linkis-python-sdk 0.2.0

Python SDK for Linkis Service

Linkis Python SDK

Linkis Python SDK是与Apache Linkis服务交互的Python客户端库。该SDK提供了简便的接口，用于执行代码并获取结果，支持同步和异步操作。

Linkis Python SDK is the Python client library for interacting with Apache Linkis services. This SDK provides convenient interfaces for executing code and retrieving results, supporting both synchronous and asynchronous operations.

安装

Installation

pip install linkis-python-sdk

功能特点

Features

• 支持同步和异步客户端

• 支持结果缓存

• 提供数据过滤、排序和分页功能

• 轻松将结果转换为Pandas DataFrame

• Support for synchronous and asynchronous clients

• Result caching capabilities

• Data filtering, sorting, and pagination

• Easy conversion of results to Pandas DataFrame

使用示例

Usage Examples

同步客户端

Synchronous Client

from linkis_python_sdk.client import LinkisClient
from linkis_python_sdk.config.config import LinkisConfig

# 创建配置
# Create configuration
config = LinkisConfig(
    base_url="http://your-linkis-server/api/rest_j/v1",
    username="your_username",
    password="your_password"
)

# 创建客户端
# Create client
client = LinkisClient(config)

# 执行代码
# Execute code
code = "SELECT * FROM my_table LIMIT 10"
query_req = {
    "columns": [{"column": "id"}, {"column": "name"}],
    "filters": [{"column": "age", "operator": "gt", "value": 18}],
    "orderbys": [{"column": "id", "ascending": True}],
    "page": 1,
    "page_size": 10
}
data, total, error = client.execute_code(
    code=code,
    query_req=query_req,
    engine_type="spark",
    run_type="sql"
)

print(f"Total records: {total}")
print(f"Data: {data}")

异步客户端

Asynchronous Client

import asyncio
from linkis_python_sdk.asyncio.client import LinkisClient
from linkis_python_sdk.config.config import LinkisConfig

async def main():
    # 创建配置
    # Create configuration
    config = LinkisConfig(
        base_url="http://your-linkis-server/api/rest_j/v1",
        username="your_username",
        password="your_password"
    )
    
    # 创建异步客户端
    # Create asynchronous client
    client = LinkisClient(config)
    
    # 执行代码
    # Execute code
    code = "SELECT * FROM my_table LIMIT 10"
    query_req = {
        "columns": [{"column": "id"}, {"column": "name"}],
        "filters": [{"column": "age", "operator": "gt", "value": 18}],
        "orderbys": [{"column": "id", "ascending": True}],
        "page": 1,
        "page_size": 10
    }
    data, total, error = await client.execute_code(
        code=code,
        query_req=query_req,
        engine_type="spark",
        run_type="sql"
    )
    
    print(f"Total records: {total}")
    print(f"Data: {data}")

if __name__ == "__main__":
    asyncio.run(main())

带缓存的执行

Execution with Caching

SDK还支持使用Redis缓存执行结果，以提高性能：

The SDK also supports caching execution results with Redis to improve performance:

from redis import ConnectionPool
from linkis_python_sdk.client import LinkisClient
from linkis_python_sdk.config.config import LinkisConfig

# 创建Redis连接池
# Create Redis connection pool
redis_pool = ConnectionPool(host="localhost", port=6379, db=0)

# 创建配置
# Create configuration
config = LinkisConfig(
    base_url="http://your-linkis-server/api/rest_j/v1",
    username="your_username",
    password="your_password"
)

# 创建带缓存的客户端
# Create client with caching
client = LinkisClient(
    linkis_conf=config,
    r_pool=redis_pool,
    cache_expire=1800,  # 缓存过期时间（秒）/ Cache expiration time (seconds)
    cache_code_prefix="linkis:codecache"  # 缓存键前缀 / Cache key prefix
)

# 执行带缓存的代码
# Execute code with caching
data, total, error = client.execute_code_with_cache(
    code="SELECT * FROM my_table LIMIT 10",
    query_req={"columns": [{"column": "id"}, {"column": "name"}]},
    engine_type="spark",
    run_type="sql"
)

依赖项

Dependencies

• Python >= 3.9
• requests
• pandas
• redis
• aiohttp

开发贡献

Development and Contribution

欢迎为Linkis Python SDK做出贡献！请参阅PUBLISH.md了解如何构建和发布该包。 Contributions to the Linkis Python SDK are welcome! Please refer to PUBLISH.md for information on how to build and publish the package.

许可证

License

该项目采用Apache License 2.0许可证。

This project is licensed under the Apache License 2.0.