Google Vertex AI embedding provider

Uses Google Vertex AI's embedding API for embedding generation. Provides access to Google's embedding models including Gecko and newer models.

Requirements

- A Google Cloud project with Vertex AI API enabled - Authentication via access token or API key

Option 1: Access Token (Recommended for development)

Get a token with: gcloud auth print-access-token - Set GOOGLE_ACCESS_TOKEN environment variable - Or pass access_token => <<"ya29...">>> in config

Option 2: API Key

Create an API key in Google Cloud Console - Set GOOGLE_API_KEY environment variable - Or pass api_key => <<"...">>> in config

Configuration

   Config = #{
       access_token => <<"ya29...">>,                  %% Access token (or use env var)
       project => <<"my-project">>,                    %% GCP project ID (required)
       region => <<"us-central1">>,                    %% GCP region (default)
       model => <<"text-embedding-004">>,              %% Model name (default, 768 dims)
       timeout => 30000,                               %% Timeout in ms (default)
       dimension => 768                                %% Vector dimension (default)
   }.

Supported Models

- <<"text-embedding-004">> - Default, 768 dims, latest - <<"text-embedding-005">> - 768 dims, newest - <<"textembedding-gecko@001">> - 768 dims, legacy - <<"textembedding-gecko@003">> - 768 dims - <<"textembedding-gecko-multilingual@001">> - 768 dims, multilingual - <<"text-multilingual-embedding-002">> - 768 dims, multilingual