Skip to content
Kitbay

pinecone-full-text-search

MITUpdated on Nov 11, 202616GitHub starsIncludes scripts

Pinecone's skill for building a full-text-search index with the document-schema API in the Pinecone Python SDK (10.1 or later).

What the skill teaches

Pinecone's skill for building a full-text-search index with the document-schema API in the Pinecone Python SDK (10.1 or later). It walks the agent through designing a schema with text, dense and sparse fields, letting Pinecone embed a text field itself, ingesting documents in bulk with a bundled helper script that batches upserts, reports errors and waits until the index is ready, and writing search calls that combine keyword, semantic and sparse scoring with phrase and metadata filters. It also explains how to turn a plain-language request into a query, and warns off the older vector and records APIs, which cannot hold text-search fields.

Capabilities

  • Design a Pinecone document schema for full-text search
  • Have Pinecone embed a text field on upsert and search
  • Bulk-ingest documents with a bundled helper script
  • Build keyword, semantic and hybrid search calls with filters