DigitalOcean Embeddings for LlamaIndex

This package provides a DigitalOceanEmbeddings implementation for LlamaIndex that calls DigitalOcean AI embeddings endpoints.

Quickstart

1. Install:

   pip install llama-index-digitalocean-embeddings
   

2. Set environment variable:

   • DIGITALOCEAN_TOKEN: DigitalOcean Personal Access Token with AI access.

   • Getting a DigitalOcean token

     Your DIGITALOCEAN_TOKEN is a Personal Access Token you create in the DigitalOcean Control Panel:

     a. Log in to https://cloud.digitalocean.com. b. Go to Settings → API → Personal Access Tokens. c. Click Generate New Token, choose the scopes you need (for embeddings, typically CRUD for GenAI and read for project), give the token a name, and click Generate. d. Copy the token and set it in your shell:

     export DIGITALOCEAN_TOKEN=your_token_here
     

     For more detailed, step‑by‑step guidance, see the official DigitalOcean documentation.

3. Use:

   import os
   from llama_index.digitalocean.embeddings import DigitalOceanEmbeddings
   from llama_index.core import VectorStoreIndex, Document
   
   # Read your token however you like (for example from an env var) and
   # pass it explicitly to the embedding class:
   token = os.environ["DIGITALOCEAN_TOKEN"]
   
   embed_model = DigitalOceanEmbeddings(model="text-embedding-3-small", api_token=token)
   docs = [Document(text="Hello from DigitalOcean")]
   index = VectorStoreIndex.from_documents(docs, embed_model=embed_model)
   print(index.as_query_engine().query("What was the greeting?"))
   

Configuration

• model: Embedding model name (e.g., text-embedding-3-small, text-embedding-3-large).
• api_token: DIGITALOCEAN_TOKEN.

License

Apache-2.0.