feat: legacy features to ruby llm (#12994)

enterprise/app/models/article_embedding.rb

@@ -19,6 +19,8 @@ class ArticleEmbedding < ApplicationRecord
 
   after_commit :update_response_embedding
 
+  delegate :account_id, to: :article
+
   private
 
   def update_response_embedding

enterprise/app/models/captain/assistant_response.rb

@@ -44,8 +44,8 @@ class Captain::AssistantResponse < ApplicationRecord
 
   enum status: { pending: 0, approved: 1 }
 
-  def self.search(query)
-    embedding = Captain::Llm::EmbeddingService.new.get_embedding(query)
+  def self.search(query, account_id: nil)
+    embedding = Captain::Llm::EmbeddingService.new(account_id: account_id).get_embedding(query)
     nearest_neighbors(:embedding, embedding, distance: 'cosine').limit(5)
   end
 

enterprise/app/models/concerns/agentable.rb

@@ -43,7 +43,7 @@ module Concerns::Agentable
   end
 
   def agent_model
-    InstallationConfig.find_by(name: 'CAPTAIN_OPEN_AI_MODEL')&.value.presence || OpenAiConstants::DEFAULT_MODEL
+    InstallationConfig.find_by(name: 'CAPTAIN_OPEN_AI_MODEL')&.value.presence || LlmConstants::DEFAULT_MODEL
   end
 
   def agent_response_schema

enterprise/app/models/enterprise/concerns/article.rb

@@ -11,7 +11,7 @@ module Enterprise::Concerns::Article
     add_article_embedding_association
 
     def self.vector_search(params)
-      embedding = Captain::Llm::EmbeddingService.new.get_embedding(params['query'])
+      embedding = Captain::Llm::EmbeddingService.new(account_id: params[:account_id]).get_embedding(params['query'])
       records = joins(
         :category
       ).search_by_category_slug(