import os import re from http import HTTPStatus from typing import Dict, List, Optional, Tuple import base64 import mimetypes import PyPDF2 import docx import cv2 import numpy as np from PIL import Image import pytesseract import requests from urllib.parse import urlparse, urljoin from bs4 import BeautifulSoup import html2text import json import time import gradio as gr from huggingface_hub import InferenceClient, HfApi, create_repo from tavily import TavilyClient # Configuration SystemPrompt = """You are a helpful coding assistant. You help users create applications by generating code based on their requirements. When asked to create an application, you should: 1. Understand the user's requirements 2. Generate clean, working code 3. Provide HTML output when appropriate for web applications 4. Include necessary comments and documentation 5. Ensure the code is functional and follows best practices For website redesign tasks: - Use the provided original HTML code as the starting point for redesign - Preserve all original content, structure, and functionality - Keep the same semantic HTML structure but enhance the styling - Reuse all original images and their URLs from the HTML code - Create a modern, responsive design with improved typography and spacing - Use modern CSS frameworks and design patterns - Ensure accessibility and mobile responsiveness - Maintain the same navigation and user flow - Enhance the visual design while keeping the original layout structure If an image is provided, analyze it and use the visual information to better understand the user's requirements. Always respond with code that can be executed or rendered directly. Always output only the HTML code inside a ```html ... ``` code block, and do not include any explanations or extra text.""" # System prompt with search capability SystemPromptWithSearch = """You are a helpful coding assistant with access to real-time web search. You help users create applications by generating code based on their requirements. When asked to create an application, you should: 1. Understand the user's requirements 2. Use web search when needed to find the latest information, best practices, or specific technologies 3. Generate clean, working code 4. Provide HTML output when appropriate for web applications 5. Include necessary comments and documentation 6. Ensure the code is functional and follows best practices For website redesign tasks: - Use the provided original HTML code as the starting point for redesign - Preserve all original content, structure, and functionality - Keep the same semantic HTML structure but enhance the styling - Reuse all original images and their URLs from the HTML code - Use web search to find current design trends and best practices for the specific type of website - Create a modern, responsive design with improved typography and spacing - Use modern CSS frameworks and design patterns - Ensure accessibility and mobile responsiveness - Maintain the same navigation and user flow - Enhance the visual design while keeping the original layout structure If an image is provided, analyze it and use the visual information to better understand the user's requirements. Always respond with code that can be executed or rendered directly. Always output only the HTML code inside a ```html ... ``` code block, and do not include any explanations or extra text.""" # Available models AVAILABLE_MODELS = [ { "name": "DeepSeek V3", "id": "deepseek-ai/DeepSeek-V3-0324", "description": "DeepSeek V3 model for code generation" }, { "name": "DeepSeek R1", "id": "deepseek-ai/DeepSeek-R1-0528", "description": "DeepSeek R1 model for code generation" }, { "name": "ERNIE-4.5-VL", "id": "baidu/ERNIE-4.5-VL-424B-A47B-Base-PT", "description": "ERNIE-4.5-VL model for multimodal code generation with image support" }, { "name": "MiniMax M1", "id": "MiniMaxAI/MiniMax-M1-80k", "description": "MiniMax M1 model for code generation and general tasks" }, { "name": "Qwen3-235B-A22B", "id": "Qwen/Qwen3-235B-A22B", "description": "Qwen3-235B-A22B model for code generation and general tasks" }, { "name": "SmolLM3-3B", "id": "HuggingFaceTB/SmolLM3-3B", "description": "SmolLM3-3B model for code generation and general tasks" } ] DEMO_LIST = [ { "title": "Todo App", "description": "Create a simple todo application with add, delete, and mark as complete functionality" }, { "title": "Calculator", "description": "Build a basic calculator with addition, subtraction, multiplication, and division" }, { "title": "Weather Dashboard", "description": "Create a weather dashboard that displays current weather information" }, { "title": "Chat Interface", "description": "Build a chat interface with message history and user input" }, { "title": "E-commerce Product Card", "description": "Create a product card component for an e-commerce website" }, { "title": "Login Form", "description": "Build a responsive login form with validation" }, { "title": "Dashboard Layout", "description": "Create a dashboard layout with sidebar navigation and main content area" }, { "title": "Data Table", "description": "Build a data table with sorting and filtering capabilities" }, { "title": "Image Gallery", "description": "Create an image gallery with lightbox functionality and responsive grid layout" }, { "title": "UI from Image", "description": "Upload an image of a UI design and I'll generate the HTML/CSS code for it" }, { "title": "Extract Text from Image", "description": "Upload an image containing text and I'll extract and process the text content" }, { "title": "Website Redesign", "description": "Enter a website URL to extract its content and redesign it with a modern, responsive layout" } ] # HF Inference Client HF_TOKEN = os.getenv('HF_TOKEN') client = InferenceClient( provider="auto", api_key=HF_TOKEN, bill_to="huggingface" ) # Type definitions History = List[Tuple[str, str]] Messages = List[Dict[str, str]] # HF API Client for deployment hf_api = HfApi(token=HF_TOKEN) # Tavily Search Client TAVILY_API_KEY = os.getenv('TAVILY_API_KEY') tavily_client = None if TAVILY_API_KEY: try: tavily_client = TavilyClient(api_key=TAVILY_API_KEY) except Exception as e: print(f"Failed to initialize Tavily client: {e}") tavily_client = None def validate_oauth_scopes(oauth_token) -> Tuple[bool, List[str]]: """ Validate that the OAuth token has the required scopes for creating spaces Returns (is_valid, missing_scopes) """ required_scopes = ['read-repos', 'write-repos', 'manage-repos'] if not oauth_token: return False, required_scopes # Try to get scopes from the token object token_scopes = [] # Check different ways scopes might be stored if hasattr(oauth_token, 'scopes'): token_scopes = oauth_token.scopes elif hasattr(oauth_token, 'scope'): token_scopes = oauth_token.scope.split(' ') if oauth_token.scope else [] elif hasattr(oauth_token, 'permissions'): token_scopes = oauth_token.permissions else: # If we can't determine scopes, assume they're missing return False, required_scopes # Convert to list if it's a string if isinstance(token_scopes, str): token_scopes = token_scopes.split(' ') # Check for required scopes missing_scopes = [scope for scope in required_scopes if scope not in token_scopes] return len(missing_scopes) == 0, missing_scopes # Deployment functions def create_space_from_html(title: str, html_content: str, prompts: List[str] = None, user_token: str = None, username: str = None) -> Dict: """ Create a Hugging Face Space from generated HTML content """ try: # Use user token if provided, otherwise fall back to server token api_token = user_token if user_token else HF_TOKEN if not api_token: return { "success": False, "error": "No Hugging Face token available", "message": "Please log in with your Hugging Face account to deploy" } # Create API client with user token user_hf_api = HfApi(token=api_token) # Debug: Test the token by getting user info try: user_info = user_hf_api.whoami() print(f"Debug: User info from API: {user_info}") if not username: username = user_info.get('name', 'user') print(f"Debug: Final username: {username}") # Additional validation if not username or username == 'user': return { "success": False, "error": "Could not retrieve username from token", "message": "The authentication token is valid but username could not be retrieved. Please try logging in again." } except Exception as e: print(f"Debug: Could not get user info: {e}") return { "success": False, "error": f"Invalid or expired token: {str(e)}", "message": "Please check your Hugging Face token and try logging in again" } # Clean the title for use as repo name clean_title = re.sub(r'[^a-zA-Z0-9_-]', '-', title.lower()) clean_title = re.sub(r'-+', '-', clean_title).strip('-') # Add timestamp to ensure uniqueness timestamp = int(time.time()) repo_name = f"{username}/{clean_title}-{timestamp}" print(f"Debug: Attempting to create space: {repo_name}") print(f"Debug: Using token for user: {username}") # Create the space try: print(f"Debug: Creating space with parameters:") print(f" repo_id: {repo_name}") print(f" repo_type: space") print(f" space_sdk: static") print(f" private: False") print(f" exist_ok: False") print(f" token length: {len(api_token) if api_token else 0}") repo_url = user_hf_api.create_repo( repo_id=repo_name, repo_type="space", space_sdk="static", private=False, exist_ok=False, token=api_token ) print(f"Debug: Successfully created space: {repo_url}") except Exception as e: print(f"Debug: Failed to create space: {e}") print(f"Debug: Exception type: {type(e)}") print(f"Debug: Exception args: {e.args}") raise e # Prepare the HTML content with proper structure html_template = """ {title}

{title}

Generated with AnyCoder - AI Code Generator

{html_content}
""" full_html = html_template.format(title=title, html_content=html_content) # Check HTML content size html_size = len(full_html.encode('utf-8')) print(f"Debug: HTML content size: {html_size} bytes") if html_size > 1000000: # 1MB limit print(f"Debug: Warning - HTML content is large ({html_size} bytes)") # Truncate if too large if len(html_content) > 500000: # 500KB limit for content html_content = html_content[:500000] + "\n" full_html = html_template.format(title=title, html_content=html_content) print(f"Debug: HTML content truncated to {len(full_html.encode('utf-8'))} bytes") # Upload the HTML file print(f"Debug: Uploading index.html to {repo_name}") user_hf_api.upload_file( path_or_fileobj=full_html.encode('utf-8'), path_in_repo="index.html", repo_id=repo_name, repo_type="space", token=api_token ) print(f"Debug: Successfully uploaded index.html") # Create README.md with project info prompts_text = "".join([f"- {prompt}\n" for prompt in (prompts or [])]) readme_template = """# {title} This project was generated using [AnyCoder](https://huggingface.co/spaces/ahsenkhaliq/anycoder), an AI-powered code generator. ## About This is a static HTML application created by describing the requirements in plain English to an AI model. ## Generated Prompts {prompts_text} ## View Live Visit: https://huggingface.co/spaces/{repo_name} --- *Generated with ❤️ using AnyCoder* """ readme_content = readme_template.format(title=title, prompts_text=prompts_text, repo_name=repo_name) print(f"Debug: Uploading README.md to {repo_name}") user_hf_api.upload_file( path_or_fileobj=readme_content.encode('utf-8'), path_in_repo="README.md", repo_id=repo_name, repo_type="space", token=api_token ) print(f"Debug: Successfully uploaded README.md") return { "success": True, "space_url": f"https://huggingface.co/spaces/{repo_name}", "repo_name": repo_name, "message": f"Successfully created space: {title}" } except Exception as e: error_msg = str(e) # Handle specific API errors if "401" in error_msg or "Unauthorized" in error_msg: return { "success": False, "error": "Authentication failed - please log in again", "message": "Your token may have expired. Please log out and log back in." } elif "403" in error_msg or "Forbidden" in error_msg: return { "success": False, "error": "Permission denied - insufficient privileges", "message": "You may not have permission to create spaces. Please check your Hugging Face account settings." } elif "409" in error_msg or "Conflict" in error_msg: return { "success": False, "error": "Repository already exists", "message": "A space with this name already exists. Please try a different title." } elif "422" in error_msg or "Validation" in error_msg: return { "success": False, "error": "Invalid repository name", "message": "The space title contains invalid characters. Please use only letters, numbers, hyphens, and underscores." } else: return { "success": False, "error": error_msg, "message": f"Failed to create space: {error_msg}" } def deploy_to_space(title: str, html_content: str, history: History, oauth_profile: gr.OAuthProfile = None, oauth_token: gr.OAuthToken = None) -> Tuple[str, str]: """ Deploy the generated HTML to a Hugging Face Space """ if not title or not title.strip(): return "❌ Please enter a title for your space.", update_oauth_status(oauth_profile, oauth_token) if not html_content or not html_content.strip(): return "❌ No HTML content to deploy. Please generate some code first.", update_oauth_status(oauth_profile, oauth_token) # Check if user is authenticated if not oauth_profile or not oauth_token: return """❌ **Authentication Required** To deploy your application, you need to be logged in with your Hugging Face account. **How to log in:** 1. Click the "Sign in with Hugging Face" button in the sidebar 2. Authorize AnyCoder to access your Hugging Face account with the required permissions: - read-repos - write-repos - manage-repos 3. Try deploying again **Why login is required:** - Deployments are created under your own Hugging Face account - You need write permissions to create spaces - This ensures you own and can manage your deployed applications **Note:** Make sure to grant all the requested permissions when authorizing the application. --- *Please log in and try again.*""", update_oauth_status(oauth_profile, oauth_token) # Get user information from OAuth profile username = oauth_profile.name user_token = oauth_token.token # Debug: Print token info (without exposing the actual token) print(f"Debug: Username from OAuth: {username}") print(f"Debug: Token type: {type(user_token)}") print(f"Debug: Token length: {len(user_token) if user_token else 0}") print(f"Debug: Token starts with: {user_token[:10] if user_token else 'None'}...") # Additional debugging for OAuth components print(f"Debug: OAuth profile type: {type(oauth_profile)}") print(f"Debug: OAuth token type: {type(oauth_token)}") print(f"Debug: OAuth profile attributes: {dir(oauth_profile) if oauth_profile else 'None'}") print(f"Debug: OAuth token attributes: {dir(oauth_token) if oauth_token else 'None'}") # Check OAuth token scopes if available if hasattr(oauth_token, 'scopes'): print(f"Debug: OAuth token scopes: {oauth_token.scopes}") else: print("Debug: OAuth token scopes not available") # Check OAuth profile permissions if available if hasattr(oauth_profile, 'permissions'): print(f"Debug: OAuth profile permissions: {oauth_profile.permissions}") else: print("Debug: OAuth profile permissions not available") # Validate OAuth scopes scopes_valid, missing_scopes = validate_oauth_scopes(oauth_token) print(f"Debug: OAuth scopes valid: {scopes_valid}") print(f"Debug: Missing scopes: {missing_scopes}") if not scopes_valid: return f"""❌ **Missing OAuth Scopes** Your OAuth token is missing the required permissions to create spaces. **Missing Scopes:** {chr(10).join([f"- `{scope}`" for scope in missing_scopes])} **Required Scopes:** - `read-repos` - Read access to repositories - `write-repos` - Write access to create repositories - `manage-repos` - Manage repository settings **Steps to fix:** 1. **Logout**: Click the logout button in the sidebar 2. **Login Again**: Click "Sign in with Hugging Face" again 3. **Grant All Permissions**: When the authorization page appears, make sure to check ALL the requested permissions: - ✅ read-repos - ✅ write-repos - ✅ manage-repos 4. **Complete Authorization**: Click "Authorize" to complete the login 5. **Try Deploying**: Try deploying again **Important:** Make sure you see all three permissions checked on the authorization page before clicking "Authorize". --- *Please log in again with full permissions.*""", update_oauth_status(oauth_profile, oauth_token) # Validate token format if not user_token or not user_token.startswith('hf_'): return """❌ **Invalid Token Format** The OAuth token appears to be invalid or in the wrong format. Please try logging out and logging back in. **Steps to fix:** 1. Click the logout button in the sidebar 2. Click "Sign in with Hugging Face" again 3. Authorize the application with the required permissions (read-repos, write-repos, manage-repos) 4. Try deploying again **Note:** Make sure you authorize all the required permissions when logging in. --- *Please try logging in again.*""", update_oauth_status(oauth_profile, oauth_token) # Test the token by making a simple API call try: test_api = HfApi(token=user_token) user_info = test_api.whoami() print(f"Debug: Token test successful - user: {user_info.get('name', 'unknown')}") # Check if user has the necessary permissions by trying to list their repos try: repos = test_api.list_repos(author=username, token=user_token) print(f"Debug: User has {len(list(repos))} repositories") # Test if user can create repositories by checking their account type try: # Try to get user info to check account capabilities user_info = test_api.whoami() print(f"Debug: User info: {user_info}") # Check if user has pro account or sufficient permissions if user_info.get('type') == 'user': print("Debug: User account type confirmed") else: print(f"Debug: User account type: {user_info.get('type', 'unknown')}") except Exception as user_info_error: print(f"Debug: Could not get detailed user info: {user_info_error}") except Exception as repo_error: print(f"Debug: Could not list repos: {repo_error}") # Check if this is a scope/permission issue error_msg = str(repo_error).lower() if "403" in error_msg or "forbidden" in error_msg or "unauthorized" in error_msg: return """❌ **Insufficient Permissions** Your Hugging Face account doesn't have the necessary permissions to create spaces. This could be because: 1. **Missing OAuth Scopes**: You didn't grant all the required permissions during login 2. **Account Restrictions**: Your account has restrictions on creating repositories 3. **Token Scope Issues**: The OAuth token doesn't include the necessary scopes **Required Permissions:** - `read-repos` - Read access to repositories - `write-repos` - Write access to create repositories - `manage-repos` - Manage repository settings **Steps to fix:** 1. **Logout**: Click the logout button in the sidebar 2. **Login Again**: Click "Sign in with Hugging Face" again 3. **Grant All Permissions**: When the authorization page appears, make sure to check ALL the requested permissions: - ✅ read-repos - ✅ write-repos - ✅ manage-repos 4. **Complete Authorization**: Click "Authorize" to complete the login 5. **Try Deploying**: Try deploying again **Important:** Make sure you see all three permissions checked on the authorization page before clicking "Authorize". --- *Please log in again with full permissions.*""", update_oauth_status(oauth_profile, oauth_token) else: return f"""❌ **Repository Access Error** Error: {str(repo_error)} This could be due to: - Network connectivity issues - Hugging Face API temporary problems - Account-specific restrictions **Steps to fix:** 1. Check your internet connection 2. Try logging out and logging back in 3. Wait a few minutes and try again 4. If the problem persists, check your Hugging Face account settings --- *Please try again or contact support if the issue persists.*""", update_oauth_status(oauth_profile, oauth_token) except Exception as token_error: print(f"Debug: Token test failed: {token_error}") return """❌ **Token Validation Failed** The OAuth token could not be validated. This could be because: 1. The token has expired 2. The token doesn't have the right permissions 3. There's an issue with the authentication **Steps to fix:** 1. Click the logout button in the sidebar 2. Click "Sign in with Hugging Face" again 3. Authorize with all required permissions 4. Try deploying again --- *Please log in again.*""", update_oauth_status(oauth_profile, oauth_token) # Extract prompts from history prompts = [] for user_msg, _ in history: if isinstance(user_msg, str) and user_msg.strip(): prompts.append(user_msg.strip()) # Test if user can create repositories by attempting a test creation try: test_api = HfApi(token=user_token) # Try to create a test repository to verify permissions test_repo_name = f"{username}/test-permissions-{int(time.time())}" print(f"Debug: Testing repository creation with: {test_repo_name}") try: # Attempt to create a test repository test_repo_url = test_api.create_repo( repo_id=test_repo_name, repo_type="model", private=True, exist_ok=False, token=user_token ) print(f"Debug: Successfully created test repository: {test_repo_url}") # Clean up the test repository try: test_api.delete_repo(repo_id=test_repo_name, token=user_token) print(f"Debug: Successfully cleaned up test repository") except Exception as cleanup_error: print(f"Debug: Could not clean up test repository: {cleanup_error}") except Exception as test_create_error: print(f"Debug: Could not create test repository: {test_create_error}") error_msg = str(test_create_error).lower() if "403" in error_msg or "forbidden" in error_msg: return """❌ **Repository Creation Permission Denied** Your Hugging Face account doesn't have permission to create repositories. This could be because: 1. **Account Type**: Your account type may not allow repository creation 2. **Organization Restrictions**: If you're part of an organization, there may be restrictions 3. **Account Status**: Your account may be limited or suspended **Steps to fix:** 1. **Check Account Status**: Visit https://huggingface.co/settings/account to check your account status 2. **Verify Account Type**: Make sure your account allows repository creation 3. **Contact Support**: If you believe this is an error, contact Hugging Face support 4. **Try Different Account**: Consider using a different Hugging Face account **Note:** Free accounts should be able to create repositories. If you're having issues, it might be a temporary restriction. --- *Please check your account settings or try with a different account.*""", update_oauth_status(oauth_profile, oauth_token) else: return f"""❌ **Repository Creation Test Failed** Error: {str(test_create_error)} This could be due to: - Network connectivity issues - Hugging Face API temporary problems - Account-specific restrictions **Steps to fix:** 1. Check your internet connection 2. Wait a few minutes and try again 3. If the problem persists, check your Hugging Face account settings --- *Please try again or contact support if the issue persists.*""", update_oauth_status(oauth_profile, oauth_token) except Exception as test_error: print(f"Debug: Repository creation test failed: {test_error}") return f"""❌ **Permission Test Failed** Could not test repository creation permissions: {str(test_error)} **Steps to fix:** 1. Check your internet connection 2. Try logging out and logging back in 3. Wait a few minutes and try again --- *Please try again or contact support if the issue persists.*""", update_oauth_status(oauth_profile, oauth_token) # Use user's OAuth token to create space under their account result = create_space_from_html(title.strip(), html_content, prompts, user_token, username) if result["success"]: return f"""✅ **Successfully deployed!** **Space URL:** {result['space_url']} Your application is now live on Hugging Face Spaces under your account. You can share this URL with others to showcase your generated application. **What's next:** - Visit the space to see your application in action - Share the URL with friends and colleagues - Make modifications and redeploy as needed - Manage your space from your Hugging Face dashboard --- *Generated with ❤️ using AnyCoder*""", update_oauth_status(oauth_profile, oauth_token) else: return f"""❌ **Deployment failed** **Error:** {result['error']} **Possible solutions:** - Ensure you have write access to create spaces on your account - Try a different title (avoid special characters) - Check your internet connection - Make sure your Hugging Face account has the necessary permissions --- *Please try again or contact support if the issue persists.*""", update_oauth_status(oauth_profile, oauth_token) def history_to_messages(history: History, system: str) -> Messages: messages = [{'role': 'system', 'content': system}] for h in history: # Handle multimodal content in history user_content = h[0] if isinstance(user_content, list): # Extract text from multimodal content text_content = "" for item in user_content: if isinstance(item, dict) and item.get("type") == "text": text_content += item.get("text", "") user_content = text_content if text_content else str(user_content) messages.append({'role': 'user', 'content': user_content}) messages.append({'role': 'assistant', 'content': h[1]}) return messages def messages_to_history(messages: Messages) -> Tuple[str, History]: assert messages[0]['role'] == 'system' history = [] for q, r in zip(messages[1::2], messages[2::2]): # Extract text content from multimodal messages for history user_content = q['content'] if isinstance(user_content, list): text_content = "" for item in user_content: if isinstance(item, dict) and item.get("type") == "text": text_content += item.get("text", "") user_content = text_content if text_content else str(user_content) history.append([user_content, r['content']]) return history def history_to_chatbot_messages(history: History) -> List[Dict[str, str]]: """Convert history tuples to chatbot message format""" messages = [] for user_msg, assistant_msg in history: # Handle multimodal content if isinstance(user_msg, list): text_content = "" for item in user_msg: if isinstance(item, dict) and item.get("type") == "text": text_content += item.get("text", "") user_msg = text_content if text_content else str(user_msg) messages.append({"role": "user", "content": user_msg}) messages.append({"role": "assistant", "content": assistant_msg}) return messages def remove_code_block(text): # Try to match code blocks with language markers patterns = [ r'```(?:html|HTML)\n([\s\S]+?)\n```', # Match ```html or ```HTML r'```\n([\s\S]+?)\n```', # Match code blocks without language markers r'```([\s\S]+?)```' # Match code blocks without line breaks ] for pattern in patterns: match = re.search(pattern, text, re.DOTALL) if match: extracted = match.group(1).strip() return extracted # If no code block is found, check if the entire text is HTML if text.strip().startswith('') or text.strip().startswith(' str: """Perform web search using Tavily with default parameters""" if not tavily_client: return "Web search is not available. Please set the TAVILY_API_KEY environment variable." try: # Use Tavily defaults with advanced search depth for better results search_params = { "search_depth": "advanced", "max_results": min(max(1, max_results), 20) } if include_domains is not None: search_params["include_domains"] = include_domains if exclude_domains is not None: search_params["exclude_domains"] = exclude_domains response = tavily_client.search(query, **search_params) search_results = [] for result in response.get('results', []): title = result.get('title', 'No title') url = result.get('url', 'No URL') content = result.get('content', 'No content') search_results.append(f"Title: {title}\nURL: {url}\nContent: {content}\n") if search_results: return "Web Search Results:\n\n" + "\n---\n".join(search_results) else: return "No search results found." except Exception as e: return f"Search error: {str(e)}" def enhance_query_with_search(query: str, enable_search: bool) -> str: """Enhance the query with web search results if search is enabled""" if not enable_search or not tavily_client: return query # Perform search to get relevant information search_results = perform_web_search(query) # Combine original query with search results enhanced_query = f"""Original Query: {query} {search_results} Please use the search results above to help create the requested application with the most up-to-date information and best practices.""" return enhanced_query def send_to_sandbox(code): # Add a wrapper to inject necessary permissions and ensure full HTML wrapped_code = f""" {code} """ encoded_html = base64.b64encode(wrapped_code.encode('utf-8')).decode('utf-8') data_uri = f"data:text/html;charset=utf-8;base64,{encoded_html}" iframe = f'' return iframe def demo_card_click(e: gr.EventData): try: # Get the index from the event data if hasattr(e, '_data') and e._data: # Try different ways to get the index if 'index' in e._data: index = e._data['index'] elif 'component' in e._data and 'index' in e._data['component']: index = e._data['component']['index'] elif 'target' in e._data and 'index' in e._data['target']: index = e._data['target']['index'] else: # If we can't get the index, try to extract it from the card data index = 0 else: index = 0 # Ensure index is within bounds if index >= len(DEMO_LIST): index = 0 return DEMO_LIST[index]['description'] except (KeyError, IndexError, AttributeError) as e: # Return the first demo description as fallback return DEMO_LIST[0]['description'] def extract_text_from_image(image_path): """Extract text from image using OCR""" try: # Check if tesseract is available try: pytesseract.get_tesseract_version() except Exception: return "Error: Tesseract OCR is not installed. Please install Tesseract to extract text from images. See install_tesseract.md for instructions." # Read image using OpenCV image = cv2.imread(image_path) if image is None: return "Error: Could not read image file" # Convert to RGB (OpenCV uses BGR) image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB) # Preprocess image for better OCR results # Convert to grayscale gray = cv2.cvtColor(image_rgb, cv2.COLOR_RGB2GRAY) # Apply thresholding to get binary image _, binary = cv2.threshold(gray, 0, 255, cv2.THRESH_BINARY + cv2.THRESH_OTSU) # Extract text using pytesseract text = pytesseract.image_to_string(binary, config='--psm 6') return text.strip() if text.strip() else "No text found in image" except Exception as e: return f"Error extracting text from image: {e}" def extract_text_from_file(file_path): if not file_path: return "" mime, _ = mimetypes.guess_type(file_path) ext = os.path.splitext(file_path)[1].lower() try: if ext == ".pdf": with open(file_path, "rb") as f: reader = PyPDF2.PdfReader(f) return "\n".join(page.extract_text() or "" for page in reader.pages) elif ext in [".txt", ".md"]: with open(file_path, "r", encoding="utf-8") as f: return f.read() elif ext == ".csv": with open(file_path, "r", encoding="utf-8") as f: return f.read() elif ext == ".docx": doc = docx.Document(file_path) return "\n".join([para.text for para in doc.paragraphs]) elif ext.lower() in [".jpg", ".jpeg", ".png", ".bmp", ".tiff", ".tif", ".gif", ".webp"]: return extract_text_from_image(file_path) else: return "" except Exception as e: return f"Error extracting text: {e}" def extract_website_content(url: str) -> str: """Extract HTML code and content from a website URL""" try: # Validate URL parsed_url = urlparse(url) if not parsed_url.scheme: url = "https://" + url parsed_url = urlparse(url) if not parsed_url.netloc: return "Error: Invalid URL provided" # Set comprehensive headers to mimic a real browser request headers = { 'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,image/apng,*/*;q=0.8', 'Accept-Language': 'en-US,en;q=0.9', 'Accept-Encoding': 'gzip, deflate, br', 'DNT': '1', 'Connection': 'keep-alive', 'Upgrade-Insecure-Requests': '1', 'Sec-Fetch-Dest': 'document', 'Sec-Fetch-Mode': 'navigate', 'Sec-Fetch-Site': 'none', 'Sec-Fetch-User': '?1', 'Cache-Control': 'max-age=0' } # Create a session to maintain cookies and handle redirects session = requests.Session() session.headers.update(headers) # Make the request with retry logic max_retries = 3 for attempt in range(max_retries): try: response = session.get(url, timeout=15, allow_redirects=True) response.raise_for_status() break except requests.exceptions.HTTPError as e: if e.response.status_code == 403 and attempt < max_retries - 1: # Try with different User-Agent on 403 session.headers['User-Agent'] = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36' continue else: raise # Get the raw HTML content with proper encoding try: # Try to get the content with automatic encoding detection response.encoding = response.apparent_encoding raw_html = response.text except: # Fallback to UTF-8 if encoding detection fails raw_html = response.content.decode('utf-8', errors='ignore') # Debug: Check if we got valid HTML if not raw_html.strip().startswith(' 10: print(f"Warning: This site has {len(script_tags)} script tags - it may be a JavaScript-heavy site") print("The content might be loaded dynamically and not available in the initial HTML") # Extract title title = soup.find('title') title_text = title.get_text().strip() if title else "No title found" # Extract meta description meta_desc = soup.find('meta', attrs={'name': 'description'}) description = meta_desc.get('content', '') if meta_desc else "" # Extract main content areas for analysis content_sections = [] main_selectors = [ 'main', 'article', '.content', '.main-content', '.post-content', '#content', '#main', '.entry-content', '.post-body' ] for selector in main_selectors: elements = soup.select(selector) for element in elements: text = element.get_text().strip() if len(text) > 100: # Only include substantial content content_sections.append(text) # Extract navigation links for analysis nav_links = [] nav_elements = soup.find_all(['nav', 'header']) for nav in nav_elements: links = nav.find_all('a') for link in links: link_text = link.get_text().strip() link_href = link.get('href', '') if link_text and link_href: nav_links.append(f"{link_text}: {link_href}") # Extract and fix image URLs in the HTML img_elements = soup.find_all('img') for img in img_elements: src = img.get('src', '') if src: # Handle different URL formats if src.startswith('//'): # Protocol-relative URL absolute_src = 'https:' + src img['src'] = absolute_src elif src.startswith('/'): # Root-relative URL absolute_src = urljoin(url, src) img['src'] = absolute_src elif not src.startswith(('http://', 'https://')): # Relative URL absolute_src = urljoin(url, src) img['src'] = absolute_src # If it's already absolute, keep it as is # Also check for data-src (lazy loading) and other common attributes data_src = img.get('data-src', '') if data_src and not src: # Use data-src if src is empty if data_src.startswith('//'): absolute_data_src = 'https:' + data_src img['src'] = absolute_data_src elif data_src.startswith('/'): absolute_data_src = urljoin(url, data_src) img['src'] = absolute_data_src elif not data_src.startswith(('http://', 'https://')): absolute_data_src = urljoin(url, data_src) img['src'] = absolute_data_src else: img['src'] = data_src # Also fix background image URLs in style attributes elements_with_style = soup.find_all(attrs={'style': True}) for element in elements_with_style: style_attr = element.get('style', '') # Find and replace relative URLs in background-image import re bg_pattern = r'background-image:\s*url\(["\']?([^"\']+)["\']?\)' matches = re.findall(bg_pattern, style_attr, re.IGNORECASE) for match in matches: if match: if match.startswith('//'): absolute_bg = 'https:' + match style_attr = style_attr.replace(match, absolute_bg) elif match.startswith('/'): absolute_bg = urljoin(url, match) style_attr = style_attr.replace(match, absolute_bg) elif not match.startswith(('http://', 'https://')): absolute_bg = urljoin(url, match) style_attr = style_attr.replace(match, absolute_bg) element['style'] = style_attr # Fix background images in