activerecord-turbopuffer-adapter 0.1.5 → 0.1.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8bb295502835167a50a215f96707b208574ac3e6641b3cd9d02c50646b2335b6
-  data.tar.gz: '09c4fcdaf392f394619a7a1b786a0be832e248edd7449a5d49e804908fd4d94d'
+  metadata.gz: dbe88ca95c2999749426d72f89dd7d0327a53d5a08bf6a0b2446340d6b17e173
+  data.tar.gz: c6c1611fe66ea9ca7dff51fde19ce3843249d8b91e955aa21ba500d9fe8efe1a
 SHA512:
-  metadata.gz: d822446b93294cbca0db8d9cb13a373588e160805d6dc1f1d8e9e88b3d6695ebff58e3cb94fae1715bf0e42850cafb6b88b807b369cc48d174dbeab893e7c35a
-  data.tar.gz: a22568481e63de3ca8a83a7367a59ac9b1e57853ed2315873c2948cd17a0bf4d4a348e77188e8d2f9deefcc25484b8e128e028948dbe11c4064adbc1909e1ffe
+  metadata.gz: 4cd6c79fd3b22a79ca2d18a8ed68765360b7e63c9a3f2d4e3bb82d644553bbf7fcd93a05489f58c3285c5967ec3f2e2f522bafa7e4c1c9debf5190b4f626c85c
+  data.tar.gz: c8f7cdae01828498d93c66d345d34bdbf5cda39ff46ff13ec7126bfe97b42bba3d9c4b2337fbce08640755f87d49d42b9283cec227f9605d6b1af02bb9b1f998

data/README.md

@@ -56,6 +56,17 @@ class Document < ApplicationRecord
 end
 ```
 
+Alternatively, turbopuffer can generate embeddings for you. Declare `embed` on a string attribute and the vector is computed on write:
+
+```ruby
+class Document < ApplicationRecord
+  turbopuffer_attribute "id",   "uuid",   not_null: 1
+  turbopuffer_attribute "body", "string", embed: "openai/text-embedding-3-small"
+end
+```
+
+turbopuffer stores the vector in a computed `embed_body` attribute. `embed` also accepts a hash (`model:`, `attribute:`, `dims:`, `dtype:`) to name the vector attribute or set its size. Embedding is billed per token by model, see https://turbopuffer.com/docs/embedding for the supported models.
+
 ### Creating records
 
 ```ruby
@@ -104,6 +115,7 @@ Document.find("018f...")
 
 Document.rank_by("vector", "ANN", query_vector).limit(10)
 Document.rank_by("text", "BM25", "quick walrus").limit(10)
+Document.rank_by("body", "ANN", ["Embed", "sea mammals"]).limit(10)
 
 Document.consistency(:eventual).where(published: true)
 

data/lib/active_record/connection_adapters/turbopuffer_adapter.rb

@@ -151,7 +151,7 @@ module ActiveRecord
           schema: model.turbopuffer_schema_hash
         }
 
-        if model.turbopuffer_attributes.any?(&:ann)
+        if model.turbopuffer_attributes.any? { |attribute| attribute.ann || attribute.embed }
           tpuf_insert_args[:distance_metric] = model.turbopuffer_distance_metric
         end
 
@@ -272,7 +272,7 @@ module ActiveRecord
 
       def turbopuffer_select(namespace, query)
         fields = if query.include_attributes == [ "*" ]
-          column_definitions(query.namespace).map(&:name)
+          column_definitions(query.namespace).reject(&:vector?).map(&:name)
         else
           query.include_attributes
         end

data/lib/turbopuffer/active_record/schema/attribute.rb

@@ -3,17 +3,22 @@ module Turbopuffer
     module Schema
       class Attribute
         TYPES = %r{\A(?:\[\])?(?:string|uuid|int|uint|float|bool|datetime)\z|\Abytes\z|\A\[\d+\](?:f16|f32|i8)\z|\A\{\}f16\z}
+        VECTOR_TYPES = %r{\A\[\d+\]|\A\{\}}
 
-        attr_reader :name, :type, :filterable, :full_text_search, :ann, :glob, :regex, :notnull
+        attr_reader :name, :type, :filterable, :full_text_search, :ann, :glob, :regex, :embed, :notnull
 
         def initialize(name, type, not_null: 0, filterable: false, full_text_search: false,
-                       ann: false, glob: false, regex: false)
+                       ann: false, glob: false, regex: false, embed: nil)
           type = type.to_s
 
           unless TYPES.match?(type)
             raise ArgumentError, "unknown turbopuffer type #{type.inspect} for attribute #{name.inspect}"
           end
 
+          if embed && type != "string"
+            raise ArgumentError, "embed is only supported on string attributes, #{name.inspect} is #{type.inspect}"
+          end
+
           @name = name
           @type = type
           @filterable = filterable
@@ -21,9 +26,12 @@ module Turbopuffer
           @ann = ann
           @glob = glob
           @regex = regex
+          @embed = embed
 
           @notnull = not_null
         end
+
+        def vector? = VECTOR_TYPES.match?(type)
       end
     end
   end

data/lib/turbopuffer/active_record/schema/model.rb

@@ -48,10 +48,10 @@ module Turbopuffer
         end
 
         class_methods do
-          def turbopuffer_attribute(name, type, not_null: 0, filterable: false,
-                                    full_text_search: false, ann: false, glob: false, regex: false)
+          def turbopuffer_attribute(name, type, not_null: 0, filterable: false, full_text_search: false,
+                                    ann: false, glob: false, regex: false, embed: nil)
             attribute = ::Turbopuffer::ActiveRecord::Schema::Attribute.new(
-              name.to_s, type, not_null:, filterable:, full_text_search:, ann:, glob:, regex:
+              name.to_s, type, not_null:, filterable:, full_text_search:, ann:, glob:, regex:, embed:
             )
 
             self.turbopuffer_attributes =
@@ -78,6 +78,10 @@ module Turbopuffer
             order(::Arel::Nodes::RankByNode.new(expression))
           end
 
+          def with_vectors
+            select(*turbopuffer_attributes.map(&:name))
+          end
+
           def glob(pattern, case_sensitive: true)
             ::Turbopuffer::ActiveRecord::Glob.new(pattern, case_sensitive:)
           end
@@ -115,6 +119,7 @@ module Turbopuffer
               attrs[:ann] = true if attribute.ann
               attrs[:glob] = true if attribute.glob
               attrs[:regex] = true if attribute.regex
+              attrs[:embed] = attribute.embed if attribute.embed
               attrs[:sparse_knn] = { distance_metric: "dot_product" } if attribute.type == "{}f16"
 
               schema[attribute.name] = attrs

data/lib/turbopuffer/active_record/version.rb

@@ -1,5 +1,5 @@
 module Turbopuffer
   module ActiveRecord
-    VERSION = "0.1.5"
+    VERSION = "0.1.7"
   end
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification
 name: activerecord-turbopuffer-adapter
 version: !ruby/object:Gem::Version
-  version: 0.1.5
+  version: 0.1.7
 platform: ruby
 authors:
 - Richard Monette
 bindir: bin
 cert_chain: []
-date: 2026-09-23 00:00:00.000000000 Z
+date: 2026-10-06 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activerecord