llm_meta_client 0.6.1 → 1.0.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 931f6b078e6df954f5ab32218cb31c4c743112ed6ae037c3e027102cec3089d4
-  data.tar.gz: 52febe0e221d964ad286ce8ad7ebb0ad595a8d80b87a4afa55af54e82e82873d
+  metadata.gz: 95bb1f421088cc0c287b4f79e961455a1a5275b860b372fb23523ddc8781d37a
+  data.tar.gz: d4f92a150de0d9996d3ea6c54e69fe24989763f4bd359ce46a6b7ff69d36ecb7
 SHA512:
-  metadata.gz: 6216c95f340ae0ef1b3e9ee54bd8a0906c2001d50ac45139df4f8da584c856962b462182e7fcc20e94c8785e6923f50f04882bf534da300bc8946fef5308cfa1
-  data.tar.gz: 2223f1a7381b9bdb484a1ec8fade18c5b83789b007b6f078680887a81db4c6a5854572310a108864eba89656cff980359b9602e14a9cdae752d8d0ad98d7a48e
+  metadata.gz: 480919bad702190fec4166d8a2659121ff62b1880123e6db168d9a1c8a8b7f80f23ca175e2e6f4a31b96e6bde97fd0723d95caa71439845a5017c351c18e373e
+  data.tar.gz: c2600201eccae124c9929eabcd8b7d755b7dbac4092cd5441c55d1caf79c839af8f0b41b49d7bf1c2e6a324e33c6bd1957d29a1ac3e837a7ab6924d727e8ae91

data/CHANGELOG.md

@@ -5,6 +5,26 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.0.0] - 2026-03-25
+
+### Changed
+
+- Replace generation settings UI from individual sliders to JSON textarea input
+- Improve prompt execution branching logic to use `execution_id` instead of message UUID
+- Use `find_by!` for proper 404 handling in controllers
+- Use URL-based chat lookup in `chats#update` instead of session-based lookup
+- Keep existing chat when switching model or LLM (update instead of creating new chat)
+- Upgrade `prompt_navigator` dependency to `~> 1.0`
+- Upgrade `chat_manager` dependency to `~> 1.0`
+
+### Fixed
+
+- Fix Turbo Stream history sidebar element ID mismatch (`history-content` → `history-sidebar`)
+- Fix `next_ann` for proper history card rendering
+- Wrap inline JavaScript in IIFE to prevent variable conflicts across Turbo Stream updates
+- Fix scroll event listener duplication across Turbo navigations
+- Validate generation settings JSON input before sending to LLM
+
 ## [0.6.1] - 2026-03-19
 
 ### Fixed

data/app/assets/stylesheets/llm_meta_client/generation_settings.css

@@ -31,60 +31,39 @@
   padding: 12px 16px;
 }
 
-.generation-setting-item {
-  margin-bottom: 14px;
-
-  &:last-child {
-    margin-bottom: 0;
-  }
+.generation-settings-label {
+  display: block;
+  font-size: 13px;
+  font-weight: 600;
+  color: #374151;
+  margin-bottom: 6px;
+}
 
-  label {
-    display: flex;
-    align-items: center;
-    gap: 8px;
-    font-size: 13px;
-    font-weight: 600;
-    color: #374151;
-    margin-bottom: 4px;
-  }
+.generation-settings-json-input {
+  width: 100%;
+  padding: 8px 10px;
+  border: 1px solid #d1d5db;
+  border-radius: 6px;
+  font-size: 13px;
+  font-family: "SFMono-Regular", "Consolas", "Liberation Mono", "Menlo", monospace;
+  background-color: white;
+  resize: vertical;
+  transition: border-color 0.2s;
+  box-sizing: border-box;
 
-  input[type="range"] {
-    width: 100%;
-    accent-color: #3b82f6;
-    cursor: pointer;
+  &:focus {
+    outline: none;
+    border-color: #3b82f6;
+    box-shadow: 0 0 0 3px rgba(59, 130, 246, 0.1);
   }
 
-  .max-tokens-input {
-    width: 100%;
-    padding: 6px 10px;
-    border: 1px solid #d1d5db;
-    border-radius: 6px;
-    font-size: 13px;
-    background-color: white;
-    transition: border-color 0.2s;
-
-    &:focus {
-      outline: none;
-      border-color: #3b82f6;
-      box-shadow: 0 0 0 3px rgba(59, 130, 246, 0.1);
-    }
-
-    &::placeholder {
-      color: #9ca3af;
-    }
+  &::placeholder {
+    color: #9ca3af;
   }
 }
 
-.setting-value {
-  font-weight: 700;
-  color: #3b82f6;
-  font-size: 13px;
-}
-
-.setting-range-labels {
-  display: flex;
-  justify-content: space-between;
+.generation-settings-hint {
   font-size: 11px;
   color: #9ca3af;
-  margin-top: 2px;
+  margin-top: 4px;
 }

data/lib/generators/llm_meta_client/scaffold/scaffold_generator.rb

@@ -56,6 +56,7 @@ module LlmMetaClient
       def add_migrations
         migration_template "db/migrate/create_chats.rb", "db/migrate/create_chats.rb"
         migration_template "db/migrate/create_messages.rb", "db/migrate/create_messages.rb"
+        migration_template "db/migrate/migrate_llm_uuid_to_prompt_executions.rb", "db/migrate/migrate_llm_uuid_to_prompt_executions.rb"
       end
 
       def configure_routes

data/lib/generators/llm_meta_client/scaffold/templates/app/controllers/chats_controller.rb

@@ -10,7 +10,8 @@ class ChatsController < ApplicationController
     # Initialize chat context
     initialize_chat current_user&.chats
 
-    @chat = current_user&.chats.includes(:messages).find_by(uuid: params[:id])
+    @chat = current_user&.chats.includes(:messages).find_by!(uuid: params[:id])
+    session[:chat_id] = @chat.id
     @messages = @chat.ordered_messages
 
     # Initialize history
@@ -60,9 +61,7 @@ class ChatsController < ApplicationController
     # Find or create chat
     @chat = Chat.find_or_switch_for_session(
       session,
-      current_user,
-      llm_uuid: params[:api_key_uuid],
-      model: params[:model]
+      current_user
     )
     add_chat @chat
     @messages = @chat&.ordered_messages || []
@@ -71,8 +70,21 @@ class ChatsController < ApplicationController
     initialize_history @chat&.ordered_by_descending_prompt_executions
 
     if params[:message].present?
+      # Validate generation settings before proceeding
+      begin
+        generation_settings = generation_settings_param
+      rescue InvalidGenerationSettingsError => e
+        @error_message = e.message
+        respond_to do |format|
+          format.turbo_stream
+          format.html { redirect_to new_chat_path, alert: e.message }
+        end
+        return
+      end
+
       # Add user message (will be rendered via turbo stream)
       @prompt_execution, @user_message = @chat.add_user_message(params[:message],
+                                                                params[:api_key_uuid],
                                                                 params[:model],
                                                                 params[:branch_from_uuid])
       # Push to history for rendering
@@ -82,7 +94,7 @@ class ChatsController < ApplicationController
 
       # Send to LLM and get assistant response
       begin
-        @assistant_message = @chat.add_assistant_response(@prompt_execution, jwt_token, tool_ids: tool_ids_param, generation_settings: generation_settings_param)
+        @assistant_message = @chat.add_assistant_response(@prompt_execution, jwt_token, tool_ids: tool_ids_param, generation_settings: generation_settings)
         # Generate chat title from the user's prompt (only if title is not yet set)
         @chat.generate_title(params[:message], jwt_token)
       rescue StandardError => e
@@ -131,9 +143,7 @@ class ChatsController < ApplicationController
 
     @chat = Chat.find_or_switch_for_session(
       session,
-      current_user,
-      llm_uuid: params[:api_key_uuid],
-      model: params[:model]
+      current_user
     )
     @messages = @chat&.ordered_messages || []
     # initialize history for the chat
@@ -149,20 +159,29 @@ class ChatsController < ApplicationController
   def update
     jwt_token = current_user.id_token if user_signed_in?
 
-    # Find or create chat
-    @chat = Chat.find_or_switch_for_session(
-      session,
-      current_user,
-      llm_uuid: params[:api_key_uuid],
-      model: params[:model]
-    )
+    # Use the chat identified by the URL, not the session
+    @chat = current_user.chats.find(params[:id])
+    session[:chat_id] = @chat.id
     @messages = @chat&.ordered_messages || []
     # initialize history for the chat
     initialize_history @chat&.ordered_by_descending_prompt_executions
 
     if params[:message].present?
+      # Validate generation settings before proceeding
+      begin
+        generation_settings = generation_settings_param
+      rescue InvalidGenerationSettingsError => e
+        @error_message = e.message
+        respond_to do |format|
+          format.turbo_stream
+          format.html { redirect_to chat_path(@chat), alert: e.message }
+        end
+        return
+      end
+
       # Add user message (will be rendered via turbo stream)
       @prompt_execution, @user_message = @chat.add_user_message(params[:message],
+                                                               params[:api_key_uuid],
                                                                params[:model],
                                                                params[:branch_from_uuid])
       # Push to history for rendering
@@ -172,7 +191,7 @@ class ChatsController < ApplicationController
 
       # Send to LLM and get assistant response
       begin
-        @assistant_message = @chat.add_assistant_response(@prompt_execution, jwt_token, tool_ids: tool_ids_param, generation_settings: generation_settings_param)
+        @assistant_message = @chat.add_assistant_response(@prompt_execution, jwt_token, tool_ids: tool_ids_param, generation_settings: generation_settings)
       rescue StandardError => e
         Rails.logger.error "Error in chat response: #{e.class} - #{e.message}\n#{e.backtrace&.join("\n")}"
         @error_message = "An error occurred while getting the response. Please try again."
@@ -192,13 +211,27 @@ class ChatsController < ApplicationController
     params[:tool_ids].presence || []
   end
 
+  ALLOWED_GENERATION_KEYS = %w[temperature top_k top_p max_tokens repeat_penalty].freeze
+
+  class InvalidGenerationSettingsError < StandardError; end
+
   def generation_settings_param
-    settings = {}
-    settings[:temperature] = params[:temperature].to_f if params[:temperature].present?
-    settings[:top_k] = params[:top_k].to_i if params[:top_k].present?
-    settings[:top_p] = params[:top_p].to_f if params[:top_p].present?
-    settings[:max_tokens] = params[:max_tokens].to_i if params[:max_tokens].present?
-    settings[:repeat_penalty] = params[:repeat_penalty].to_f if params[:repeat_penalty].present?
-    settings
+    return {} if params[:generation_settings_json].blank?
+
+    parsed = JSON.parse(params[:generation_settings_json])
+    raise InvalidGenerationSettingsError, "Generation settings must be a JSON object" unless parsed.is_a?(Hash)
+
+    settings = parsed.slice(*ALLOWED_GENERATION_KEYS)
+    invalid_keys = parsed.keys - ALLOWED_GENERATION_KEYS
+    raise InvalidGenerationSettingsError, "Unknown keys: #{invalid_keys.join(', ')}" if invalid_keys.any?
+
+    non_numeric = settings.reject { |_k, v| v.is_a?(Numeric) }
+    if non_numeric.any?
+      raise InvalidGenerationSettingsError, "Values must be numeric: #{non_numeric.keys.join(', ')}"
+    end
+
+    settings.symbolize_keys
+  rescue JSON::ParserError => e
+    raise InvalidGenerationSettingsError, "Invalid JSON: #{e.message}"
   end
 end

data/lib/generators/llm_meta_client/scaffold/templates/app/controllers/prompts_controller.rb

@@ -4,8 +4,8 @@ class PromptsController < ApplicationController
   skip_before_action :authenticate_user!, raise: false
 
   def show
-    @prompt_execution = PromptNavigator::PromptExecution.includes(:messages).find_by(execution_id: params[:id])
-    @message = @prompt_execution.messages.first
+    @prompt_execution = PromptNavigator::PromptExecution.find_by!(execution_id: params[:id])
+    @message = Message.where(prompt_navigator_prompt_execution: @prompt_execution).order(:created_at).first
     @chat = @message.chat
     @messages = @chat.ordered_messages
 
@@ -25,6 +25,9 @@ class PromptsController < ApplicationController
     # Set active UUID for history sidebar highlighting
     set_active_message_uuid(@prompt_execution.execution_id)
 
+    # Set branch_from_uuid so the form knows which message to branch from
+    @branch_from_uuid = @prompt_execution.execution_id
+
     render "chats/edit"
   rescue StandardError => e
     Rails.logger.error "Error in PromptsController#show_by_uuid: #{e.class} - #{e.message}\n#{e.backtrace&.join("\n")}"

data/lib/generators/llm_meta_client/scaffold/templates/app/javascript/controllers/generation_settings_controller.js

@@ -6,15 +6,7 @@ export default class extends Controller {
     "toggleButton",
     "toggleIcon",
     "panel",
-    "temperatureRange",
-    "temperatureValue",
-    "topKRange",
-    "topKValue",
-    "topPRange",
-    "topPValue",
-    "maxTokensInput",
-    "repeatPenaltyRange",
-    "repeatPenaltyValue",
+    "jsonInput",
   ]
 
   connect() {
@@ -32,20 +24,4 @@ export default class extends Controller {
       this.toggleIconTarget.classList.toggle("bi-chevron-up", this.expanded)
     }
   }
-
-  updateTemperature() {
-    this.temperatureValueTarget.textContent = this.temperatureRangeTarget.value
-  }
-
-  updateTopK() {
-    this.topKValueTarget.textContent = this.topKRangeTarget.value
-  }
-
-  updateTopP() {
-    this.topPValueTarget.textContent = this.topPRangeTarget.value
-  }
-
-  updateRepeatPenalty() {
-    this.repeatPenaltyValueTarget.textContent = this.repeatPenaltyRangeTarget.value
-  }
 }

data/lib/generators/llm_meta_client/scaffold/templates/app/models/chat.rb

@@ -6,25 +6,14 @@ class Chat < ApplicationRecord
 
   before_create :set_uuid
 
-  validates :llm_uuid, presence: true
-  validates :model, presence: true
-
   # Find existing chat from session or create new one
   class << self
-    def find_or_switch_for_session(session, current_user, llm_uuid: nil, model: nil)
+    def find_or_switch_for_session(session, current_user)
       chat = find_by_session_chat_id(session, current_user)
-      return chat if llm_uuid.nil? || model.nil?
-
-      # Create new chat if it doesn't exist or LLM/model has changed
-      if llm_uuid.present? && model.present? && (chat.nil? || (chat.present? && chat.needs_reset?(llm_uuid, model)))
-        chat = create!(
-          user: current_user,
-          llm_uuid: llm_uuid,
-          model: model
-        )
-        session[:chat_id] = chat.id
-      end
+      return chat if chat.present?
 
+      chat = create!(user: current_user)
+      session[:chat_id] = chat.id
       chat
     end
 
@@ -41,21 +30,19 @@ class Chat < ApplicationRecord
     end
   end
 
-  # Get the LLM type for this chat
-  def llm_type(jwt_token)
-    llm_options = LlmMetaClient::ServerResource.available_llm_options(jwt_token)
-    selected_llm = llm_options.find { |opt| opt[:uuid] == llm_uuid }
-    selected_llm&.dig(:llm_type) || "unknown"
-  end
-
   # Add a user message to the chat
-  def add_user_message(message, model, branch_from_uuid = nil)
-    parent_message = branch_from_uuid.present? ? messages.find_by(uuid: branch_from_uuid) : nil
+  def add_user_message(message, llm_uuid, model, branch_from_execution_id = nil)
+    previous_id = if branch_from_execution_id.present?
+      PromptNavigator::PromptExecution.find_by(execution_id: branch_from_execution_id)&.id
+    else
+      messages.where(role: "user").order(:created_at).last&.prompt_navigator_prompt_execution_id
+    end
     prompt_execution = PromptNavigator::PromptExecution.create!(
       prompt: message,
+      llm_uuid: llm_uuid,
       model: model,
       configuration: "",
-      previous_id: parent_message&.prompt_navigator_prompt_execution_id
+      previous_id: previous_id
     )
 
     new_message = messages.create!(
@@ -68,9 +55,9 @@ class Chat < ApplicationRecord
 
   # Add assistant response by sending to LLM
   def add_assistant_response(prompt_execution, jwt_token, tool_ids: [], generation_settings: {})
-    response_content = send_to_llm(jwt_token, tool_ids: tool_ids, generation_settings: generation_settings)
+    response_content = send_to_llm(prompt_execution, jwt_token, tool_ids: tool_ids, generation_settings: generation_settings)
     prompt_execution.update!(
-      llm_platform: llm_type(jwt_token),
+      llm_platform: resolve_llm_type(prompt_execution.llm_uuid, jwt_token),
       response: response_content
     )
     new_message = messages.create!(
@@ -98,19 +85,24 @@ class Chat < ApplicationRecord
       .map(&:prompt_navigator_prompt_execution)
   end
 
-  # Check if chat needs to be reset due to LLM or model change
-  def needs_reset?(new_llm_uuid, new_model)
-    llm_uuid != new_llm_uuid || model != new_model
-  end
-
   private
 
+  # Resolve the LLM type (e.g. "openai", "google") from a given llm_uuid
+  def resolve_llm_type(llm_uuid, jwt_token)
+    llm_options = LlmMetaClient::ServerResource.available_llm_options(jwt_token)
+    selected_llm = llm_options.find { |opt| opt[:uuid] == llm_uuid }
+    selected_llm&.dig(:llm_type) || "unknown"
+  end
+
   # Summarize the user's prompt into a short title via LLM (required by ChatManager::TitleGeneratable)
   def summarize_for_title(prompt_text, jwt_token)
+    latest_pe = ordered_by_descending_prompt_executions.first
+    return nil unless latest_pe&.llm_uuid && latest_pe&.model
+
     LlmMetaClient::ServerQuery.new.call(
       jwt_token,
-      llm_uuid,
-      model,
+      latest_pe.llm_uuid,
+      latest_pe.model,
       "No context available.",
       { role: "user", prompt: "Please summarize the following text into a short title (max 50 characters). Respond with only the title, nothing else: #{prompt_text}" }
     )
@@ -122,7 +114,10 @@ class Chat < ApplicationRecord
   end
 
   # Send messages to LLM and get response
-  def send_to_llm(jwt_token, tool_ids: [], generation_settings: {})
+  def send_to_llm(prompt_execution, jwt_token, tool_ids: [], generation_settings: {})
+    llm_uuid = prompt_execution.llm_uuid
+    model = prompt_execution.model
+
     # Get LLM options
     llm_options = LlmMetaClient::ServerResource.available_llm_options(jwt_token)
 

data/lib/generators/llm_meta_client/scaffold/templates/app/views/chats/create.turbo_stream.erb

@@ -33,12 +33,12 @@
 <%% # Update history sidebar - replace entire content to ensure update %>
 <%% if @prompt_execution %>
   <%%= turbo_stream.replace "history-sidebar" do %>
-    <div id="history-content">
+    <div id="history-sidebar">
       <h2>History</h2>
       <div class="history-stack" id="history-stack" data-controller="history">
         <%%= render 'prompt_navigator/history_card', locals: {
           ann: @prompt_execution,
-          next_ann: nil,
+          next_ann: (@chat&.ordered_by_descending_prompt_executions || [])[1],
           is_active: @prompt_execution.execution_id == @active_message_uuid,
           card_path: ->(uuid) { prompt_path(uuid) }
         } %>
@@ -60,25 +60,33 @@
 <turbo-stream action="after" target="messages-list">
   <template>
     <script>
-      // Clear and refocus message input
-      const messageInput = document.getElementById('message-input');
-      if (messageInput) {
-        messageInput.value = '';
-        messageInput.focus();
-      }
+      (function() {
+        // Clear and refocus message input
+        const messageInput = document.getElementById('message-input');
+        if (messageInput) {
+          messageInput.value = '';
+          messageInput.focus();
+        }
 
-      // Update submit button state
-      const form = document.querySelector('[data-controller="chats-form"]');
-      if (form && messageInput) {
-        const event = new Event('input', { bubbles: true });
-        messageInput.dispatchEvent(event);
-      }
+        // Update submit button state
+        const form = document.querySelector('[data-controller="chats-form"]');
+        if (form && messageInput) {
+          const event = new Event('input', { bubbles: true });
+          messageInput.dispatchEvent(event);
+        }
 
-      // Scroll to bottom
-      const chatMessages = document.getElementById('chat-messages');
-      if (chatMessages) {
-        chatMessages.scrollTop = chatMessages.scrollHeight;
-      }
+        // Update branch_from_uuid to the latest prompt execution
+        const branchField = document.getElementById('branch_from_uuid');
+        if (branchField) {
+          branchField.value = '<%%= @prompt_execution&.execution_id %>';
+        }
+
+        // Scroll to bottom
+        const chatMessages = document.getElementById('chat-messages');
+        if (chatMessages) {
+          chatMessages.scrollTop = chatMessages.scrollHeight;
+        }
+      })();
     </script>
   </template>
 </turbo-stream>

data/lib/generators/llm_meta_client/scaffold/templates/app/views/chats/edit.html.erb

@@ -30,7 +30,7 @@
             id: "message-input",
             data: { "chats-form-target": "prompt", action: "input->chats-form#updateSubmitButton" } %>
       </div>
-      <%%= f.hidden_field :branch_from_uuid, value: params.dig(:chat, :branch_from_uuid) %>
+      <%%= f.hidden_field :branch_from_uuid, value: @branch_from_uuid || params.dig(:chat, :branch_from_uuid) %>
       <div class="button-wrapper">
         <%%= f.button type: "submit",
             class: "send-button",
@@ -47,16 +47,18 @@
 
 
 <script>
-  // Function to handle scrolling for both initial load and Turbo navigation
-  function scrollChatMessages() {
+  // Remove previous listener to prevent duplicates across Turbo navigations
+  if (window._scrollChatMessages) {
+    document.removeEventListener('turbo:load', window._scrollChatMessages);
+  }
+
+  window._scrollChatMessages = function() {
     const chatMessages = document.getElementById('chat-messages');
     if (chatMessages) {
       <%% if @target_message_id.present? %>
-        // Scroll to target message
         const targetMessage = document.getElementById('message-<%%= @target_message_id %>');
         if (targetMessage) {
           targetMessage.scrollIntoView({ behavior: 'smooth', block: 'center' });
-          // Highlight the target message
           targetMessage.style.backgroundColor = '#fef3c7';
           targetMessage.style.border = '2px solid #fbbf24';
           setTimeout(() => {
@@ -69,10 +71,8 @@
         chatMessages.scrollTop = chatMessages.scrollHeight;
       <%% end %>
     }
-  }
+  };
 
-  // Scroll chat messages to bottom or to target message
-  // Support both regular page loads and Turbo navigation
-  document.addEventListener('DOMContentLoaded', scrollChatMessages);
-  document.addEventListener('turbo:load', scrollChatMessages);
+  document.addEventListener('DOMContentLoaded', window._scrollChatMessages);
+  document.addEventListener('turbo:load', window._scrollChatMessages);
 </script>

data/lib/generators/llm_meta_client/scaffold/templates/app/views/chats/new.html.erb

@@ -30,7 +30,7 @@
             id: "message-input",
             data: { "chats-form-target": "prompt", action: "input->chats-form#updateSubmitButton" } %>
       </div>
-      <%%= f.hidden_field :branch_from_uuid, value: params.dig(:chat, :branch_from_uuid) %>
+      <%%= f.hidden_field :branch_from_uuid, value: @branch_from_uuid || params.dig(:chat, :branch_from_uuid) %>
       <div class="button-wrapper">
         <%%= f.button type: "submit",
             class: "send-button",

data/lib/generators/llm_meta_client/scaffold/templates/app/views/chats/update.turbo_stream.erb

@@ -26,13 +26,12 @@
 <%% # Update history sidebar - replace entire content to ensure update %>
 <%% if @prompt_execution %>
   <%%= turbo_stream.replace "history-sidebar" do %>
-
-    <div id="history-content">
+    <div id="history-sidebar">
       <h2>History</h2>
       <div class="history-stack" id="history-stack" data-controller="history">
         <%%= render 'prompt_navigator/history_card', locals: {
           ann: @prompt_execution,
-          next_ann: nil,
+          next_ann: (@chat&.ordered_by_descending_prompt_executions || [])[1],
           is_active: @prompt_execution.execution_id == @active_message_uuid,
           card_path: ->(uuid) { prompt_path(uuid) }
         } %>
@@ -54,25 +53,33 @@
 <turbo-stream action="after" target="messages-list">
   <template>
     <script>
-      // Clear and refocus message input
-      const messageInput = document.getElementById('message-input');
-      if (messageInput) {
-        messageInput.value = '';
-        messageInput.focus();
-      }
+      (function() {
+        // Clear and refocus message input
+        const messageInput = document.getElementById('message-input');
+        if (messageInput) {
+          messageInput.value = '';
+          messageInput.focus();
+        }
+
+        // Update submit button state
+        const form = document.querySelector('[data-controller="chats-form"]');
+        if (form && messageInput) {
+          const event = new Event('input', { bubbles: true });
+          messageInput.dispatchEvent(event);
+        }
 
-      // Update submit button state
-      const form = document.querySelector('[data-controller="chats-form"]');
-      if (form && messageInput) {
-        const event = new Event('input', { bubbles: true });
-        messageInput.dispatchEvent(event);
-      }
+        // Update branch_from_uuid to the latest prompt execution
+        const branchField = document.getElementById('branch_from_uuid');
+        if (branchField) {
+          branchField.value = '<%%= @prompt_execution&.execution_id %>';
+        }
 
-      // Scroll to bottom
-      const chatMessages = document.getElementById('chat-messages');
-      if (chatMessages) {
-        chatMessages.scrollTop = chatMessages.scrollHeight;
-      }
+        // Scroll to bottom
+        const chatMessages = document.getElementById('chat-messages');
+        if (chatMessages) {
+          chatMessages.scrollTop = chatMessages.scrollHeight;
+        }
+      })();
     </script>
   </template>
 </turbo-stream>

data/lib/generators/llm_meta_client/scaffold/templates/app/views/shared/_generation_settings_field.html.erb

@@ -13,75 +13,16 @@
     </button>
   </div>
   <div class="generation-settings-panel" data-<%%= stimulus_controller %>-target="panel" style="display: none;">
-    <div class="generation-setting-item">
-      <label for="temperature">
-        Temperature
-        <span class="setting-value" data-<%%= stimulus_controller %>-target="temperatureValue">0.7</span>
-      </label>
-      <input type="range" name="temperature" id="temperature"
-             min="0" max="2" step="0.1" value="0.7"
-             data-<%%= stimulus_controller %>-target="temperatureRange"
-             data-action="input-><%%= stimulus_controller %>#updateTemperature">
-      <div class="setting-range-labels">
-        <span>0 (deterministic)</span>
-        <span>2 (creative)</span>
-      </div>
-    </div>
-
-    <div class="generation-setting-item">
-      <label for="top_k">
-        Top-K
-        <span class="setting-value" data-<%%= stimulus_controller %>-target="topKValue">40</span>
-      </label>
-      <input type="range" name="top_k" id="top_k"
-             min="1" max="100" step="1" value="40"
-             data-<%%= stimulus_controller %>-target="topKRange"
-             data-action="input-><%%= stimulus_controller %>#updateTopK">
-      <div class="setting-range-labels">
-        <span>1 (focused)</span>
-        <span>100 (diverse)</span>
-      </div>
-    </div>
-
-    <div class="generation-setting-item">
-      <label for="top_p">
-        Top-P
-        <span class="setting-value" data-<%%= stimulus_controller %>-target="topPValue">0.9</span>
-      </label>
-      <input type="range" name="top_p" id="top_p"
-             min="0" max="1" step="0.05" value="0.9"
-             data-<%%= stimulus_controller %>-target="topPRange"
-             data-action="input-><%%= stimulus_controller %>#updateTopP">
-      <div class="setting-range-labels">
-        <span>0 (narrow)</span>
-        <span>1 (broad)</span>
-      </div>
-    </div>
-
-    <div class="generation-setting-item">
-      <label for="max_tokens">
-        Max Tokens
-      </label>
-      <input type="number" name="max_tokens" id="max_tokens"
-             min="1" max="128000" step="1" value=""
-             placeholder="Default (model-dependent)"
-             class="max-tokens-input"
-             data-<%%= stimulus_controller %>-target="maxTokensInput">
-    </div>
-
-    <div class="generation-setting-item">
-      <label for="repeat_penalty">
-        Repeat Penalty
-        <span class="setting-value" data-<%%= stimulus_controller %>-target="repeatPenaltyValue">1.1</span>
-      </label>
-      <input type="range" name="repeat_penalty" id="repeat_penalty"
-             min="1" max="2" step="0.05" value="1.1"
-             data-<%%= stimulus_controller %>-target="repeatPenaltyRange"
-             data-action="input-><%%= stimulus_controller %>#updateRepeatPenalty">
-      <div class="setting-range-labels">
-        <span>1.0 (no penalty)</span>
-        <span>2.0 (strong)</span>
-      </div>
+    <label for="generation_settings_json" class="generation-settings-label">
+      JSON format
+    </label>
+    <textarea name="generation_settings_json" id="generation_settings_json"
+              class="generation-settings-json-input"
+              rows="8"
+              placeholder='{"temperature": 0.7, "top_k": 40, "top_p": 0.9, "max_tokens": 4096, "repeat_penalty": 1.1}'
+              data-<%%= stimulus_controller %>-target="jsonInput"></textarea>
+    <div class="generation-settings-hint">
+      Available keys: temperature, top_k, top_p, max_tokens, repeat_penalty
     </div>
   </div>
 </div>

data/lib/generators/llm_meta_client/scaffold/templates/db/migrate/create_chats.rb

@@ -4,8 +4,6 @@ class CreateChats < ActiveRecord::Migration[8.1]
       t.references :user, null: true, foreign_key: true
       t.string :uuid, null: false
       t.string :title
-      t.string :llm_uuid
-      t.string :model
 
       t.timestamps
     end

data/lib/generators/llm_meta_client/scaffold/templates/db/migrate/migrate_llm_uuid_to_prompt_executions.rb

@@ -0,0 +1,7 @@
+class MigrateLlmUuidToPromptExecutions < ActiveRecord::Migration[8.1]
+  def change
+    add_column :prompt_navigator_prompt_executions, :llm_uuid, :string unless column_exists?(:prompt_navigator_prompt_executions, :llm_uuid)
+    remove_column :chats, :llm_uuid, :string if column_exists?(:chats, :llm_uuid)
+    remove_column :chats, :model, :string if column_exists?(:chats, :model)
+  end
+end

data/lib/llm_meta_client/server_resource.rb

@@ -117,7 +117,7 @@ module LlmMetaClient
       def ollama_options
         ollama_list = llms.filter { it["family"] == "ollama" }
         raise LlmMetaClient::Exceptions::OllamaUnavailableError if ollama_list.empty?
-        ollama_list
+        ollama_list.each { it["llm_type"] ||= "ollama" }
       end
 
       # Builds normalized option hashes from an array of prompts by slicing common keys

data/lib/llm_meta_client/version.rb

@@ -1,3 +1,3 @@
 module LlmMetaClient
-  VERSION = "0.6.1"
+  VERSION = "1.0.1"
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: llm_meta_client
 version: !ruby/object:Gem::Version
-  version: 0.6.1
+  version: 1.0.1
 platform: ruby
 authors:
 - dhq_boiler
@@ -49,28 +49,28 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.2'
+        version: '1.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.2'
+        version: '1.0'
 - !ruby/object:Gem::Dependency
   name: chat_manager
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.2'
+        version: '1.0'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '0.2'
+        version: '1.0'
 description: llm_meta_client provides a Rails Engine with scaffold and authentication
   generators for building LLM-powered chat applications. Supports OpenAI, Anthropic,
   Google, and Ollama providers.
@@ -130,6 +130,7 @@ files:
 - lib/generators/llm_meta_client/scaffold/templates/config/initializers/llm_service.rb
 - lib/generators/llm_meta_client/scaffold/templates/db/migrate/create_chats.rb
 - lib/generators/llm_meta_client/scaffold/templates/db/migrate/create_messages.rb
+- lib/generators/llm_meta_client/scaffold/templates/db/migrate/migrate_llm_uuid_to_prompt_executions.rb
 - lib/llm_meta_client.rb
 - lib/llm_meta_client/chat_manageable.rb
 - lib/llm_meta_client/engine.rb