Skip to main content

Python (upcoming)

The Python client is not yet generally available. The current package on pypi.org (distribution name dataspine) ships only the configuration and token-provider scaffolding shared with the other Dataspine client libraries; ingest, outlet, and API clients are not yet exposed.

What ships today​

  • dataspine.config — ConfigLoader, behavior versions, token providers (static, file-backed, AWS SigV4 token exchange) and HOCON-style interpolation helpers. The runtime reads the same ~/.dataspine/config file and DATASPINE_* environment variables as the Rust client.
  • Dependencies: pydantic, requests, optional boto3/botocore (for the AWS SigV4 exchange).

You can use the configuration loader today to validate that your environment can talk to the Dataspine STS endpoint and to obtain a JWT for use with a direct client (e.g. raw grpcio with the gRPC wire reference, or requests/httpx with the REST wire reference).

What is upcoming​

  • An IngestClient and OutletClient based on grpcio / grpcio-tools, mirroring the Rust and Java APIs.
  • Generated per-data-product clients (placeholder package: acme-orders-client, exposing a typed OrdersClient for the com.acme.orders data product owned by the placeholder organization acme).
  • A RestApiClient for supports_raw_response operations.
  • Async-first APIs (asyncio-based) consistent with the other clients.

We will add per-language usage examples to this page once the typed client surface is published. Until then, prefer the TypeScript, Java, or Rust clients.

Provisional usage​

If you must consume Dataspine from Python today:

from dataspine.config.config_loader import ConfigLoader

config = ConfigLoader().load()

token = config.token_provider.get_token()

Use token as a Bearer token against the gRPC or REST transport directly. Treat the result as best-effort; the surface here will change before the typed client is GA.