Create a Gemini-Powered Chatbot with Function Calling

Build a chatbot using the Gemini API with function calling capabilities. The chatbot should:

**1. Core Configuration**:
- Model: gemini-2.0-flash or gemini-2.0-pro (specify trade-offs)
- System instruction defining the chatbot's persona and capabilities
- Conversation history management with token budgeting
- Temperature and safety settings appropriate for the use case

**2. Function Declarations** (implement these tools):
- search_knowledge_base(query: string, filters?: object) → searches internal docs
- get_user_info(user_id: string) → retrieves user account details
- create_ticket(title: string, description: string, priority: string) → creates support ticket
- check_order_status(order_id: string) → looks up order information
- schedule_callback(phone: string, preferred_time: string) → books a callback

**3. Function Calling Flow**:
- Parse Gemini's function call response
- Execute the function with provided arguments
- Return results to Gemini for natural language formatting
- Handle multi-turn function calls (tool calls that depend on previous results)
- Implement parallel function calling when queries need multiple data sources

**4. Error Handling**:
- Graceful fallback when functions fail
- User-friendly error messages
- Retry logic with exponential backoff for API calls

**5. Code**: Provide complete implementation in Python or TypeScript.