A comprehensive chatbot system built with Streamlit, LangChain, and Groq that can answer questions from uploaded documents and collect user information through conversational forms for appointment booking.

• Multi-format Support: Upload PDF, TXT, and DOCX files
• Intelligent Chunking: Documents are split into optimal chunks for better retrieval
• Vector Search: Uses Weviate for efficient similarity search
• Conversational Memory: Maintains context across conversations
• Source Citations: Shows relevant document sections for answers

• Natural Language Processing: Understands requests like "call me" or "book appointment"
• Conversational Forms: Step-by-step information collection
• Smart Date Extraction: Handles natural date inputs like "next Monday", "tomorrow"
• Input Validation: Real-time validation for emails, phone numbers, and names
• Confirmation Process: Users can review and confirm their information

• Agent-based Architecture: Uses LangChain agents for specialized tasks
• Multiple Tools: Date extraction, validation, form management, and search
• Context Awareness: Switches between document Q&A and form collection seamlessly

├── logs
├── uploaded_docs
├── vectors_store      
├── app.py                    # Main Streamlit application
├── config.py                # Configuration settings
├── document_processor.py     # Document loading and vector store management
├── chatbot.py               # Main chatbot logic
├── conversational_form.py   # Form collection and management
├── tool_agents.py           # Agent tools and integration
├── validators.py            # Input validation and date extraction
├── requirements.txt         # Python dependencies
└── README.md               # This file

1. Clone or download all the files to a directory

2. Install dependencies:

   pip install -r requirements.txt

3. Set up your Groq API key:

   • Get your API key from Groq Console
   • Either set it as an environment variable as creating .env and inside .env set this:

     GROQ_API_KEY=your-api-key-here

4. Use docker to run Weviate:

   • Start Weaviate locally:
   • Use docker for this

     docker run -p 8080:8080 -p 50051:50051 semitechnologies/weaviate:latest

5. Run the application:

   streamlit run app.py

Edit config.py to customize:

• MODEL_NAME: Groq model to use (default: "llama-3.3-70b-versatile2")
• CHUNK_SIZE: Document chunk size (default: 1000)
• CHUNK_OVERLAP: Overlap between chunks (default: 200)
• TEMPERATURE: Model creativity (default: 0.3)
• MAX_TOKENs: (default: 4000)
• EMBEDDING_MODEL: Sentence transformer model (default: "sentence-transformers/all-mpnet-base-v2")

1. Upload Documents:

   • Use the sidebar to upload PDF, TXT, or DOCX files
   • Click "Process Documents" to create the vector store
   • Documents are chunked and indexed automatically

2. Ask Questions:

   • Type questions about your documents in the chat
   • Get answers with relevant source excerpts
   • Maintain conversation context

3. Switch back to appointment booking to document QA

   • Clear chat as it has history saving it can answer based on previous chat
   • for better performance

1. Trigger Form Collection:

   • Say things like "call me", "book appointment", "schedule a meeting"
   • The system will detect the intent and start form collection

2. Provide Information:

   • Name: Full name validation
   • Email: Email format validation
   • Phone: phone number validation
   • Date: Natural language date parsing (e.g., "next Monday", "tomorrow")

3. Confirmation:

   • Review your information
   • Confirm or request changes
   • Receive confirmation message

The system understands various date formats:

• "tomorrow", "today", "yesterday"
• "next Monday", "next Friday"
• "next week", "next month"
• "2024-01-15", "01/15/2024"
• "January 15th", "Jan 15"

The system includes several specialized tools:

1. schedule_appointment: Handles appointment booking workflow
2. extract_date: Converts natural language to YYYY-MM-DD format
3. get_form_status: Shows current form collection progress
4. validate_contact_info: Validates emails, phones, and names
5. search_appointments: Searches through completed appointments

• Email: Verifying Email with regex
• Phone: Character validation and length requirements of Phone numbers
• Name: Character validation and length requirements
• Date: Comprehensive natural language date parsing

• Context Switching: Seamlessly switches between document Q&A and form collection
• Memory Management: Maintains conversation history and form state
• Error Handling: Graceful error recovery with helpful messages
• Progress Tracking: Shows form completion progress

• Clean Design: Intuitive Streamlit interface
• Sidebar Organization: Document upload, form status, and help sections
• Chat History: Persistent conversation display
• Real-time Updates: Dynamic form progress and status updates

• No Persistent Storage: Forms are only stored in session state
• API Key Security: Keys are handled securely through environment variables
• Input Sanitization: All user inputs are validated and sanitized
• Error Boundaries: Comprehensive error handling prevents crashes

1. "Agent is not available"

   • Check your Groq API key
   • Ensure internet connectivity
   • Verify model availability

2. Document processing fails

   • Check file format (PDF, TXT, DOCX only)
   • Ensure files are not corrupted
   • Check file size limitations

3. Weviate Issue

   • Check you weviate docker conatiner is running or not

• Document Size: Keep documents under 10MB for best performance
• Chunk Size: Adjust chunk size based on document complexity
• Memory Usage: Clear chat history periodically for long sessions
• API Limits: Monitor Groq API usage to avoid rate limits