linkis-python-sdk 0.1.0

Python SDK for Apache Linkis

Linkis Python SDK

Python SDK for Apache Linkis, providing a simple interface to submit and manage jobs.

中文文档

Installation

pip install linkis-python-sdk

Features

• User authentication
• Job submission and execution
• Job status monitoring
• Results retrieval (with pandas DataFrame support)
• Job termination

Quick Start

Submitting a Job and Getting Results

from linkis_python_sdk import LinkisClient

# Create a client
client = LinkisClient(
    address="http://linkis-gateway:9001",
    username="your-username",
    password="your-password"
)

# Login
client.login()

# Submit a job and wait for results
result = client.execute(
    code="SELECT * FROM my_table LIMIT 10",
    run_type="sql",
    engine_type="spark-2.4.3"
)

# Convert results to pandas DataFrame
df = client.get_result_dataframe(result)
print(df)

Killing a Running Job

from linkis_python_sdk import LinkisClient

# Create a client
client = LinkisClient(
    address="http://linkis-gateway:9001",
    username="your-username",
    password="your-password"
)

# Login
client.login()

# Submit a job (don't wait)
result = client.execute(
    code="SELECT * FROM my_big_table",
    run_type="sql",
    engine_type="spark-2.4.3",
    wait=False
)

# Kill the job
client.kill_job(result['exec_id'])

API Documentation

LinkisClient

The main client class for interacting with Linkis.

Constructor

LinkisClient(
    address: str,
    token_key: str = None,
    token_value: str = None, 
    username: str = None,
    password: str = None,
    timeout: int = 60,
    api_version: str = "v1"
)

• address: Linkis gateway address (e.g., "http://127.0.0.1:9001")
• token_key: Authentication token key (optional)
• token_value: Authentication token value (optional)
• username: Username for login (required if token not provided)
• password: Password for login (required if token not provided)
• timeout: Request timeout in seconds
• api_version: API version to use

Methods

• login(): Authenticate with Linkis server
• submit_job(code, run_type, engine_type, source, params): Submit a job
• get_job_info(task_id): Get job status and information
• get_job_results(task_id): Get result file paths
• get_result_content(file_path): Get content of a result file
• kill_job(exec_id): Kill a running job
• execute(code, run_type, engine_type, source, params, wait, interval, timeout, callback): Submit and optionally wait for job completion
• get_result_dataframe(result): Convert execution result to pandas DataFrame

License

MIT