Skip to content

Latest commit

 

History

History
166 lines (128 loc) · 6.04 KB

services.md

File metadata and controls

166 lines (128 loc) · 6.04 KB

Services

The services provide the core business logic for the application. They are implemented using the singleton pattern for global access.

ChatService

Path: src/services/chat-service.ts

Description: Manages chat conversations and message handling.

Key Methods:

  • getInstance(): Returns the singleton instance
  • initialize(): Initializes the service
  • loadConversations(): Loads all conversations from the database
  • loadConversation(conversationId): Loads a specific conversation
  • createConversation(title): Creates a new conversation
  • sendMessage(content, conversationId, isStreaming, conversationUpdate): Sends a message with streaming support
  • editMessage(messageId, conversationId, newContent, isStreaming, conversationUpdate): Edits a message
  • regenerateAiMessage(messageId, conversationId, isStreaming, conversationUpdate): Regenerates an AI response
  • stopStreaming(conversationId): Stops the current streaming message
  • isCurrentlyStreaming(conversationId): Checks if streaming is active
  • getCurrentProviderModelCapabilities(): Gets capabilities of current provider/model combination

Usage Example:

const chatService = ChatService.getInstance();
await chatService.initialize();
const conversations = await chatService.loadConversations();

AIService

Path: src/services/ai-service.ts

Description: Provides integration with AI providers.

Key Methods:

  • getInstance(): Returns the singleton instance
  • getProvider(name): Gets a provider by name
  • getAllProviders(): Gets all registered providers
  • getChatCompletion(messages, options, streamController): Gets a streaming chat completion
  • generateImage(prompt, options): Generates an image from a prompt
  • getCachedAllModels(): Gets all models from cache
  • getModelsForProvider(providerName): Gets models for a specific provider
  • refreshModels(): Refreshes the model cache

Usage Example:

const aiService = AIService.getInstance();
const provider = aiService.getProvider('openai');
const models = await aiService.getModelsForProvider('openai');

SettingsService

Path: src/services/settings-service.ts

Description: Manages application settings and user preferences.

Key Methods:

  • getInstance(): Returns the singleton instance
  • getSettings(): Gets all settings
  • saveSettings(settings): Saves settings
  • getSelectedProvider(): Gets the currently selected provider
  • setSelectedProvider(provider): Sets the selected provider
  • getSelectedModel(): Gets the currently selected model
  • setSelectedModel(model): Sets the selected model
  • getProviderSettings(provider): Gets settings for a specific provider
  • setProviderApiKey(provider, apiKey): Sets the API key for a provider
  • getWebSearchActive(): Gets the web search status
  • setWebSearchEnabled(enabled): Enables/disables web search

Usage Example:

const settingsService = SettingsService.getInstance();
const selectedProvider = settingsService.getSelectedProvider();
await settingsService.setSelectedModel('gpt-4');

DatabaseIntegrationService

Path: src/services/database-integration.ts

Description: Provides database integration for persistent storage.

Key Methods:

  • getInstance(): Returns the singleton instance
  • initialize(): Initializes the database connection
  • loadConversationsList(): Loads the list of all conversations
  • loadConversation(conversationId): Loads a specific conversation with messages
  • createConversation(title): Creates a new conversation
  • updateConversation(conversation): Updates a conversation
  • deleteConversation(conversationId): Deletes a conversation
  • updateChatMessage(messageId, message, conversationId): Updates a message
  • loadFoldersList(): Loads all conversation folders
  • createFolder(folder): Creates a new folder
  • updateFolder(folderId, folder): Updates a folder
  • deleteFolder(folderId): Deletes a folder

Usage Example:

const dbService = DatabaseIntegrationService.getInstance();
await dbService.initialize();
const conversations = await dbService.loadConversationsList();

MessageHelper

Path: src/services/message-helper.ts

Description: Provides utility functions for message handling.

Key Methods:

  • mapMessagesTreeToList(conversation, includeSystemMessages): Maps message tree to flat list
  • addUserMessageToConversation(content, conversation): Adds a user message to conversation
  • insertAssistantMessageToConversation(userMessage, aiResponse, conversation): Inserts AI response
  • getPlaceholderMessage(model, provider, conversationId): Creates a placeholder message for streaming

Usage Example:

const messages = MessageHelper.mapMessagesTreeToList(conversation, false);
const { conversation: updated, message: userMsg } = await MessageHelper.addUserMessageToConversation(content, conversation);

StreamControlHandler

Path: src/services/streaming-control.ts

Description: Handles streaming control for AI responses.

Key Methods:

  • constructor(conversation, placeholderMessage, onChunk, onFinish): Initializes a handler
  • getAbortSignal(): Gets the abort signal for cancellation
  • abortStreaming(): Aborts the current streaming
  • handleChunk(chunk): Handles a streaming chunk
  • finalizeStreaming(): Finalizes the streaming process

Usage Example:

const streamController = new StreamControlHandler(
  conversation,
  placeholderMessage,
  (updated) => { /* handle chunk */ },
  (aiResponse) => { /* handle final response */ }
);

Provider System

Path: src/services/providers/

Description: Provider implementations for different AI services.

The provider system uses a factory pattern to create provider instances:

ProviderFactory:

  • getNewProvider(providerName): Creates a new provider instance

Provider Interface:

  • getChatCompletion(messages, options, streamController): Gets chat completion
  • getAvailableModels(): Gets available models for the provider

Implemented Providers:

  • OpenAI
  • Anthropic
  • Custom providers