Skip to content

README.md

Latest commit

 

History

History
96 lines (74 loc) · 3.37 KB

README.md

File metadata and controls

96 lines (74 loc) · 3.37 KB
Raw

cubejs

An unofficial async Python client for Cube, making it easy to integrate your Python applications with Cube's semantic layer.

What is Cube?

CubeJS is an open-source analytics API platform that enables data engineers and application developers to build consistent, accessible, and secure data applications. It serves as a semantic layer between your data sources and applications, providing a unified API for data access.

What is a Semantic Layer?

A semantic layer is an abstraction that sits between you data warehouse and your analytical consumers. It:

  • Translates complex data structures into an easy API endpoints
  • Ensures consistent business metrics and KPIs across all applications, vizualizations, and tools
  • Centralizes data access controls and security policies
  • Optimizes queries for performance

Installation

pip install "git+https://github.com/wandercom/cubejs.git"

Usage

Basic Query Example

import asyncio
import cubejs
from loguru import logger

async def main() -> None:
    # Define the query using the request model
    request = cubejs.CubeJSRequest(
        measures=["orders.count"],
        dimensions=["orders.status"],
        time_dimensions=[
            cubejs.TimeDimension(
                dimension="orders.created_at",
                date_range="last 30 days",
                granularity="day"
            )
        ],
        limit=100
    )

    # Execute the query and get results
    response = await cubejs.get_measures(
        auth=cubejs.CubeJSAuth(
            token="your-api-token",
            host="https://your-cube-instance.example.com"
        ),
        request=request
    )
    
    logger.info(f"Query Results: {response}")

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

Client features

  • pydantic model types defined by the CubeJS API
  • tenacity handles retries and exponential backoff for server instability and continue wait errors.
  • Async requests for performance

About Wander

This client is maintained by Wander, a company revolutionizing how people experience travel and discover accommodations.

How Wander Uses a Semantic Layer

At Wander, we use a semantic layer to:

  • Centralize business metrics meaning and definition in a single repository
  • Dashboards, ML services, data pipelines and operational tools, both internal and client-facing, all access a single source of truth from our semantic layer.
  • Our data warehouse becomes a technical detail so we can focus on building great products.

Dependency management

This project uses uv with pyproject.toml for dependency management. A uv.lock lockfile is committed to pin exact resolved versions for reproducible environments.

To guard against Python supply chain attacks, the [tool.uv] section in pyproject.toml sets a 7-day dependency cooldown via exclude-newer. This prevents uv from resolving package versions published within the last week — a window during which most malicious PyPI uploads are detected and yanked before they can reach any environment.

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.