deltastream-connector

A Python client library for DeltaStream - a SQL streaming processing engine based on Apache Flink.

Features

• Asynchronous API client for DeltaStream
• Support for SQL statements execution
• Streaming result sets
• File attachments for SQL queries (e.g., JAR files for UDFs)
• API Token authentication
• Python 3.11+ support

Installation

pip install deltastream-connector

Quick Start

import asyncio
from deltastream.api.conn import APIConnection

# Initialize connection with API token
auth_token = os.getenv("DELTASTREAM_AUTH_TOKEN")

if not auth_token:
    raise ValueError("Environment variable 'DELTASTREAM_AUTH_TOKEN' is not set")

# Use the token to construct the DSN and create the connection
dsn = f"https://:{auth_token}@api.deltastream.io/v2"
conn = APIConnection.from_dsn(dsn)

async def main():
    # Execute SQL queries
    rows = await conn.query("SELECT 1;")
    
    # Process results asynchronously
    async for row in rows:
        print(row)

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

File Attachments

The connector provides convenient helper functions to attach files (like JAR files for UDFs) to SQL queries without manually creating blob objects.

Simple File Attachment

import asyncio
from deltastream.api.conn import APIConnection

async def main():
    conn = APIConnection.from_dsn("https://:{token}@api.deltastream.io/v2")
    
    # Create a function source with a JAR file attachment
    await conn.exec_with_files(
        "CREATE FUNCTION_SOURCE \"my_udf\" WITH ('file' = 'my_function.jar', 'description' = 'My custom UDF');",
        ["/path/to/my_function.jar"]
    )
    
    # Query function sources and get results
    rows = await conn.query_with_files("SHOW FUNCTION_SOURCES;")
    async for row in rows:
        print(f"Function: {row[0]}, Status: {row[1]}")

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

Advanced File Attachment Configuration

# Custom file names and content types
await conn.exec_with_files(
    "CREATE FUNCTION_SOURCE \"advanced_udf\" WITH ('file' = 'custom_name.jar');",
    [{
        "path": "/path/to/actual_file.jar",
        "name": "custom_name.jar",
        "content_type": "application/java-archive"
    }]
)

# Multiple files
await conn.exec_with_files(
    "CREATE FUNCTION_SOURCE \"multi_file_udf\" WITH ('file' = 'main.jar', 'lib' = 'dependency.jar');",
    [
        "/path/to/main.jar",
        {"path": "/path/to/lib.jar", "name": "dependency.jar"}
    ]
)

Authentication

The connector uses API token authentication. You can obtain an API token from the DeltaStream platform by running CREATE API_TOKEN api_token_name; using the console.

Support

For support, please contact support@deltastream.com or open an issue on our GitHub repository.