Add Milvus#remove_texts() method (#689)

* Add Milvus#remove_texts() method

* Bump milvus gem

Gemfile.lock

@@ -214,7 +214,7 @@ GEM
       net-smtp
     matrix (0.4.2)
     method_source (1.1.0)
-    milvus (0.9.2)
+    milvus (0.9.3)
       faraday (>= 2.0.1, < 3)
     mini_mime (1.1.5)
     mini_portile2 (2.8.6)
@@ -479,7 +479,7 @@ DEPENDENCIES
   langchainrb!
   llama_cpp (~> 0.9.4)
   mail (~> 2.8)
-  milvus (~> 0.9.2)
+  milvus (~> 0.9.3)
   mistral-ai
   nokogiri (~> 1.13)
   open-weather-ruby-client (~> 0.4.0)

README.md

@@ -343,7 +343,7 @@ You can instantiate any other supported vector search database:
 client = Langchain::Vectorsearch::Chroma.new(...)   # `gem "chroma-db", "~> 0.6.0"`
 client = Langchain::Vectorsearch::Epsilla.new(...)  # `gem "epsilla-ruby", "~> 0.0.3"`
 client = Langchain::Vectorsearch::Hnswlib.new(...)  # `gem "hnswlib", "~> 0.8.1"`
-client = Langchain::Vectorsearch::Milvus.new(...)   # `gem "milvus", "~> 0.9.2"`
+client = Langchain::Vectorsearch::Milvus.new(...)   # `gem "milvus", "~> 0.9.3"`
 client = Langchain::Vectorsearch::Pinecone.new(...) # `gem "pinecone", "~> 0.1.6"`
 client = Langchain::Vectorsearch::Pgvector.new(...) # `gem "pgvector", "~> 0.2"`
 client = Langchain::Vectorsearch::Qdrant.new(...)   # `gem "qdrant-ruby", "~> 0.9.3"`

langchain.gemspec

@@ -59,7 +59,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "google_search_results", "~> 2.0.0"
   spec.add_development_dependency "hnswlib", "~> 0.8.1"
   spec.add_development_dependency "hugging-face", "~> 0.3.4"
-  spec.add_development_dependency "milvus", "~> 0.9.2"
+  spec.add_development_dependency "milvus", "~> 0.9.3"
   spec.add_development_dependency "llama_cpp", "~> 0.9.4"
   spec.add_development_dependency "nokogiri", "~> 1.13"
   spec.add_development_dependency "mail", "~> 2.8"

lib/langchain/vectorsearch/milvus.rb

@@ -6,7 +6,7 @@ class Milvus < Base
     # Wrapper around Milvus REST APIs.
     #
     # Gem requirements:
-    #     gem "milvus", "~> 0.9.2"
+    #     gem "milvus", "~> 0.9.3"
     #
     # Usage:
     # milvus = Langchain::Vectorsearch::Milvus.new(url:, index_name:, llm:, api_key:)
@@ -39,6 +39,21 @@ def add_texts(texts:)
       )
     end
 
+    # Deletes a list of texts in the index
+    #
+    # @param ids [Array<Integer>] The ids of texts to delete
+    # @return [Boolean] The response from the server
+    def remove_texts(ids:)
+      raise ArgumentError, "ids must be an array" unless ids.is_a?(Array)
+      # Convert ids to integers if strings are passed
+      ids = ids.map(&:to_i)
+
+      client.entities.delete(
+        collection_name: index_name,
+        expression: "id in #{ids}"
+      )
+    end
+
     # TODO: Add update_texts method
 
     # Create default schema
@@ -83,7 +98,7 @@ def create_default_schema
     # @return [Boolean] The response from the server
     def create_default_index
       client.indices.create(
-        collection_name: "Documents",
+        collection_name: index_name,
         field_name: "vectors",
         extra_params: [
           {key: "metric_type", value: "L2"},

spec/langchain/vectorsearch/milvus_spec.rb

@@ -89,6 +89,18 @@
     end
   end
 
+  describe "remove_texts" do
+    let(:response) { {"status" => {}, "IDs" => {"IdField" => nil}, "delete_cnt" => 1} }
+
+    before do
+      allow(subject.client).to receive_message_chain(:entities, :delete).and_return(response)
+    end
+
+    it "adds texts" do
+      expect(subject.remove_texts(ids: [450847466900986695])).to eq(response)
+    end
+  end
+
   describe "#similarity_search_by_vector" do
     before do
       allow(subject.client.collections).to receive(:load).and_return(true)