DEV: Improve embedding configs validations (#1101)

Before this change, we let you set the embeddings selected model back to " " even with embeddings enabled. This will leave the site in a broken state.

Additionally, it adds a fail-safe for these scenarios to avoid errors on the topics page.

config/locales/server.en.yml

@@ -434,6 +434,7 @@ en:
       cannot_edit_builtin: "You can't edit a built-in model."
       configuration:
         disable_embeddings: "You have to disable 'ai embeddings enabled' first."
+        invalid_config: "You selected a invalid option."
         choose_model: "Set 'ai embeddings selected model' first."
 
     llm_models:

lib/configuration/embedding_defs_validator.rb

@@ -8,10 +8,19 @@ module DiscourseAi
       end
 
       def valid_value?(val)
-        val.blank? || EmbeddingDefinition.exists?(id: val)
+        if val.blank?
+          @module_enabled = SiteSetting.ai_embeddings_enabled
+
+          !@module_enabled
+        else
+          EmbeddingDefinition.exists?(id: val).tap { |def_exists| @invalid_option = !def_exists }
+        end
       end
 
       def error_message
+        return I18n.t("discourse_ai.embeddings.configuration.disable_embeddings") if @module_enabled
+        return I18n.t("discourse_ai.embeddings.configuration.invalid_config") if @invalid_option
+
         ""
       end
     end

lib/embeddings/semantic_related.rb

@@ -10,6 +10,7 @@ module DiscourseAi
 
       def related_topic_ids_for(topic)
         return [] if SiteSetting.ai_embeddings_semantic_related_topics < 1
+        return [] if SiteSetting.ai_embeddings_selected_model.blank? # fail-safe in case something end up in a broken state.
 
         cache_for = results_ttl(topic)
 

lib/embeddings/semantic_search.rb

@@ -10,8 +10,8 @@ module DiscourseAi
           "semantic-search-#{digest}-#{SiteSetting.ai_embeddings_semantic_search_hyde_model}"
 
         Discourse.cache.delete(hyde_key)
-        Discourse.cache.delete("#{hyde_key}-#{SiteSetting.ai_embeddings_model}")
-        Discourse.cache.delete("-#{SiteSetting.ai_embeddings_model}")
+        Discourse.cache.delete("#{hyde_key}-#{SiteSetting.ai_embeddings_selected_model}")
+        Discourse.cache.delete("-#{SiteSetting.ai_embeddings_selected_model}")
       end
 
       def initialize(guardian)
@@ -24,7 +24,7 @@ module DiscourseAi
           build_embedding_key(
             digest,
             SiteSetting.ai_embeddings_semantic_search_hyde_model,
-            SiteSetting.ai_embeddings_model,
+            SiteSetting.ai_embeddings_selected_model,
           )
 
         Discourse.cache.read(embedding_key).present?
@@ -42,7 +42,7 @@ module DiscourseAi
           build_embedding_key(
             digest,
             SiteSetting.ai_embeddings_semantic_search_hyde_model,
-            SiteSetting.ai_embeddings_model,
+            SiteSetting.ai_embeddings_selected_model,
           )
 
         hypothetical_post =
@@ -57,7 +57,7 @@ module DiscourseAi
 
       def embedding(search_term)
         digest = OpenSSL::Digest::SHA1.hexdigest(search_term)
-        embedding_key = build_embedding_key(digest, "", SiteSetting.ai_embeddings_model)
+        embedding_key = build_embedding_key(digest, "", SiteSetting.ai_embeddings_selected_model)
 
         Discourse.cache.fetch(embedding_key, expires_in: 1.week) { vector.vector_from(search_term) }
       end
@@ -120,7 +120,7 @@ module DiscourseAi
           build_embedding_key(
             digest,
             SiteSetting.ai_embeddings_semantic_search_hyde_model,
-            SiteSetting.ai_embeddings_model,
+            SiteSetting.ai_embeddings_selected_model,
           )
 
         search_term_embedding =

spec/models/rag_document_fragment_spec.rb

@@ -97,7 +97,7 @@ RSpec.describe RagDocumentFragment do
       vector.generate_representation_from(rag_document_fragment_1)
     end
 
-    it "regenerates all embeddings if ai_embeddings_model changes" do
+    it "regenerates all embeddings if ai_embeddings_selected_model changes" do
       old_id = rag_document_fragment_1.id
 
       UploadReference.create!(upload_id: upload_1.id, target: persona)