Hey!

Semantic search is finally here! Read more in Semantic search with Torus guide.
Shortly - it allows you to generate embeddings using a configurable (and chainable) adapters and use them to compare against the ones stored in your database.

Supported adapters (for now):

• Torus.Embeddings.OpenAI - uses OpenAI’s API to generate embeddings.

• Torus.Embeddings.HuggingFace - uses HuggingFace’s API to generate embeddings.

• Torus.Embeddings.LocalNxServing - generate embeddings on your local machine using a variety of models available on Hugging Face

• Torus.Embeddings.PostgresML - uses PostgreSQL PostgresML extension to generate embeddings

• Torus.Embeddings.Batcher - a long‑running GenServer that collects individual embedding calls, groups them into a single batch, and forwards the batch to the configured embedding_module (any from the above or your custom one).

• Torus.Embeddings.NebulexCache - a wrapper around Nebulex cache, allowing you to cache the embedding calls in memory, so you save the resources/cost of calling the embedding module multiple times for the same input.

And you can easily create your own adapter by implementing the Torus.Embedding behaviour.

So after you’ll pick your favorite embedding adapter, you can add semantic search:

def search(term) do
  search_vector = Torus.to_vector(term)

  Post
  |> Torus.semantic([p], p.embedding, search_vector, distance: :l2_distance, pre_filter: 0.7)
  |> Repo.all()
end

Future plans:

• Release torus_example app that would allow experimenting with options and different search types to pick the best one
• Add support for highlighting search results. (Base off of a ts_headline function)
• Extend similarity search to support fuzzystrmatch extension distance options.

Links

• Semantic search — Torus v0.4.0
• GitHub - dimamik/torus: Torus bridges Ecto and PostgreSQL, simplifying building search queries.

Torus v0.5.2 is released!

New

• New demo page where you can explore different search types and their options. It also includes semantic search, so if you’re hesitant - go check it out!
• Other documentation improvements

Fixes

• Correctly handles order: :none in Torus.semantic/5 search.
• Updates Torus.Embeddings.HuggingFace to point to the updated feature extraction endpoint.
• Suppresses warnings for missing ecto_sql dependency by adding it to the required dependencies. Most of us already had it, but now it’ll be explicit.
• Correctly parses an array of integers in Torus.QueryInspector.substituted_sql/3 and Torus.QueryInspector.tap_substituted_sql/3. Now we should be able to handle all possible query variations.

What is the difference between the ecto like/ilike and the one provided by your library?

There are a few options, either in postgresql.conf or per-database in a migration:

defmodule YourApp.Repo.Migrations.SetWordSimilarityThreshold do
  use Ecto.Migration

  @database Application.compile_env!(:your_app_name, YourApp.Repo)[:database]

  def up do
    execute("""
    ALTER DATABASE #{@database} SET pg_trgm.word_similarity_threshold = 0.5;
    """)
  end

  def down do
    execute("""
    ALTER DATABASE #{@database} RESET pg_trgm.word_similarity_threshold;
    """)
  end
end

Looks interesting! I saw that all the examples were using “function-style queries.” Does Torus support “macro-style” queries? If not, are you planning to add support for them?