# 以管理员身份运行PowerShell，然后执行：
Start-Process -FilePath "mysql-installer-web-community-8.0.*.msi" -Verb RunAs

# 打开新的PowerShell窗口测试连接
mysql -u root -p

mysql>

SHOW DATABASES;
exit;

# Python虚拟环境可以隔离不同项目的依赖
python -m venv langchain_env
.\langchain_env\Scripts\activate

pip install langchain openai mysql-connector-python python-dotenv pymysql tiktoken

pip install langchain==0.0.* openai==0.*.* 

CREATE DATABASE langchain_demo;
USE langchain_demo;

CREATE TABLE documents (
    id INT AUTO_INCREMENT PRIMARY KEY,
    content TEXT NOT NULL,
    metadata JSON,
    created_at TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);

INSERT INTO documents (content, metadata) VALUES 
('LangChain是一个用于开发语言模型应用的框架', '{"source": "官网", "type": "介绍"}'),
('MySQL是最流行的开源关系型数据库之一', '{"source": "维基百科", "type": "定义"}');

import os
from dotenv import load_dotenv
from langchain.document_loaders import SQLDatabaseLoader
from langchain.indexes import VectorstoreIndexCreator

# 加载环境变量（推荐将敏感信息放在.env文件中）
load_dotenv()

# MySQL连接配置（实际项目中应使用环境变量） 
db_config = {
    "host": "localhost",
    "user": "root",
    "password": os.getenv("MYSQL_PASSWORD"), # .env文件中设置MYSQL_PASSWORD=你的密码 
    "database": "langchain_demo",
    "port": 3306,
}

def basic_query_example():
    """基础SQL查询示例"""
    from mysql.connector import connect

    conn = connect(**db_config)
    cursor = conn.cursor(dictionary=True)

    cursor.execute("SELECT * FROM documents LIMIT 2")
    results = cursor.fetchall()

    print("=== SQL查询结果 ===")
    for row in results:
        print(f"ID: {row['id']}")
        print(f"内容: {row['content'][:50]}...")
        print("-" * 30)

    cursor.close()
    conn.close()

def langchain_integration():
    """LangChain集成示例"""
    from langchain.sql_database import SQLDatabase

    # SQLDatabase是LangChain的数据库抽象层 
    db = SQLDatabase.from_uri(
        f"mysql+mysqlconnector://{db_config['user']}:{db_config['password']}@{db_config['host']}:{db_config['port']}/{db_config['database']}"
    )

    # SQLDatabaseLoader可以加载数据库内容作为文档源 
    loader = SQLDatabaseLoader(db, table_name="documents")

    # VectorstoreIndexCreator会自动处理文档索引创建 
    index = VectorstoreIndexCreator().from_loaders([loader])

    # Now you can query the index like a vector database 
    query = "什么是LangChain?"
    print(f"\n=== LangChain查询: '{query}' ===")

    result = index.query(query)
    print("回答:", result)

if __name__ == "__main__":
    basic_query_example()

    if input("\n运行LangChain集成示例? (y/n): ").lower() == 'y':
        langchain_integration()

=== SQL查询结果 ===  
ID: 1  
内容: LangChain是一个用于开发语言模型应用的框架...  
------------------------------  
ID: 2  
内容: MySQL是最流行的开源关系型数据库之一...  

运行LangChain集成示例? (y/n): y  

=== LangChain查询: '什么是LangChain?' ===  
回答: LangChain是一个用于开发语言模型应用的框架。

from langchain.chains import LLMChain, SimpleSequentialChain   
from langchain.prompts import PromptTemplate   

def advanced_langchain_example():     
"""结合LLM进行数据处理的完整示例"""     
from langchain.chat_models import ChatOpenAI      
from sqlalchemy import create_engine      

# .env中设置OPENAI_API_KEY=sk-your-key      
engine = create_engine(           
f'mysql+mysqlconnector://{db_config["user"]}:{db_config["password"]}@{db_config["host"]}:{db_config["port"]}/{db_config["database"]}'      
)      

llm = ChatOpenAI(temperature=0, model_name="gpt-3")      

# Chain :分析文档内容      
analyze_template = """基于以下文档内容进行分析总结:
{documents}  

请用中文回答并突出关键信息"""      
analyze_prompt = PromptTemplate(input_variables=["documents"], template=analyze_template)      
analyze_chain = LLMChain(llm=llm, prompt=analyze_prompt)      

# Chain :生成SQL查询建议      
sql_template = """根据用户问题生成SQL查询:
问题:{question}  

已知表结构:
CREATE TABLE documents (
id INT PRIMARY KEY,
content TEXT,
metadata JSON);  

请返回可直接执行的SQL语句"""      
sql_prompt = PromptTemplate(input_variables=["question"], template=sql_template)      
sql_chain = LLMChain(llm=llm, prompt=sql_prompt)      

#组合链执行      
overall_chain = SimpleSequentialChain(chains=[analyze_chain, sql_chain], verbose=True)     

result = overall_chain.run("查找包含'框架'关键字的文档")     
print("\n最终结果:", result)   

if __name__ == "__main__":     
advanced_langchain_example() 

Get-Counter '\Process(*)\% Processor Time' | Select-Object -ExpandProperty CounterSamples | Where-Object {$_.InstanceName -match "(mysqld|python)"} | Sort-Object CookedValue -Descending | Select-Object InstanceName,CookedValue | Format-Table 

Get-Counter '\Memory\Available MBytes'

$action = New-ScheduledTaskAction -Execute 'cmd.exe' -Argument '/c mysqldump --user=root --password=[YOUR_PASSWORD] --databases langchain_demo > C:\backups\langchain_%date:~10,4%%date:~4,2%%date:~7,2%.sql && exit /b %errorlevel%' 

$trigger = New-ScheduledTaskTrigger -Daily -At '02:00' 

Register-ScheduledTask `
-TaskName 'MySQL Daily Backup' `
-Action $action `
-Trigger $trigger `
-RunLevel Highest `
-Password [YOUR_WINDOWS_PASSWORD]

# WSL中允许远程连接(不安全!仅开发用) 
sudo sed -i 's/bind\-address.*/bind\-address\t\t= /etc/mysql/mysql.conf.d/mysqld.cnf 

sudo service mysql restart 

# Windows防火墙放行WSL IP(每次重启WSL IP会变!) 
Get-NetIPAddress | Where {$_.InterfaceAlias -like "*WSL*"} | ForEach { New-NetFirewallRule `
-DisplayName WSL-MySQL-Inbound `
-Direction Inbound `
-LocalPort @(3306) `
-Action Allow `
-RemoteAddress $_.IPAddress }