diff --git a/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts b/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts index 152fce03aa..33b881e2cc 100644 --- a/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts +++ b/src/renderer/src/aiCore/chunk/handleTooCallChunk.ts @@ -8,6 +8,7 @@ import { ToolCallUnion, ToolResultUnion, ToolSet } from '@cherrystudio/ai-core' import Logger from '@renderer/config/logger' import { BaseTool, MCPToolResponse, ToolCallResponse } from '@renderer/types' import { Chunk, ChunkType } from '@renderer/types/chunk' +import { type ProviderMetadata } from 'ai' // import type { // AnthropicSearchOutput, // WebSearchPluginConfig @@ -40,7 +41,27 @@ export class ToolCallChunkHandler { // this.onChunk = callback // } - handleToolCallCreated(chunk: { type: 'tool-input-start' | 'tool-input-delta' | 'tool-input-end' }): void { + handleToolCallCreated( + chunk: + | { + type: 'tool-input-start' + id: string + toolName: string + providerMetadata?: ProviderMetadata + providerExecuted?: boolean + } + | { + type: 'tool-input-end' + id: string + providerMetadata?: ProviderMetadata + } + | { + type: 'tool-input-delta' + id: string + delta: string + providerMetadata?: ProviderMetadata + } + ): void { switch (chunk.type) { case 'tool-input-start': { // 能拿到说明是mcpTool diff --git a/src/renderer/src/aiCore/plugins/smoothReasoningPlugin.ts b/src/renderer/src/aiCore/plugins/smoothReasoningPlugin.ts deleted file mode 100644 index 62ed5e9e7e..0000000000 --- a/src/renderer/src/aiCore/plugins/smoothReasoningPlugin.ts +++ /dev/null @@ -1,152 +0,0 @@ -// // 可能会废弃，在流上做delay还是有问题 - -// import { definePlugin } from '@cherrystudio/ai-core' - -// const chunkingRegex = /([\u4E00-\u9FFF])|\S+\s+/ -// const delayInMs = 50 - -// export default definePlugin({ -// name: 'reasoningPlugin', - -// transformStream: () => () => { -// // === smoothing 状态 === -// let buffer = '' - -// // === 时间跟踪状态 === -// let thinkingStartTime = performance.now() -// let hasStartedThinking = false -// let accumulatedThinkingContent = '' - -// // === 日志计数器 === -// let chunkCount = 0 -// let delayCount = 0 - -// const delay = (ms: number) => new Promise((resolve) => setTimeout(resolve, ms)) - -// // 收集所有当前可匹配的chunks -// const collectMatches = (inputBuffer: string) => { -// const matches: string[] = [] -// let tempBuffer = inputBuffer -// let match - -// // 重置regex状态 -// chunkingRegex.lastIndex = 0 - -// while ((match = chunkingRegex.exec(tempBuffer)) !== null) { -// matches.push(match[0]) -// tempBuffer = tempBuffer.slice(match.index + match[0].length) -// // 重置regex以从头开始匹配剩余内容 -// chunkingRegex.lastIndex = 0 -// } - -// return { -// matches, -// remaining: tempBuffer -// } -// } - -// return new TransformStream({ -// async transform(chunk, controller) { -// if (chunk.type !== 'reasoning') { -// // === 处理 reasoning 结束 === -// if (hasStartedThinking && accumulatedThinkingContent) { -// console.log( -// `[ReasoningPlugin] Ending reasoning. Final stats: chunks=${chunkCount}, delays=${delayCount}, efficiency=${(chunkCount / Math.max(delayCount, 1)).toFixed(2)}x` -// ) - -// // 先输出剩余的 buffer -// if (buffer.length > 0) { -// console.log(`[ReasoningPlugin] Flushing remaining buffer: "${buffer}"`) -// controller.enqueue({ -// type: 'reasoning', -// textDelta: buffer, -// thinking_millsec: performance.now() - thinkingStartTime -// }) -// buffer = '' -// } - -// // 生成 reasoning-signature -// controller.enqueue({ -// type: 'reasoning-signature', -// text: accumulatedThinkingContent, -// thinking_millsec: performance.now() - thinkingStartTime -// }) - -// // 重置状态 -// accumulatedThinkingContent = '' -// hasStartedThinking = false -// thinkingStartTime = 0 -// chunkCount = 0 -// delayCount = 0 -// } - -// controller.enqueue(chunk) -// return -// } - -// // === 处理 reasoning 类型 === - -// // 1. 时间跟踪逻辑 -// if (!hasStartedThinking) { -// hasStartedThinking = true -// thinkingStartTime = performance.now() -// console.log(`[ReasoningPlugin] Starting reasoning session`) -// } -// accumulatedThinkingContent += chunk.textDelta - -// // 2. 动态Smooth处理逻辑 -// const beforeBuffer = buffer -// buffer += chunk.textDelta - -// console.log(`[ReasoningPlugin] Received chunk: "${chunk.textDelta}", buffer: "${beforeBuffer}" → "${buffer}"`) - -// // 收集所有当前可以匹配的chunks -// const { matches, remaining } = collectMatches(buffer) - -// if (matches.length > 0) { -// console.log( -// `[ReasoningPlugin] Collected ${matches.length} matches: [${matches.map((m) => `"${m}"`).join(', ')}], remaining: "${remaining}"` -// ) - -// // 批量输出所有匹配的chunks -// for (const matchText of matches) { -// controller.enqueue({ -// type: 'reasoning', -// textDelta: matchText, -// thinking_millsec: performance.now() - thinkingStartTime -// }) -// chunkCount++ -// } - -// // 更新buffer为剩余内容 -// buffer = remaining - -// // 只等待一次，而不是每个chunk都等待 -// delayCount++ -// console.log( -// `[ReasoningPlugin] Delaying ${delayInMs}ms (delay #${delayCount}, efficiency: ${(chunkCount / delayCount).toFixed(2)} chunks/delay)` -// ) -// const delayStart = performance.now() -// await delay(delayInMs) -// const actualDelay = performance.now() - delayStart -// console.log(`[ReasoningPlugin] Delay completed: expected=${delayInMs}ms, actual=${actualDelay.toFixed(1)}ms`) -// } else { -// console.log(`[ReasoningPlugin] No matches found, keeping in buffer: "${buffer}"`) -// } -// // 如果没有匹配，保留在buffer中等待下次数据 -// }, - -// // === flush 处理剩余 buffer === -// flush(controller) { -// if (buffer.length > 0) { -// console.log(`[ReasoningPlugin] Final flush: "${buffer}"`) -// controller.enqueue({ -// type: 'reasoning', -// textDelta: buffer, -// thinking_millsec: hasStartedThinking ? performance.now() - thinkingStartTime : 0 -// }) -// } -// } -// }) -// } -// }) diff --git a/src/renderer/src/aiCore/plugins/textPlugin.ts b/src/renderer/src/aiCore/plugins/textPlugin.ts deleted file mode 100644 index b35156eae8..0000000000 --- a/src/renderer/src/aiCore/plugins/textPlugin.ts +++ /dev/null @@ -1,13 +0,0 @@ -// // 可能会废弃，在流上做delay还是有问题 - -// import { definePlugin, smoothStream } from '@cherrystudio/ai-core' - -// export default definePlugin({ -// name: 'textPlugin', -// transformStream: () => -// smoothStream({ -// delayInMs: 50, -// // 中文3个字符一个chunk,英文一个单词一个chunk -// chunking: /([\u4E00-\u9FFF]{3})|\S+\s+/ -// }) -// }) diff --git a/src/renderer/src/aiCore/tools/MemorySearchTool.ts b/src/renderer/src/aiCore/tools/MemorySearchTool.ts new file mode 100644 index 0000000000..cd62f7de8b --- /dev/null +++ b/src/renderer/src/aiCore/tools/MemorySearchTool.ts @@ -0,0 +1,140 @@ +import store from '@renderer/store' +import { selectCurrentUserId, selectGlobalMemoryEnabled, selectMemoryConfig } from '@renderer/store/memory' +import type { Assistant } from '@renderer/types' +import { type InferToolOutput, tool } from 'ai' +import { z } from 'zod' + +import { MemoryProcessor } from '../../services/MemoryProcessor' + +/** + * 🧠 基础记忆搜索工具 + * AI 可以主动调用的简单记忆搜索 + */ +export const memorySearchTool = () => { + return tool({ + name: 'builtin_memory_search', + description: 'Search through conversation memories and stored facts for relevant context', + inputSchema: z.object({ + query: z.string().describe('Search query to find relevant memories'), + limit: z.number().min(1).max(20).default(5).describe('Maximum number of memories to return') + }), + execute: async ({ query, limit = 5 }) => { + console.log('🧠 [memorySearchTool] Searching memories:', { query, limit }) + + try { + const globalMemoryEnabled = selectGlobalMemoryEnabled(store.getState()) + if (!globalMemoryEnabled) { + return [] + } + + const memoryConfig = selectMemoryConfig(store.getState()) + if (!memoryConfig.llmApiClient || !memoryConfig.embedderApiClient) { + console.warn('Memory search skipped: embedding or LLM model not configured') + return [] + } + + const currentUserId = selectCurrentUserId(store.getState()) + const processorConfig = MemoryProcessor.getProcessorConfig(memoryConfig, 'default', currentUserId) + + const memoryProcessor = new MemoryProcessor() + const relevantMemories = await memoryProcessor.searchRelevantMemories(query, processorConfig, limit) + + if (relevantMemories?.length > 0) { + console.log('🧠 [memorySearchTool] Found memories:', relevantMemories.length) + return relevantMemories + } + return [] + } catch (error) { + console.error('🧠 [memorySearchTool] Error:', error) + return [] + } + } + }) +} + +/** + * 🧠 智能记忆搜索工具（带上下文提取） + * 从用户消息和对话历史中自动提取关键词进行记忆搜索 + */ +export const memorySearchToolWithExtraction = (assistant: Assistant) => { + return tool({ + name: 'memory_search_with_extraction', + description: 'Search memories with automatic keyword extraction from conversation context', + inputSchema: z.object({ + userMessage: z.object({ + content: z.string().describe('The main content of the user message'), + role: z.enum(['user', 'assistant', 'system']).describe('Message role') + }), + lastAnswer: z + .object({ + content: z.string().describe('The main content of the last assistant response'), + role: z.enum(['user', 'assistant', 'system']).describe('Message role') + }) + .optional() + }), + execute: async ({ userMessage, lastAnswer }) => { + console.log('🧠 [memorySearchToolWithExtraction] Processing:', { userMessage, lastAnswer }) + + try { + const globalMemoryEnabled = selectGlobalMemoryEnabled(store.getState()) + if (!globalMemoryEnabled || !assistant.enableMemory) { + return { + extractedKeywords: 'Memory search disabled', + searchResults: [] + } + } + + const memoryConfig = selectMemoryConfig(store.getState()) + if (!memoryConfig.llmApiClient || !memoryConfig.embedderApiClient) { + console.warn('Memory search skipped: embedding or LLM model not configured') + return { + extractedKeywords: 'Memory models not configured', + searchResults: [] + } + } + + // 🔍 使用用户消息内容作为搜索关键词 + const content = userMessage.content + + if (!content) { + return { + extractedKeywords: 'No content to search', + searchResults: [] + } + } + + const currentUserId = selectCurrentUserId(store.getState()) + const processorConfig = MemoryProcessor.getProcessorConfig(memoryConfig, assistant.id, currentUserId) + + const memoryProcessor = new MemoryProcessor() + const relevantMemories = await memoryProcessor.searchRelevantMemories( + content, + processorConfig, + 5 // Limit to top 5 most relevant memories + ) + + if (relevantMemories?.length > 0) { + console.log('🧠 [memorySearchToolWithExtraction] Found memories:', relevantMemories.length) + return { + extractedKeywords: content, + searchResults: relevantMemories + } + } + + return { + extractedKeywords: content, + searchResults: [] + } + } catch (error) { + console.error('🧠 [memorySearchToolWithExtraction] Error:', error) + return { + extractedKeywords: 'Search failed', + searchResults: [] + } + } + } + }) +} + +export type MemorySearchToolOutput = InferToolOutput> +export type MemorySearchToolWithExtractionOutput = InferToolOutput> diff --git a/src/renderer/src/types/tool.ts b/src/renderer/src/types/tool.ts index 52a58063bd..96522ba8e8 100644 --- a/src/renderer/src/types/tool.ts +++ b/src/renderer/src/types/tool.ts @@ -1,6 +1,3 @@ -import { WebSearchToolOutputSchema } from '@cherrystudio/ai-core/built-in/plugins' -import type { WebSearchToolOutput } from '@renderer/aiCore/tools/WebSearchTool' - import type { MCPToolInputSchema } from './index' export type ToolType = 'builtin' | 'provider' | 'mcp' @@ -12,15 +9,15 @@ export interface BaseTool { type: ToolType } -export interface ToolCallResponse { - id: string - toolName: string - arguments: Record | undefined - status: 'invoking' | 'completed' | 'error' - result?: any // AI SDK的工具执行结果 - error?: string - providerExecuted?: boolean // 标识是Provider端执行还是客户端执行 -} +// export interface ToolCallResponse { +// id: string +// toolName: string +// arguments: Record | undefined +// status: 'invoking' | 'completed' | 'error' +// result?: any // AI SDK的工具执行结果 +// error?: string +// providerExecuted?: boolean // 标识是Provider端执行还是客户端执行 +// } export interface BuiltinTool extends BaseTool { inputSchema: MCPToolInputSchema @@ -33,5 +30,3 @@ export interface MCPTool extends BaseTool { inputSchema: MCPToolInputSchema type: 'mcp' } - -export type WebSearchToolOutputSchema = WebSearchToolOutput | WebSearchToolOutputSchema