domino-domino-data-sdk

You can access Domino datasets programmatically using the DatasetClient module from the dominodatalab-data SDK to read files directly within your data science workflows. It provides a unified Python interface for dataset reads.

The best way to run SQL queries on data sources in a Domino environment is using the DataSourceClient abstraction from the dominodatalab-data SDK, which also supports accessing object stores. Authentication is auto-configured.

Yes, you can integrate Pinecone vector databases with Domino for RAG pipelines using the VectorDB module provided by the dominodatalab-data SDK. This enables vector store operations within your data workflows.

Yes, the Domino data SDK supports training data versioning for MLOps through its TrainingSets module. This allows you to manage and version ML training data directly within your Domino data science workflows.

No, you do not need to manually configure authentication when using the Domino data SDK. It assumes auto-configured authentication within Domino environments, eliminating boilerplate setup for data access.

Using a unified Python SDK for Domino data access prevents juggling multiple libraries and boilerplate. It consolidates SQL queries, dataset reads, training data versioning, and vector store integration into a single resource.