value="",
                        info="Define the assistant's role, purpose, and behavior in a single prompt"
                    )
                    
                    # Assistant configuration options
                    enable_research_assistant = gr.Checkbox(
                        label="Research Template",
                        value=False,
                        info="Enable to use pre-configured research assistant settings"
                    )
                    
                    
                
                examples_text = gr.Textbox(
                    label="Example Prompts (one per line)",
                    placeholder="Can you analyze this research paper: https://example.com/paper.pdf\nWhat are the latest findings on climate change adaptation?\nHelp me fact-check claims about renewable energy efficiency",
                    lines=3,
                    info="These will appear as clickable examples in the chat interface"
                )
                
                with gr.Accordion("Tool Settings", open=False):

                    enable_dynamic_urls = gr.Checkbox(
                        label="Enable Dynamic URL Fetching",
                        value=False,
                        info="Allow the assistant to fetch additional URLs mentioned in conversations"
                    )
                    
                    enable_vector_rag = gr.Checkbox(
                        label="Enable Document RAG",
                        value=False,
                        info="Upload documents for context-aware responses (PDF, DOCX, TXT, MD)",
                        visible=HAS_RAG
                    )
                    
                    with gr.Column(visible=False) as rag_section:
                        gr.Markdown("### Document Upload")
                        file_upload = gr.File(
                            label="Upload Documents",
                            file_types=[".pdf", ".docx", ".txt", ".md"],
                            file_count="multiple"
                        )
                        process_btn = gr.Button("Process Documents", variant="secondary")
                        rag_status = gr.Markdown()
                        
                        # State to store RAG tool
                        rag_tool_state = gr.State(None)
                    
                    with gr.Accordion("URL Grounding (Optional)", open=False):
                        gr.Markdown("Add URLs to provide context. Content will be fetched and added to the system prompt.")
                        
                        # Initial URL fields
                        url1 = gr.Textbox(
                            label="URL 1",
                            placeholder="https://example.com/page1",
                            info="First URL for context grounding"
                        )
                        
                        url2 = gr.Textbox(
                            label="URL 2", 
                            placeholder="https://example.com/page2",
                            info="Second URL for context grounding"
                        )
                        
                        # Additional URL fields (initially hidden)
                        url3 = gr.Textbox(
                            label="URL 3",
                            placeholder="https://example.com/page3", 
                            info="Third URL for context grounding",
                            visible=False
                        )
                        
                        url4 = gr.Textbox(
                            label="URL 4",
                            placeholder="https://example.com/page4",
                            info="Fourth URL for context grounding",
                            visible=False
                        )
                        
                        # URL management buttons
                        with gr.Row():
                            add_url_btn = gr.Button("+ Add URLs", size="sm")
                            remove_url_btn = gr.Button("- Remove URLs", size="sm", visible=False)
                        url_count = gr.State(2)  # Track number of visible URLs
                
                with gr.Accordion("Advanced Settings", open=False):
                    with gr.Row():
                        temperature = gr.Slider(
                            label="Temperature",
                            minimum=0,
                            maximum=2,
                            value=0.7,
                            step=0.1,
                            info="Higher = more creative, Lower = more focused"
                        )
                        
                        max_tokens = gr.Slider(
                            label="Max Response Tokens",
                            minimum=50,
                            maximum=4096,
                            value=500,
                            step=50
                        )
                
                with gr.Row():
                    preview_btn = gr.Button("Preview Deployment Package", variant="secondary")
                    generate_btn = gr.Button("Generate Deployment Package", variant="primary")
                
                status = gr.Markdown(visible=False)
                download_file = gr.File(label="Download your zip package", visible=False)
            
            
            
            # Connect the research assistant checkbox
            enable_research_assistant.change(
                toggle_research_assistant,
                inputs=[enable_research_assistant],
                outputs=[system_prompt, enable_dynamic_urls]
            )
            
            
            
            # Connect the URL management buttons
            add_url_btn.click(
                add_urls,
                inputs=[url_count],
                outputs=[url3, url4, add_url_btn, remove_url_btn, url_count]
            )
            
            remove_url_btn.click(
                remove_urls,
                inputs=[url_count],
                outputs=[url3, url4, add_url_btn, remove_url_btn, url_count]
            )
            
            # Connect RAG functionality
            enable_vector_rag.change(
                toggle_rag_section,
                inputs=[enable_vector_rag],
                outputs=[rag_section]
            )
            
            process_btn.click(
                process_documents,
                inputs=[file_upload, rag_tool_state],
                outputs=[rag_status, rag_tool_state]
            )
            
            
            # Connect the generate button
            generate_btn.click(
                on_generate,
                inputs=[name, description, system_prompt, enable_research_assistant, model, api_key_var, temperature, max_tokens, examples_text, access_code, enable_dynamic_urls, url1, url2, url3, url4, enable_vector_rag, rag_tool_state],
                outputs=[status, download_file, sandbox_state]
            )
            
        
        with gr.Tab("Preview"):
            gr.Markdown("# Sandbox Preview")
            gr.Markdown("Preview your generated assistant exactly as it will appear in the deployed HuggingFace Space.")
            
            with gr.Column():
                # Preview status - assign to global variable
                preview_status_comp = gr.Markdown("**Status:** Configure your space in the Configuration tab and click 'Preview Deployment Package' to see your assistant here.", visible=True)
                
                # Simulated chat interface for preview
                with gr.Column(visible=False) as preview_chat_section_comp:
                    preview_chatbot = gr.Chatbot(
                        value=[],
                        label="Preview Chat Interface", 
                        height=400
                    )
                    preview_msg = gr.Textbox(
                        label="Test your assistant",
                        placeholder="Type a message to test your assistant...",
                        lines=2
                    )
                    
                    # URL context fields for preview testing
                    with gr.Accordion("Test URL Context (Optional)", open=False):
                        gr.Markdown("Add URLs to test context grounding in the preview")
                        
                        with gr.Row():
                            preview_url1 = gr.Textbox(
                                label="URL 1",
                                placeholder="https://example.com/page1",
                                scale=1
                            )
                            preview_url2 = gr.Textbox(
                                label="URL 2", 
                                placeholder="https://example.com/page2",
                                scale=1
                            )
                        
                        with gr.Row():
                            preview_url3 = gr.Textbox(
                                label="URL 3",
                                placeholder="https://example.com/page3",
                                scale=1,
                                visible=False
                            )
                            preview_url4 = gr.Textbox(
                                label="URL 4",
                                placeholder="https://example.com/page4",
                                scale=1,
                                visible=False
                            )
                        
                        # URL management for preview
                        with gr.Row():
                            preview_add_url_btn = gr.Button("+ Add URLs", size="sm")
                            preview_remove_url_btn = gr.Button("- Remove URLs", size="sm", visible=False)
                        preview_url_count = gr.State(2)
                    
                    with gr.Row():
                        preview_send = gr.Button("Send", variant="primary")
                        preview_clear = gr.Button("Clear")
                        export_btn = gr.Button("Export Conversation", variant="secondary")
                    
                    # Export functionality
                    export_file = gr.File(label="Download Conversation", visible=False)
                
                # Configuration display - assign to global variable
                config_display_comp = gr.Markdown("Configuration will appear here after preview generation.")
                
            
            
            # Connect preview chat functionality
            preview_send.click(
                preview_chat_response,
                inputs=[preview_msg, preview_chatbot, preview_config_state, preview_url1, preview_url2, preview_url3, preview_url4],
                outputs=[preview_msg, preview_chatbot]
            )
            
            preview_msg.submit(
                preview_chat_response,
                inputs=[preview_msg, preview_chatbot, preview_config_state, preview_url1, preview_url2, preview_url3, preview_url4],
                outputs=[preview_msg, preview_chatbot]
            )
            
            preview_clear.click(
                clear_preview_chat,
                outputs=[preview_msg, preview_chatbot]
            )
            
            export_btn.click(
                export_preview_conversation,
                inputs=[preview_chatbot],
                outputs=[export_file]
            )
            
            # Connect preview URL management buttons
            preview_add_url_btn.click(
                add_chat_urls,
                inputs=[preview_url_count],
                outputs=[preview_url3, preview_url4, preview_add_url_btn, preview_remove_url_btn, preview_url_count]
            )
            
            preview_remove_url_btn.click(
                remove_chat_urls,
                inputs=[preview_url_count],
                outputs=[preview_url3, preview_url4, preview_add_url_btn, preview_remove_url_btn, preview_url_count]
            )
        
        with gr.Tab("Support"):
            create_support_docs()
    
    # Connect cross-tab functionality after all components are defined
    preview_btn.click(
        on_preview_combined,
        inputs=[name, description, system_prompt, enable_research_assistant, role_purpose, intended_audience, key_tasks, additional_context, model, temperature, max_tokens, examples_text, enable_dynamic_urls, enable_vector_rag],
        outputs=[preview_config_state, preview_status_comp, preview_chat_section_comp, config_display_comp]
    )

if __name__ == "__main__":
    # Check if running in local development with dev tunnels
    if os.environ.get('CODESPACES') or 'devtunnels.ms' in os.environ.get('GRADIO_SERVER_NAME', ''):
        demo.launch(share=True, allowed_paths=[], server_name="0.0.0.0")
    else:
        demo.launch(share=True)