---

name: llm-app-patterns description: "构建 LLM 应用的生产就绪模式，灵感来自 Dify 和行业最佳实践。" risk: unknown source: community date_added: "2026-02-27"

LLM 应用模式

构建 LLM 应用的生产就绪模式，灵感来自 Dify 和行业最佳实践。

适用场景

在以下情况使用本技能：

• 设计 LLM 驱动的应用
• 实现 RAG（检索增强生成）
• 构建带工具的 AI 代理
• 设置 LLMOps 监控
• 在代理架构之间做选择

---

1. RAG 管道架构

概述

RAG（检索增强生成）将 LLM 响应建立在你的数据基础上。

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│   Ingest    │────▶│   Retrieve  │────▶│   Generate  │
│  Documents  │     │   Context   │     │   Response  │
└─────────────┘     └─────────────┘     └─────────────┘
      │                   │                   │
      ▼                   ▼                   ▼
 ┌─────────┐       ┌───────────┐       ┌───────────┐
 │ Chunking│       │  Vector   │       │    LLM    │
 │Embedding│       │  Search   │       │  + Context│
 └─────────┘       └───────────┘       └───────────┘

1.1 文档摄取

# Chunking strategies
class ChunkingStrategy:
    # Fixed-size chunks (simple but may break context)
    FIXED_SIZE = "fixed_size"  # e.g., 512 tokens

    # Semantic chunking (preserves meaning)
    SEMANTIC = "semantic"      # Split on paragraphs/sections

    # Recursive splitting (tries multiple separators)
    RECURSIVE = "recursive"    # ["\n\n", "\n", " ", ""]

    # Document-aware (respects structure)
    DOCUMENT_AWARE = "document_aware"  # Headers, lists, etc.

# Recommended settings
CHUNK_CONFIG = {
    "chunk_size": 512,       # tokens
    "chunk_overlap": 50,     # token overlap between chunks
    "separators": ["\n\n", "\n", ". ", " "],
}

1.2 嵌入与存储

# Vector database selection
VECTOR_DB_OPTIONS = {
    "pinecone": {
        "use_case": "Production, managed service",
        "scale": "Billions of vectors",
        "features": ["Hybrid search", "Metadata filtering"]
    },
    "weaviate": {
        "use_case": "Self-hosted, multi-modal",
        "scale": "Millions of vectors",
        "features": ["GraphQL API", "Modules"]
    },
    "chromadb": {
        "use_case": "Development, prototyping",
        "scale": "Thousands of vectors",
        "features": ["Simple API", "In-memory option"]
    },
    "pgvector": {
        "use_case": "Existing Postgres infrastructure",
        "scale": "Millions of vectors",
        "features": ["SQL integration", "ACID compliance"]
    }
}

# Embedding model selection
EMBEDDING_MODELS = {
    "openai/text-embedding-3-small": {
        "dimensions": 1536,
        "cost": "$0.02/1M tokens",
        "quality": "Good for most use cases"
    },
    "openai/text-embedding-3-large": {
        "dimensions": 3072,
        "cost": "$0.13/1M tokens",
        "quality": "Best for complex queries"
    },
    "local/bge-large": {
        "dimensions": 1024,
        "cost": "Free (compute only)",
        "quality": "Comparable to OpenAI small"
    }
}

1.3 检索策略

# Basic semantic search
def semantic_search(query: str, top_k: int = 5):
    query_embedding = embed(query)
    results = vector_db.similarity_search(
        query_embedding,
        top_k=top_k
    )
    return results

# Hybrid search (semantic + keyword)
def hybrid_search(query: str, top_k: int = 5, alpha: float = 0.5):
    """
    alpha=1.0: Pure semantic
    alpha=0.0: Pure keyword (BM25)
    alpha=0.5: Balanced
    """
    semantic_results = vector_db.similarity_search(query)
    keyword_results = bm25_search(query)

    # Reciprocal Rank Fusion
    return rrf_merge(semantic_results, keyword_results, alpha)

# Multi-query retrieval
def multi_query_retrieval(query: str):
    """Generate multiple query variations for better recall"""
    queries = llm.generate_query_variations(query, n=3)
    all_results = []
    for q in queries:
        all_results.extend(semantic_search(q))
    return deduplicate(all_results)

# Contextual compression
def compressed_retrieval(query: str):
    """Retrieve then compress to relevant parts only"""
    docs = semantic_search(query, top_k=10)
    compressed = llm.extract_relevant_parts(docs, query)
    return compressed

1.4 带上下文的生成

RAG_PROMPT_TEMPLATE = """
Answer the user's question based ONLY on the following context.
If the context doesn't contain enough information, say "I don't have enough information to answer that."

Context:
{context}

Question: {question}

Answer:"""

def generate_with_rag(question: str):
    # Retrieve
    context_docs = hybrid_search(question, top_k=5)
    context = "\n\n".join([doc.content for doc in context_docs])