Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

title: “Guidance” sidebar_label: “Guidance” description: “Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidanc…”

{/* 本页面由 website/scripts/generate-skill-docs.py 根据该技能的 SKILL.md 文件自动生成。请直接编辑源文件 SKILL.md，而非此页面。 */}

Guidance

借助 Microsoft Research 开发的受限生成框架 Guidance，您可以通过正则表达式和语法来控制大语言模型的输出，确保生成有效的 JSON/XML/代码，强制采用结构化格式，并构建多步骤工作流。

技能元数据
来源可选 — 通过 hermes skills install official/mlops/guidance 安装
路径optional-skills/mlops/guidance
版本1.0.0
开发者Orchestra Research
许可证MIT
依赖项guidance, transformers
支持平台linux、macos、windows
标签提示工程、Guidance、受限生成、结构化输出、JSON 验证、语法、Microsoft Research、格式强制、多步骤工作流

参考：完整的 SKILL.md 文件

:::info 以下是当触发该技能时 Hermes 会加载的完整技能定义。当该技能处于激活状态时，智能体看到的指令即为此内容。
::

Guidance：受限大语言模型生成

何时使用此技能

在以下情况下可使用 Guidance：

  • 利用正则表达式或语法控制大语言模型的输出语法
  • 确保生成有效的 JSON/XML/代码
  • 相比传统提示方法降低延迟
  • 强制要求采用结构化格式（日期、电子邮件、编号等）
  • 借助 Python 风格的控制流构建多步骤工作流
  • 通过语法约束防止无效输出

GitHub 星标数：18,000+ | 来源：Microsoft Research

安装方式

# Base installation
pip install guidance

# With specific backends
pip install guidance[transformers]  # Hugging Face models
pip install guidance[llama_cpp]     # llama.cpp models

快速入门

基本示例：结构化生成

from guidance import models, gen

# Load model (supports OpenAI, Transformers, llama.cpp)
lm = models.OpenAI("gpt-4")

# Generate with constraints
result = lm + "The capital of France is " + gen("capital", max_tokens=5)

print(result["capital"])  # "Paris"

集成 Anthropic Claude

from guidance import models, gen, system, user, assistant

# Configure Claude
lm = models.Anthropic("claude-sonnet-4-5-20250929")

# Use context managers for chat format
with system():
    lm += "You are a helpful assistant."

with user():
    lm += "What is the capital of France?"

with assistant():
    lm += gen(max_tokens=20)

核心概念

1. 上下文管理器

该指南采用符合 Python 风格的上下文管理器来实现类似对话式的交互。

from guidance import system, user, assistant, gen

lm = models.Anthropic("claude-sonnet-4-5-20250929")

# System message
with system():
    lm += "You are a JSON generation expert."

# User message
with user():
    lm += "Generate a person object with name and age."

# Assistant response
with assistant():
    lm += gen("response", max_tokens=100)

print(lm["response"])

优势：

  • 自然的对话流程
  • 清晰的角色分工
  • 易于阅读与维护

2. 约束生成

通过规则引导，利用正则表达式或语法确保输出符合指定模式。

正则表达式约束

from guidance import models, gen

lm = models.Anthropic("claude-sonnet-4-5-20250929")

# Constrain to valid email format
lm += "Email: " + gen("email", regex=r"[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}")

# Constrain to date format (YYYY-MM-DD)
lm += "Date: " + gen("date", regex=r"\d{4}-\d{2}-\d{2}")

# Constrain to phone number
lm += "Phone: " + gen("phone", regex=r"\d{3}-\d{3}-\d{4}")

print(lm["email"])  # Guaranteed valid email
print(lm["date"])   # Guaranteed YYYY-MM-DD format

工作原理：

  • 正则表达式在分词层面被转换为语法结构
  • 在生成过程中会过滤掉无效的分词
  • 模型仅能输出符合规则的响应内容

选择约束条件

from guidance import models, gen, select

lm = models.Anthropic("claude-sonnet-4-5-20250929")

# Constrain to specific choices
lm += "Sentiment: " + select(["positive", "negative", "neutral"], name="sentiment")

# Multiple-choice selection
lm += "Best answer: " + select(
    ["A) Paris", "B) London", "C) Berlin", "D) Madrid"],
    name="answer"
)

print(lm["sentiment"])  # One of: positive, negative, neutral
print(lm["answer"])     # One of: A, B, C, or D

3. 令牌修复功能

该功能可自动“修复”提示词与生成内容之间的令牌边界问题。

问题： 分词处理会形成不自然的边界。

# Without token healing
prompt = "The capital of France is "
# Last token: " is "
# First generated token might be " Par" (with leading space)
# Result: "The capital of France is  Paris" (double space!)

解决方案： Guidance功能会备份一个令牌并重新生成。

from guidance import models, gen

lm = models.Anthropic("claude-sonnet-4-5-20250929")

# Token healing enabled by default
lm += "The capital of France is " + gen("capital", max_tokens=5)
# Result: "The capital of France is Paris" (correct spacing)

优势：

  • 自然的文本分界
  • 无尴尬的间距问题
  • 更出色的模型性能（能够识别自然的标记序列）

4. 基于语法的生成

利用上下文无关语法来定义复杂结构。

from guidance import models, gen

lm = models.Anthropic("claude-sonnet-4-5-20250929")

# JSON grammar (simplified)
json_grammar = """
{
    "name": <gen name regex="[A-Za-z ]+" max_tokens=20>,
    "age": <gen age regex="[0-9]+" max_tokens=3>,
    "email": <gen email regex="[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}" max_tokens=50>
}
"""

# Generate valid JSON
lm += gen("person", grammar=json_grammar)

print(lm["person"])  # Guaranteed valid JSON structure

应用场景：

  • 复杂的结构化输出
  • 嵌套的数据结构
  • 编程语言语法
  • 领域专用语言

5. 指导函数

通过 @guidance 装饰器创建可重复使用的生成模式。

from guidance import guidance, gen, models

@guidance
def generate_person(lm):
    """Generate a person with name and age."""
    lm += "Name: " + gen("name", max_tokens=20, stop="\n")
    lm += "\nAge: " + gen("age", regex=r"[0-9]+", max_tokens=3)
    return lm

# Use the function
lm = models.Anthropic("claude-sonnet-4-5-20250929")
lm = generate_person(lm)

print(lm["name"])
print(lm["age"])

有状态函数：

@guidance(stateless=False)
def react_agent(lm, question, tools, max_rounds=5):
    """ReAct agent with tool use."""
    lm += f"Question: {question}\n\n"

    for i in range(max_rounds):
        # Thought
        lm += f"Thought {i+1}: " + gen("thought", stop="\n")

        # Action
        lm += "\nAction: " + select(list(tools.keys()), name="action")

        # Execute tool
        tool_result = tools[lm["action"]]()
        lm += f"\nObservation: {tool_result}\n\n"

        # Check if done
        lm += "Done? " + select(["Yes", "No"], name="done")
        if lm["done"] == "Yes":
            break

    # Final answer
    lm += "\nFinal Answer: " + gen("answer", max_tokens=100)
    return lm

后端配置

Anthropic Claude

from guidance import models

lm = models.Anthropic(
    model="claude-sonnet-4-5-20250929",
    api_key="your-api-key"  # Or set ANTHROPIC_API_KEY env var
)

OpenAI

lm = models.OpenAI(
    model="gpt-4o-mini",
    api_key="your-api-key"  # Or set OPENAI_API_KEY env var
)

本地模型（Transformer架构）

from guidance.models import Transformers

lm = Transformers(
    "microsoft/Phi-4-mini-instruct",
    device="cuda"  # Or "cpu"
)

本地模型（llama.cpp）

from guidance.models import LlamaCpp

lm = LlamaCpp(
    model_path="/path/to/model.gguf",
    n_ctx=4096,
    n_gpu_layers=35
)

常见模式

模式 1：JSON 生成

from guidance import models, gen, system, user, assistant

lm = models.Anthropic("claude-sonnet-4-5-20250929")

with system():
    lm += "You generate valid JSON."

with user():
    lm += "Generate a user profile with name, age, and email."

with assistant():
    lm += """{
    "name": """ + gen("name", regex=r'"[A-Za-z ]+"', max_tokens=30) + """,
    "age": """ + gen("age", regex=r"[0-9]+", max_tokens=3) + """,
    "email": """ + gen("email", regex=r'"[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}"', max_tokens=50) + """
}"""

print(lm)  # Valid JSON guaranteed

模式 2：分类任务

from guidance import models, gen, select

lm = models.Anthropic("claude-sonnet-4-5-20250929")

text = "This product is amazing! I love it."

lm += f"Text: {text}\n"
lm += "Sentiment: " + select(["positive", "negative", "neutral"], name="sentiment")
lm += "\nConfidence: " + gen("confidence", regex=r"[0-9]+", max_tokens=3) + "%"

print(f"Sentiment: {lm['sentiment']}")
print(f"Confidence: {lm['confidence']}%")

模式 3：多步骤推理

from guidance import models, gen, guidance

@guidance
def chain_of_thought(lm, question):
    """Generate answer with step-by-step reasoning."""
    lm += f"Question: {question}\n\n"

    # Generate multiple reasoning steps
    for i in range(3):
        lm += f"Step {i+1}: " + gen(f"step_{i+1}", stop="\n", max_tokens=100) + "\n"

    # Final answer
    lm += "\nTherefore, the answer is: " + gen("answer", max_tokens=50)

    return lm

lm = models.Anthropic("claude-sonnet-4-5-20250929")
lm = chain_of_thought(lm, "What is 15% of 200?")

print(lm["answer"])

模式 4：ReAct 智能体

from guidance import models, gen, select, guidance

@guidance(stateless=False)
def react_agent(lm, question):
    """ReAct agent with tool use."""
    tools = {
        "calculator": lambda expr: eval(expr),
        "search": lambda query: f"Search results for: {query}",
    }

    lm += f"Question: {question}\n\n"

    for round in range(5):
        # Thought
        lm += f"Thought: " + gen("thought", stop="\n") + "\n"

        # Action selection
        lm += "Action: " + select(["calculator", "search", "answer"], name="action")

        if lm["action"] == "answer":
            lm += "\nFinal Answer: " + gen("answer", max_tokens=100)
            break

        # Action input
        lm += "\nAction Input: " + gen("action_input", stop="\n") + "\n"

        # Execute tool
        if lm["action"] in tools:
            result = tools[lm["action"]](lm["action_input"])
            lm += f"Observation: {result}\n\n"

    return lm

lm = models.Anthropic("claude-sonnet-4-5-20250929")
lm = react_agent(lm, "What is 25 * 4 + 10?")
print(lm["answer"])

模式 5：数据提取

from guidance import models, gen, guidance

@guidance
def extract_entities(lm, text):
    """Extract structured entities from text."""
    lm += f"Text: {text}\n\n"

    # Extract person
    lm += "Person: " + gen("person", stop="\n", max_tokens=30) + "\n"

    # Extract organization
    lm += "Organization: " + gen("organization", stop="\n", max_tokens=30) + "\n"

    # Extract date
    lm += "Date: " + gen("date", regex=r"\d{4}-\d{2}-\d{2}", max_tokens=10) + "\n"

    # Extract location
    lm += "Location: " + gen("location", stop="\n", max_tokens=30) + "\n"

    return lm

text = "Tim Cook announced at Apple Park on 2024-09-15 in Cupertino."

lm = models.Anthropic("claude-sonnet-4-5-20250929")
lm = extract_entities(lm, text)

print(f"Person: {lm['person']}")
print(f"Organization: {lm['organization']}")
print(f"Date: {lm['date']}")
print(f"Location: {lm['location']}")

最佳实践

1. 使用正则表达式进行格式验证

# ✅ Good: Regex ensures valid format
lm += "Email: " + gen("email", regex=r"[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}")

# ❌ Bad: Free generation may produce invalid emails
lm += "Email: " + gen("email", max_tokens=50)

2. 使用 select() 函数处理固定类别

# ✅ Good: Guaranteed valid category
lm += "Status: " + select(["pending", "approved", "rejected"], name="status")

# ❌ Bad: May generate typos or invalid values
lm += "Status: " + gen("status", max_tokens=20)

3. 充分利用令牌修复功能

# Token healing is enabled by default
# No special action needed - just concatenate naturally
lm += "The capital is " + gen("capital")  # Automatic healing

4. 使用停止序列

# ✅ Good: Stop at newline for single-line outputs
lm += "Name: " + gen("name", stop="\n")

# ❌ Bad: May generate multiple lines
lm += "Name: " + gen("name", max_tokens=50)

5. 创建可复用函数

# ✅ Good: Reusable pattern
@guidance
def generate_person(lm):
    lm += "Name: " + gen("name", stop="\n")
    lm += "\nAge: " + gen("age", regex=r"[0-9]+")
    return lm

# Use multiple times
lm = generate_person(lm)
lm += "\n\n"
lm = generate_person(lm)

6. 平衡约束条件

# ✅ Good: Reasonable constraints
lm += gen("name", regex=r"[A-Za-z ]+", max_tokens=30)

# ❌ Too strict: May fail or be very slow
lm += gen("name", regex=r"^(John|Jane)$", max_tokens=10)

与其他方案的对比

功能特性GuidanceInstructorOutlinesLMQL
正则表达式限制✅ 支持❌ 不支持✅ 支持✅ 支持
语法支持✅ 基于CFG❌ 不支持✅ 基于CFG✅ 基于CFG
Pydantic验证❌ 不支持✅ 支持✅ 支持❌ 不支持
令牌修复功能✅ 支持❌ 不支持✅ 支持❌ 不支持
本地模型支持✅ 支持⚠️ 功能有限✅ 支持✅ 支持
API模型支持✅ 支持✅ 支持⚠️ 功能有限✅ 支持
Python风格语法✅ 支持✅ 支持✅ 支持❌ 类SQL语法
学习曲线较低较低中等较高

何时选择 Guidance：

  • 需要正则表达式或语法限制
  • 需要令牌修复功能
  • 需要构建包含控制流的复杂工作流
  • 使用本地模型（如Transformers、llama.cpp）
  • 偏好Python风格的语法

何时选择其他方案：

  • Instructor：需要支持Pydantic验证及自动重试功能
  • Outlines：需要JSON模式验证
  • LMQL：偏好声明式查询语法

性能特点

延迟降低：

  • 对于有约束的输出，其速度比传统提示方式快30-50%
  • 令牌修复功能可减少不必要的重新生成
  • 语法限制能有效避免无效令牌的产生

内存占用：

  • 相较于无约束生成模式，内存开销极低
  • 语法规则会在首次使用后进行缓存
  • 推理时能高效过滤令牌

令牌效率：

  • 避免因无效输出而浪费令牌
  • 无需重复尝试循环 | 能直接生成有效结果 |

相关资源

  • 文档：https://guidance.readthedocs.io
  • GitHub仓库：https://github.com/guidance-ai/guidance（星标数超1.8万）
  • 示例笔记本：https://github.com/guidance-ai/guidance/tree/main/notebooks
  • Discord社区：提供用户支持

相关参考

  • references/constraints.md - 详尽的正则表达式及语法模式说明
  • references/backends.md - 各后端特定的配置选项
  • references/examples.md - 可直接用于生产环境的示例代码