Query Embeddings

Description

This operation searches the embeddings from an index. It retrieves the IDs, vectors, and associated metadata (e.g. text) of the most similar items in an index, along with their similarity scores.

Project Compatibility

Windows | Cross-platform

Configuration

• ConnectionId - The connection established in Integration Service. Access the drop-down to choose, add or manage connections.
• Embeddings - The query vectors you want to find similarities for. This should be the same length as the dimension of the index being queried. Check what model was used to create the embeddings.
• Number of results - The number of results you want to return for each query.
• Namespace - The namespace is a subset within the index that the Pinecone connection is associated with. Namespaces can be used for indexing different types of content. If not identified, the entire index is referenced.

Manage Properties Wizard

The Manage Properties Wizard can be used to configure or use any of the object’s standard or custom fields by selecting and adding them to the activity canvas. The added standard/custom fields are available in the Properties panel (in Studio Desktop) or under Additional options (in Studio Web).

Advanced

Options

• Include values - Indicates whether vector values are included in the response.
• Include metadata - Indicates whether metadata is included in the response with Vector IDs.

Output

• IDs - The IDs of topK matches.
• Scores - The score of topK matches.
• Query Embeddings - The query embeddings. Within this object is the metadata associated with the similar vectors. This is where things like text may be depending on the schema.