Get Best Available Embeddings

Auto-detects and uses the best available embedding provider with the following priority:

1. sentence-transformers (local Python) - if Python environment is set up

2. OpenAI API - if OPENAI_API_KEY is set

3. Gemini API - if GEMINI_API_KEY is set

Usage

get_best_embeddings(
  texts,
  provider = "auto",
  model = NULL,
  api_key = NULL,
  verbose = TRUE
)

Arguments

texts

Character vector of texts to embed

provider

Character string: "auto" (default), "sentence-transformers", "openai", or "gemini". Use "auto" for automatic detection.

model

Character string specifying the embedding model. If NULL, uses default model for the selected provider.

api_key

Optional API key for OpenAI or Gemini providers. If NULL, falls back to environment variables (OPENAI_API_KEY, GEMINI_API_KEY).

verbose

Logical, whether to print progress messages (default: TRUE)

Value

Matrix with embeddings (rows = texts, columns = dimensions)

Examples

if (interactive()) {
data(SpecialEduTech)
texts <- SpecialEduTech$abstract[1:5]

# Auto-detect best available provider
embeddings <- get_best_embeddings(texts)

# Force specific provider
embeddings <- get_best_embeddings(texts, provider = "openai")

dim(embeddings)
}
#> Using sentence-transformers embeddings (local Python)
#> Generating embeddings using model: all-MiniLM-L6-v2
#> Embeddings generated successfully
#> No OpenAI API key found. Provide your key using one of these methods:
#>   1. Sys.setenv(OPENAI_API_KEY = "your-key-here")
#>   2. Add OPENAI_API_KEY=your-key-here to your .Renviron file
#>   3. Pass directly via the api_key parameter
#> NULL