/**
* 工具注册表 - 统一管理所有 MCP 工具的元数据和处理函数
*/
import { searchArxivPapers, downloadArxivPdfTool } from './arxiv-tools.js';
import {
parsePdfToText,
parsePdfToMarkdown,
convertToWechatArticle,
convertToAcademicReviewEnhanced,
processArxivPaper
} from './processing-tools.js';
import {
batchDownloadPdfsTool,
batchAnalyzePapersTool,
generateUnifiedLiteratureReview
} from './batch-tools.js';
import { searchAcademicPapers } from './search-tools.js';
import {
exportToNotionFullTool,
exportToNotionUpdateTool,
exportIndividualReviewToMd,
batchExportIndividualReviews
} from './export-tools.js';
import { clearWorkdir } from './utility-tools.js';
/**
* 工具定义接口
*/
export interface ToolDefinition {
name: string;
description: string;
inputSchema: any;
handler: (args: any) => Promise<any>;
}
/**
* 所有工具的注册表
*/
export const TOOL_REGISTRY: ToolDefinition[] = [
// arXiv 工具
{
name: 'search_arxiv_papers',
description: '搜索 arXiv 论文',
inputSchema: {
type: 'object',
properties: {
query: {
type: 'string',
description: '搜索英文关键词'
},
max_results: {
type: 'number',
description: '最大结果数量',
default: 5
}
},
required: ['query']
},
handler: async (args) => await searchArxivPapers(args.query, args.max_results || 5)
},
{
name: 'download_arxiv_pdf',
description: '下载 arXiv PDF 文件',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文URL（如：http://arxiv.org/abs/2403.15137v1）或 arXiv ID（如：2403.15137v1），优先使用URL'
}
},
required: ['arxiv_id']
},
handler: async (args) => await downloadArxivPdfTool(args.arxiv_id)
},
// 论文处理工具
{
name: 'parse_pdf_to_text',
description: '解析 PDF 并返回原始文本内容',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
},
paperInfo: {
type: 'object',
description: '论文信息（可选，用于添加论文元数据）',
properties: {
title: { type: 'string' },
summary: { type: 'string' },
published: { type: 'string' },
authors: { type: 'array' }
}
}
},
required: ['arxiv_id']
},
handler: async (args) => await parsePdfToText(args.arxiv_id, args.paperInfo)
},
{
name: 'parse_pdf_to_markdown',
description: '解析 PDF 并返回 LLM 翻译后的中文 Markdown 文件',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
},
paperInfo: {
type: 'object',
description: '论文信息（可选，用于添加论文元数据）',
properties: {
title: { type: 'string' },
summary: { type: 'string' },
published: { type: 'string' },
authors: { type: 'array' }
}
}
},
required: ['arxiv_id']
},
handler: async (args) => await parsePdfToMarkdown(args.arxiv_id, args.paperInfo)
},
{
name: 'convert_to_wechat_article',
description: '转换为微信文章格式',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
}
},
required: ['arxiv_id']
},
handler: async (args) => await convertToWechatArticle(args.arxiv_id)
},
{
name: 'convert_to_academic_review_enhanced',
description: '转换为增强版学术文献综述（包含博士思维分析和Mermaid可视化）',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
}
},
required: ['arxiv_id']
},
handler: async (args) => await convertToAcademicReviewEnhanced(args.arxiv_id)
},
{
name: 'process_arxiv_paper',
description: '完整流程处理 arXiv 论文（搜索、下载、解析、转换）',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
},
include_wechat: {
type: 'boolean',
description: '是否生成微信文章',
default: true
}
},
required: ['arxiv_id']
},
handler: async (args) => await processArxivPaper(args.arxiv_id, args.include_wechat)
},
// 批量处理工具
{
name: 'batch_download_pdfs',
description: '批量并发下载多篇论文的 PDF 文件',
inputSchema: {
type: 'object',
properties: {
arxiv_ids: {
type: 'array',
items: { type: 'string' },
description: 'arXiv ID 列表'
},
maxConcurrent: {
type: 'number',
description: '最大并发数',
default: 5
},
maxRetries: {
type: 'number',
description: '最大重试次数',
default: 3
}
},
required: ['arxiv_ids']
},
handler: async (args) => await batchDownloadPdfsTool(args.arxiv_ids, args.maxConcurrent, args.maxRetries)
},
{
name: 'batch_analyze_papers',
description: '批量并发分析多篇论文，生成单篇深度分析并保存到数据库',
inputSchema: {
type: 'object',
properties: {
arxiv_ids: {
type: 'array',
items: { type: 'string' },
description: 'arXiv ID 列表'
},
include_wechat: {
type: 'boolean',
description: '是否生成微信文章',
default: true
},
maxConcurrent: {
type: 'number',
description: '最大并发数',
default: 3
},
temperature: {
type: 'number',
description: 'AI 温度参数（0-1）',
default: 0.3
},
skipExisting: {
type: 'boolean',
description: '跳过已有分析的论文',
default: true
}
},
required: ['arxiv_ids']
},
handler: async (args) => await batchAnalyzePapersTool(
args.arxiv_ids,
args.maxConcurrent,
args.temperature,
args.skipExisting
)
},
{
name: 'generate_unified_literature_review',
description: '基于多篇论文的单篇分析，生成跨论文的综合文献综述',
inputSchema: {
type: 'object',
properties: {
arxiv_ids: {
type: 'array',
items: { type: 'string' },
description: 'arXiv ID 列表'
},
topic: {
type: 'string',
description: '综述主题'
},
temperature: {
type: 'number',
description: 'AI 温度参数（0-1）',
default: 0.4
},
focusArea: {
type: 'string',
description: '研究焦点领域'
}
},
required: ['arxiv_ids']
},
handler: async (args) => await generateUnifiedLiteratureReview(
args.arxiv_ids,
args.temperature,
args.focusArea
)
},
// 多源搜索工具
{
name: 'search_academic_papers',
description: '跨多个学术数据源搜索论文（DBLP、OpenReview、Papers With Code），返回 Notion 友好格式',
inputSchema: {
type: 'object',
properties: {
query: {
type: 'string',
description: '搜索关键词'
},
max_results: {
type: 'number',
description: '最大结果数量',
default: 10
},
sources: {
type: 'array',
items: {
type: 'string',
enum: ['dblp', 'openreview', 'paperswithcode']
},
description: '数据源列表',
default: ['dblp', 'openreview', 'paperswithcode']
},
export_to_notion: {
type: 'boolean',
description: '是否导出到 Notion',
default: false
},
minQualityScore: {
type: 'number',
description: '最低质量评分（0-100）',
default: 0
}
},
required: ['query']
},
handler: async (args) => await searchAcademicPapers(
args.query,
args.max_results,
args.sources,
args.minQualityScore
)
},
// 导出工具
{
name: 'export_to_notion_full',
description: '导出完整的 Notion 元数据（论文库 + 单篇综述 + 综合综述）',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
},
database_id: {
type: 'string',
description: 'Notion 数据库 ID'
}
},
required: ['arxiv_id', 'database_id']
},
handler: async (args) => await exportToNotionFullTool(args.arxiv_id, args.database_id)
},
{
name: 'export_to_notion_update',
description: '只导出增量内容的 Notion 元数据（需要提供已存在的论文列表）',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
},
page_id: {
type: 'string',
description: 'Notion 页面 ID'
}
},
required: ['arxiv_id', 'page_id']
},
handler: async (args) => await exportToNotionUpdateTool(
args.arxiv_id,
args.page_id
)
},
{
name: 'export_individual_review_to_md',
description: '将数据库中的单篇文献综述导出为 Markdown 文件',
inputSchema: {
type: 'object',
properties: {
arxiv_id: {
type: 'string',
description: 'arXiv 论文ID'
}
},
required: ['arxiv_id']
},
handler: async (args) => await exportIndividualReviewToMd(args.arxiv_id)
},
{
name: 'batch_export_individual_reviews',
description: '批量导出所有有单篇综述的论文为 Markdown 文件',
inputSchema: {
type: 'object',
properties: {},
required: []
},
handler: async () => await batchExportIndividualReviews()
},
// 实用工具
{
name: 'clear_workdir',
description: '清空工作区所有文件（危险操作）',
inputSchema: {
type: 'object',
properties: {},
required: []
},
handler: async () => await clearWorkdir()
}
];
/**
* 获取所有工具的元数据（用于 ListToolsRequestSchema）
*/
export function getAllToolsMetadata() {
return TOOL_REGISTRY.map(tool => ({
name: tool.name,
description: tool.description,
inputSchema: tool.inputSchema
}));
}
/**
* 根据工具名称获取处理函数
*/
export function getToolHandler(toolName: string): ((args: any) => Promise<any>) | undefined {
const tool = TOOL_REGISTRY.find(t => t.name === toolName);
return tool?.handler;
}
