生产级代理AI系统（ Agentic AI System）设计与实现：Production-Grade Agentic AI System Design and Implementation

文章目录

• Production-Grade Agentic AI System Design and Implementation: Building Agentic AI Systems Using FastAPI and LangGraph
• Table of Contents
• Preface
• Part I: Foundations of the Modern AI Stack
• Chapter 1: The Shift to Agentic AI
• • 1.1 From Chains to Agents: Understanding the Paradigm Shift
  • 1.2 The Cognitive Architecture of Agents
  • 1.3 Industry Use Cases: Where Agents Succeed
  • 1.4 The Technology Stack: Why FastAPI and LangGraph
• Chapter 2: Modern FastAPI Fundamentals for AI
• • 2.1 Deep Dive into Asynchronous Python
  • 2.2 Advanced Pydantic Patterns for AI Applications
  • 2.3 Dependency Injection for AI Resources
  • 2.4 Background Tasks and Long-Running Operations
• Chapter 3: LangGraph Core Concepts
• • 3.1 The Evolution from LangChain to LangGraph
  • 3.2 The StateGraph: Defining Your Agent's Mind
  • 3.3 Nodes and Edges: Building the Workflow
  • 3.4 Checkpointers: Persistence and Time Travel
• Part II: Designing the Agent
• Chapter 4: State Management and Memory Design
• • 4.1 Short-Term vs. Long-Term Memory Architectures
  • 4.2 Designing the Agent State Schema
  • 4.3 Implementing Semantic Memory with Vector Stores
  • 4.4 Managing Context Windows: Strategies for Long Interactions
• Chapter 5: Tool Use and Environment Interaction
• • 5.1 Defining Tools: Schema-Based Tool Specifications
  • 5.2 The ToolNode Pattern: Safe Tool Execution
  • 5.3 Error Handling and Self-Correction Loops