Release v0.6.0

Author: dimamik

README.md

@@ -17,7 +17,7 @@ The package can be installed by adding `torus` to your list of dependencies in `
 ```elixir
 def deps do
   [
-    {:torus, "~> 0.5"}
+    {:torus, "~> 0.6"}
   ]
 end
 ```
@@ -72,7 +72,7 @@ See [`full_text/5`](https://hexdocs.pm/torus/Torus.html#full_text/5) for more de
 
    See [`similarity/5`](https://hexdocs.pm/torus/Torus.html#similarity/5) for more details.
 
-1. **Full text**: Uses term-document matrix vectors for, enabling efficient querying and ranking based on term frequency. Supports prefix search and is great for large datasets to quickly return relevant results. See [PostgreSQL Full Text Search](https://www.postgresql.org/docs/current/textsearch.html) for internal implementation details.
+1. **Full text**: Uses term-document matrix vectors, enabling efficient querying and ranking based on term frequency. Supports prefix search and is great for large datasets to quickly return relevant results. See [PostgreSQL Full Text Search](https://www.postgresql.org/docs/current/textsearch.html) for internal implementation details.
 
    ```elixir
    insert_post!(title: "Hogwarts Shocker", body: "A spell disrupts the Quidditch Cup.")

config/config.exs

@@ -14,11 +14,11 @@ config :torus, Torus.Test.Repo,
 
 config :torus, ecto_repos: [Torus.Test.Repo]
 
-config :torus, batcher: Torus.Embeddings.Batcher
+config :torus, embedding_module: Torus.Embeddings.Batcher
 
 config :torus, Torus.Embeddings.Batcher,
   max_batch_size: 10,
-  default_batch_timeout: 100,
+  batch_timeout: 100,
   embedding_module: Torus.Embeddings.HuggingFace
 
 if Mix.env() == :test do

guides/semantic_search.md

@@ -110,7 +110,7 @@ To use it:
   end
   ```
 
-- Add an API token for Hugging Face to your `runtime.exs`. You can get your token [here](https://aistudio.google.com/app/apikey).
+- Add an API token for Gemini to your `runtime.exs`. You can get your token [here](https://aistudio.google.com/app/apikey).
 
   ```elixir
   # `config/runtime.exs`
@@ -151,7 +151,7 @@ To use it:
   config :torus, Torus.Embeddings.OpenAI, token: System.get_env("OPEN_AI_API_KEY")
   ```
 
-By default, it uses the `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` in the configuration or when calling `Torus.to_vector/1` function via `model` option:
+By default, it uses the `text-embedding-ada-002` model, but you can specify a different model by explicitly passing `model` in the configuration or when calling `Torus.to_vector/1` function via `model` option:
 
 ```elixir
 config :torus, Torus.Embeddings.OpenAI, model: "your/model"
@@ -168,7 +168,7 @@ config :torus, embedding_module: Torus.Embeddings.PostgresML
 config :torus, Torus.Embeddings.PostgresML, repo: YourApp.Repo
 ```
 
-By default, it uses the `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` in the configuration or when calling `Torus.to_vector/1` function via `model` option:
+By default, it uses the `sentence-transformers/paraphrase-MiniLM-L3-v2` model, but you can specify a different model by explicitly passing `model` in the configuration or when calling `Torus.to_vector/1` function via `model` option:
 
 ```elixir
 config :torus, Torus.Embeddings.PostgresML, model: "your/model"
@@ -216,7 +216,7 @@ To use it:
 
 You can pass all options directly to the `Nx.Serving.start_link/1` function by passing them to `Torus.Embeddings.LocalNxServing` when starting.
 
-By default, it uses the `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` in the configuration:
+By default, it uses the `sentence-transformers/paraphrase-MiniLM-L3-v2` model, but you can specify a different model by explicitly passing `model` in the configuration:
 
 ```elixir
 config :torus, Torus.Embeddings.LocalNxServing, model: "your/model"
@@ -237,7 +237,7 @@ To use it:
 
   config :torus, Torus.Embeddings.Batcher,
     max_batch_size: 10,
-    default_batch_timeout: 100,
+    batch_timeout: 100,
     embedding_module: Torus.Embeddings.HuggingFace
   ```
 

lib/torus.ex

@@ -307,7 +307,7 @@ defmodule Torus do
       - `0` (default for `ts_rank`) - ignores the document length
       - `1`  - divides the rank by 1 + the logarithm of the document length
       - `2`  - divides the rank by the document length
-      - `4` (default for `ts_rank_cs`)  - divides the rank by the mean harmonic
+      - `4` (default for `ts_rank_cd`)  - divides the rank by the mean harmonic
       distance between extents (this is implemented only by `ts_rank_cd`)
       - `8`  - divides the rank by the number of unique words in document
       - `16` -  divides the rank by 1 + the logarithm of the number of unique words in
@@ -467,29 +467,6 @@ defmodule Torus do
   - You're on PostgreSQL < 17
   - You need the `concat` filter type
 
-  ## Multi-column search workaround
-
-  Since BM25 indexes work on single columns, you can create a generated column:
-
-  ```sql
-  ALTER TABLE posts
-  ADD COLUMN searchable_text TEXT
-  GENERATED ALWAYS AS (title || ' ' || body) STORED;
-
-  CREATE INDEX posts_searchable_bm25_idx
-  ON posts USING bm25(searchable_text)
-  WITH (text_config='english');
-  ```
-
-  Then search the generated column:
-
-  ```elixir
-  Post
-  |> Torus.bm25([p], p.searchable_text, "search term")
-  |> limit(10)
-  |> Repo.all()
-  ```
-
   ## Index options
 
   BM25 indexes support these parameters in the `WITH` clause:

lib/torus/embeddings/batch_server.ex

@@ -39,11 +39,11 @@ defmodule Torus.Embeddings.Batcher do
   - Add the following to your `config.exs`:
 
   ```elixir
-   config :torus, batcher: Torus.Embeddings.Batcher
+   config :torus, embedding_module: Torus.Embeddings.Batcher
 
    config :torus, Torus.Embeddings.Batcher,
       max_batch_size: 10,
-      default_batch_timeout: 100,
+      batch_timeout: 100,
       embedding_module: Torus.Embeddings.HuggingFace
   ```
 

lib/torus/embeddings/hugging_face.ex

@@ -21,7 +21,7 @@ if Code.ensure_loaded?(Req) do
       end
       ```
 
-    - Add an API token for hugging face to your `runtime.exs`. You can get your token [here](https://huggingface.co/settings/tokens).
+    - Add an API token for Hugging Face to your `runtime.exs`. You can get your token [here](https://huggingface.co/settings/tokens).
 
       ```elixir
       config :torus, Torus.Embeddings.HuggingFace, token: System.get_env("HUGGING_FACE_API_KEY")

lib/torus/embeddings/local_nx_serving.ex

@@ -42,7 +42,7 @@ if Code.ensure_loaded?(Bumblebee) and Code.ensure_loaded?(Nx) do
 
     You can pass all options directly to `Nx.Serving.start_link/1` function by passing them to `Torus.Embeddings.LocalNxServing` when starting.
 
-    By default, it uses `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` to the config:
+    By default, it uses `sentence-transformers/paraphrase-MiniLM-L3-v2` model, but you can specify a different model by explicitly passing `model` to the config:
 
     ```elixir
     config :torus, Torus.Embeddings.LocalNxServing, model: "your/model"
@@ -115,7 +115,7 @@ else
     @error_message """
     `Torus.Embeddings.LocalNxServing` is not available. Please add `:bumblebee` and `:nx` to your dependencies.
 
-    See `Torus.SemanticSearch` docs for more info.
+    See `Torus.semantic/5` docs for more info.
     """
 
     def child_spec(_opts) do

lib/torus/embeddings/nebulex_cache.ex

@@ -10,7 +10,7 @@ if Code.ensure_loaded?(Nebulex) and Code.ensure_loaded(Decorator.Decorate) do
     - Add the following to your `config.exs`:
 
       ```elixir
-      config :torus, cache: Torus.Embeddings.NebulexCache
+      config :torus, embedding_module: Torus.Embeddings.NebulexCache
       config :torus, Torus.Embeddings.NebulexCache,
         embedding_module: Torus.Embeddings.PostgresML
         cache: Nebulex.Cache,

lib/torus/embeddings/open_ai.ex

@@ -27,7 +27,7 @@ if Code.ensure_loaded?(Req) do
       config :torus, Torus.Embeddings.OpenAI, token: System.get_env("OPEN_AI_API_KEY")
       ```
 
-    By default, it uses `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` to the config:
+    By default, it uses `text-embedding-ada-002` model, but you can specify a different model by explicitly passing `model` to the config:
 
     ```elixir
     config :torus, Torus.Embeddings.OpenAI, model: "your/model"

lib/torus/embeddings/postgres_ml.ex

@@ -1,6 +1,6 @@
 defmodule Torus.Embeddings.PostgresML do
   @moduledoc """
-  `Torus.Embeddings.PostgresML` uses PostgreSQL [PostgresML extension](https://PostgresML.org/docs) to generate embeddings. It allows you to generate embeddings using a variety of models and performs inference directly in the database. This would require your database to have GPU support.
+  `Torus.Embeddings.PostgresML` uses the PostgreSQL [PostgresML extension](https://postgresml.org/docs) to generate embeddings. It allows you to generate embeddings using a variety of models and performs inference directly in the database. This would require your database to have GPU support.
 
   To use it, add the following to your `config.exs`:
 
@@ -9,13 +9,13 @@ defmodule Torus.Embeddings.PostgresML do
   config :torus, Torus.Embeddings.PostgresML, repo: YourApp.Repo
   ```
 
-  By default, it uses `sentence-transformers/all-MiniLM-L6-v2` model, but you can specify a different model by explicitly passing `model` to the config:
+  By default, it uses `sentence-transformers/paraphrase-MiniLM-L3-v2` model, but you can specify a different model by explicitly passing `model` to the config:
 
   ```elixir
   config :torus, Torus.Embeddings.PostgresML, model: "your/model"
   ```
 
-  Read more about in [PostgresML](https://PostgresML.org/blog/semantic-search-in-postgres-in-15-minutes).
+  Read more in the [PostgresML blog](https://postgresml.org/blog/semantic-search-in-postgres-in-15-minutes).
 
   See `Torus.semantic/5` on how to use this module to introduce semantic search in your application.
   """