@sanity/embeddings-index-cli NPM

Sanity Embeddings Index CLI

Using this feature requires Sanity to send data to OpenAI.com, and Pinecone.io for storing vector interpretations of documents.

CLI for the Sanity Embeddings Index API.

The Embeddings Index API enables creating named embeddings vector indexes.

An embeddings index contains embeddings for all Sanity documents matching a configured GROQ filter in a dataset. A GROQ projection is applied to matching documents before vectorization.

You can query indexes using semantic text search to obtain a list of matching document IDs sorted by relevance.

When an index is first created, all documents matching the configured filter are synced into the index. Creating an index can take time, depending on the number of existing documents and the indexer load.

Installation

npm i -D @sanity/embeddings-index-cli

Usage

 $ embeddings-index --help
 
 Usage
    $ embeddings-index [--help] [--debug] <command> [<args>]

   These are common commands for CRUD operations on embeddings indexes:

     create          Create a new embeddings index
     delete          Delete an existing embeddings index
     list            List all embedding indexes
     get             Get a specific embeddings index
     query           Query an embeddings index
     version         Show the currently installed version of Embeddings Index CLI

   Run the commands in a Sanity project directory with a 'sanity.cli.(ts|js)' file.

   Options
     --verbose       Log everything. This option conflicts with --silent
     --silent        Do not print info and warning messages. This option conflicts with --verbose
     --debug         Print stack trace on errors
     --version       Output the version number
     --help          Output usage information

See the individual commands for more, ie embeddings-index create --help.