deepfish-ai 1.0.29 → 2.0.3

package/src/AgentRobot/BaseAgentRobot/lazy-tools/pdf.js

@@ -1,736 +0,0 @@
-/**
- * @Author: Roman 306863030@qq.com
- * @Description: PDF 文档处理工具集（pdf-parse / pdf-lib / pdfkit）
- */
-const path = require('path')
-const fs = require('fs-extra')
-const pdfParse = require('pdf-parse')
-const { PDFDocument, degrees, rgb, StandardFonts } = require('pdf-lib')
-const PDFKit = require('pdfkit')
-const { pdf } = require('pdf-to-img')
-const sharp = require('sharp')
-
-// ─── 统一返回结构 ─────────────────────────────────────────────────────────────
-
-function ok(data = null) {
-  return { success: true, data }
-}
-
-function fail(error, data = null) {
-  return { success: false, error: error?.message || String(error), data }
-}
-
-function resolvePath(filePath) {
-  return path.resolve(process.cwd(), filePath)
-}
-
-// ─── 工具函数 ─────────────────────────────────────────────────────────────────
-
-/**
- * 读取 PDF 中的全部文本内容
- */
-async function readPdfText(filePath) {
-  try {
-    const fullPath = resolvePath(filePath)
-    if (!fs.existsSync(fullPath)) return fail(`File does not exist: ${fullPath}`, { filePath: fullPath })
-    const dataBuffer = fs.readFileSync(fullPath)
-    const data = await pdfParse(dataBuffer)
-    return ok({
-      filePath: fullPath,
-      text: data.text,
-      pageCount: data.numpages,
-      wordCount: data.text.trim().split(/\s+/).filter(Boolean).length,
-      charCount: data.text.length,
-    })
-  } catch (error) {
-    return fail(error, { filePath })
-  }
-}
-
-/**
- * 获取 PDF 文档元信息（页数、作者、标题、创建时间等）
- */
-async function getPdfInfo(filePath) {
-  try {
-    const fullPath = resolvePath(filePath)
-    if (!fs.existsSync(fullPath)) return fail(`File does not exist: ${fullPath}`, { filePath: fullPath })
-    const dataBuffer = fs.readFileSync(fullPath)
-    const parseData = await pdfParse(dataBuffer)
-    const pdfDoc = await PDFDocument.load(dataBuffer)
-    const stat = fs.statSync(fullPath)
-    return ok({
-      filePath: fullPath,
-      pageCount: parseData.numpages,
-      size: stat.size,
-      mtime: stat.mtime,
-      title: pdfDoc.getTitle() || null,
-      author: pdfDoc.getAuthor() || null,
-      subject: pdfDoc.getSubject() || null,
-      keywords: pdfDoc.getKeywords() || null,
-      creator: pdfDoc.getCreator() || null,
-      producer: pdfDoc.getProducer() || null,
-      creationDate: pdfDoc.getCreationDate() || null,
-      modificationDate: pdfDoc.getModificationDate() || null,
-      pdfVersion: parseData.info?.PDFFormatVersion || null,
-    })
-  } catch (error) {
-    return fail(error, { filePath })
-  }
-}
-
-/**
- * 仅获取 PDF 总页数（轻量调用）
- */
-async function getPdfPageCount(filePath) {
-  try {
-    const fullPath = resolvePath(filePath)
-    if (!fs.existsSync(fullPath)) return fail(`File does not exist: ${fullPath}`, { filePath: fullPath })
-    const pdfDoc = await PDFDocument.load(fs.readFileSync(fullPath))
-    return ok({ filePath: fullPath, pageCount: pdfDoc.getPageCount() })
-  } catch (error) {
-    return fail(error, { filePath })
-  }
-}
-
-/**
- * 在 PDF 中搜索关键词，返回包含该词的行
- */
-async function searchPdfText(filePath, keyword) {
-  try {
-    const readResult = await readPdfText(filePath)
-    if (!readResult.success) return readResult
-    const lines = readResult.data.text.split('\n')
-    const matches = lines
-      .map((line, idx) => ({ line: idx + 1, text: line }))
-      .filter(({ text }) => text.includes(keyword))
-    return ok({
-      filePath: resolvePath(filePath),
-      keyword,
-      matchCount: matches.length,
-      matches,
-    })
-  } catch (error) {
-    return fail(error, { filePath, keyword })
-  }
-}
-
-/**
- * 合并多个 PDF 文件为一个
- */
-async function mergePdfs(inputPaths, outputPath) {
-  try {
-    const fullOutput = resolvePath(outputPath)
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const mergedDoc = await PDFDocument.create()
-    for (const inputPath of inputPaths) {
-      const fullInput = resolvePath(inputPath)
-      if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-      const srcDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-      const pages = await mergedDoc.copyPages(srcDoc, srcDoc.getPageIndices())
-      pages.forEach((p) => mergedDoc.addPage(p))
-    }
-    fs.writeFileSync(fullOutput, await mergedDoc.save())
-    return ok({ outputPath: fullOutput, pageCount: mergedDoc.getPageCount(), inputCount: inputPaths.length })
-  } catch (error) {
-    return fail(error, { outputPath })
-  }
-}
-
-/**
- * 将 PDF 按每页拆分为独立文件，保存到输出目录
- */
-async function splitPdf(filePath, outputDir) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutputDir = resolvePath(outputDir)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(fullOutputDir)
-    const srcDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    const pageCount = srcDoc.getPageCount()
-    const baseName = path.basename(fullInput, path.extname(fullInput))
-    const results = []
-    for (let i = 0; i < pageCount; i++) {
-      const newDoc = await PDFDocument.create()
-      const [page] = await newDoc.copyPages(srcDoc, [i])
-      newDoc.addPage(page)
-      const outFile = path.join(fullOutputDir, `${baseName}_page${i + 1}.pdf`)
-      fs.writeFileSync(outFile, await newDoc.save())
-      results.push(outFile)
-    }
-    return ok({ filePath: fullInput, outputDir: fullOutputDir, pageCount, files: results })
-  } catch (error) {
-    return fail(error, { filePath, outputDir })
-  }
-}
-
-/**
- * 从 PDF 中提取指定页码（1-based）生成新文件
- */
-async function extractPdfPages(filePath, outputPath, pages) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutput = resolvePath(outputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const srcDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    const total = srcDoc.getPageCount()
-    const indices = (pages || []).map((p) => p - 1).filter((i) => i >= 0 && i < total)
-    if (indices.length === 0) return fail('No valid page numbers provided', { filePath: fullInput, pages })
-    const newDoc = await PDFDocument.create()
-    const copied = await newDoc.copyPages(srcDoc, indices)
-    copied.forEach((p) => newDoc.addPage(p))
-    fs.writeFileSync(fullOutput, await newDoc.save())
-    return ok({ filePath: fullInput, outputPath: fullOutput, extractedPages: indices.map((i) => i + 1) })
-  } catch (error) {
-    return fail(error, { filePath, outputPath })
-  }
-}
-
-/**
- * 旋转 PDF 页面（可指定特定页或全部页）
- */
-async function rotatePdfPages(filePath, outputPath, rotateDegrees, pageNumbers) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutput = resolvePath(outputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const pdfDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    const deg = rotateDegrees || 90
-    const total = pdfDoc.getPageCount()
-    const targets = Array.isArray(pageNumbers) && pageNumbers.length > 0
-      ? pageNumbers.map((p) => p - 1).filter((i) => i >= 0 && i < total)
-      : Array.from({ length: total }, (_, i) => i)
-    for (const idx of targets) {
-      const page = pdfDoc.getPage(idx)
-      page.setRotation(degrees((page.getRotation().angle + deg) % 360))
-    }
-    fs.writeFileSync(fullOutput, await pdfDoc.save())
-    return ok({ filePath: fullInput, outputPath: fullOutput, degrees: deg, rotatedPages: targets.map((i) => i + 1) })
-  } catch (error) {
-    return fail(error, { filePath, outputPath })
-  }
-}
-
-/**
- * 向 PDF 全部页面添加文字水印
- */
-async function addWatermarkText(filePath, outputPath, text, options = {}) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutput = resolvePath(outputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const pdfDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    const font = await pdfDoc.embedFont(StandardFonts.Helvetica)
-    const fontSize = options.fontSize || 48
-    const opacity = options.opacity !== undefined ? options.opacity : 0.2
-    const color = options.color
-      ? rgb(options.color.r / 255, options.color.g / 255, options.color.b / 255)
-      : rgb(0.5, 0.5, 0.5)
-    const rotateAngle = options.angle !== undefined ? options.angle : 45
-    const pages = pdfDoc.getPages()
-    for (const page of pages) {
-      const { width, height } = page.getSize()
-      page.drawText(text, {
-        x: width / 2 - (font.widthOfTextAtSize(text, fontSize) / 2),
-        y: height / 2,
-        size: fontSize,
-        font,
-        color,
-        opacity,
-        rotate: degrees(rotateAngle),
-      })
-    }
-    fs.writeFileSync(fullOutput, await pdfDoc.save())
-    return ok({ filePath: fullInput, outputPath: fullOutput, text, pageCount: pages.length })
-  } catch (error) {
-    return fail(error, { filePath, outputPath })
-  }
-}
-
-/**
- * 创建新 PDF（支持文本段落、标题、分页）
- * content 为数组，每项：{ type: 'text'|'heading'|'pageBreak', text, fontSize, bold, color, align }
- */
-async function createPdf(outputPath, content = [], options = {}) {
-  try {
-    const fullOutput = resolvePath(outputPath)
-    fs.ensureDirSync(path.dirname(fullOutput))
-    return await new Promise((resolve, reject) => {
-      const doc = new PDFKit({
-        size: options.pageSize || 'A4',
-        margin: options.margin || 50,
-        info: {
-          Title: options.title || '',
-          Author: options.author || '',
-          Subject: options.subject || '',
-        },
-      })
-      const stream = fs.createWriteStream(fullOutput)
-      doc.pipe(stream)
-
-      for (const sec of content) {
-        switch (sec.type) {
-          case 'heading':
-            doc.font('Helvetica-Bold').fontSize(sec.fontSize || 18)
-            if (sec.color) doc.fillColor(sec.color)
-            doc.text(sec.text || '', { align: sec.align || 'left' })
-            doc.fillColor('#000000').font('Helvetica')
-            doc.moveDown(0.5)
-            break
-          case 'pageBreak':
-            doc.addPage()
-            break
-          case 'text':
-          default:
-            doc.font(sec.bold ? 'Helvetica-Bold' : 'Helvetica').fontSize(sec.fontSize || 12)
-            if (sec.color) doc.fillColor(sec.color)
-            doc.text(sec.text || '', { align: sec.align || 'left' })
-            doc.fillColor('#000000').font('Helvetica')
-            doc.moveDown(0.3)
-            break
-        }
-      }
-
-      doc.end()
-      stream.on('finish', () => resolve(ok({ outputPath: fullOutput, itemCount: content.length })))
-      stream.on('error', (err) => reject(err))
-    })
-  } catch (error) {
-    return fail(error, { outputPath })
-  }
-}
-
-/**
- * 将纯文本文件转换为 PDF
- */
-async function textToPdf(inputPath, outputPath, options = {}) {
-  try {
-    const fullInput = resolvePath(inputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    const text = fs.readFileSync(fullInput, 'utf8')
-    const content = [{ type: 'text', text }]
-    return await createPdf(outputPath, content, options)
-  } catch (error) {
-    return fail(error, { inputPath, outputPath })
-  }
-}
-
-/**
- * 在 PDF 末尾追加空白页
- */
-async function appendBlankPages(filePath, outputPath, count) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutput = resolvePath(outputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const pdfDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    const n = count || 1
-    for (let i = 0; i < n; i++) pdfDoc.addPage()
-    fs.writeFileSync(fullOutput, await pdfDoc.save())
-    return ok({ filePath: fullInput, outputPath: fullOutput, addedPages: n, totalPages: pdfDoc.getPageCount() })
-  } catch (error) {
-    return fail(error, { filePath, outputPath })
-  }
-}
-
-/**
- * 设置 PDF 文档元信息（标题、作者、主题、关键词）
- */
-async function setPdfMetadata(filePath, outputPath, metadata) {
-  try {
-    const fullInput = resolvePath(filePath)
-    const fullOutput = resolvePath(outputPath)
-    if (!fs.existsSync(fullInput)) return fail(`File does not exist: ${fullInput}`, { filePath: fullInput })
-    fs.ensureDirSync(path.dirname(fullOutput))
-    const pdfDoc = await PDFDocument.load(fs.readFileSync(fullInput))
-    if (metadata.title !== undefined) pdfDoc.setTitle(metadata.title)
-    if (metadata.author !== undefined) pdfDoc.setAuthor(metadata.author)
-    if (metadata.subject !== undefined) pdfDoc.setSubject(metadata.subject)
-    if (metadata.keywords !== undefined) pdfDoc.setKeywords(Array.isArray(metadata.keywords) ? metadata.keywords : [metadata.keywords])
-    if (metadata.creator !== undefined) pdfDoc.setCreator(metadata.creator)
-    fs.writeFileSync(fullOutput, await pdfDoc.save())
-    return ok({ filePath: fullInput, outputPath: fullOutput, metadata })
-  } catch (error) {
-    return fail(error, { filePath, outputPath })
-  }
-}
-
-/**
- * 统计 PDF 各页文字字符数（基于整体文本按页估算）
- */
-async function getPdfTextStats(filePath) {
-  try {
-    const readResult = await readPdfText(filePath)
-    if (!readResult.success) return readResult
-    const { text, pageCount } = readResult.data
-    const totalChars = text.length
-    const totalWords = text.trim().split(/\s+/).filter(Boolean).length
-    const lines = text.split('\n').filter(Boolean)
-    return ok({
-      filePath: resolvePath(filePath),
-      pageCount,
-      totalChars,
-      totalWords,
-      totalLines: lines.length,
-      avgCharsPerPage: pageCount > 0 ? Math.round(totalChars / pageCount) : 0,
-      avgWordsPerPage: pageCount > 0 ? Math.round(totalWords / pageCount) : 0,
-    })
-  } catch (error) {
-    return fail(error, { filePath })
-  }
-}
-
-
-async function pdfToLongImage(pdfPath, outputImagePath) {
-  try {
-    pdfPath = resolvePath(pdfPath)
-    outputImagePath = resolvePath(outputImagePath)
-    if (!fs.existsSync(pdfPath)) return fail(`File does not exist: ${pdfPath}`, { pdfPath })
-    // 0. 确保输出目录存在
-    fs.ensureDirSync(path.dirname(outputImagePath));
-    // 1. 读取 PDF 文档，设置 scale 参数可提高图片清晰度
-    const document = await pdf(pdfPath, { scale: 2 });
-    const pages = [];
-
-    // 2. 遍历每一页，将图片 Buffer 收集起来
-    for await (const pageBuffer of document) {
-      pages.push(pageBuffer);
-    }
-
-    if (pages.length === 0) {
-      return fail(`PDF has no pages`, { pdfPath })
-    }
-
-    // 3. 获取所有页面的尺寸，计算最大宽度和总高度
-    let maxWidth = 0;
-    let totalHeight = 0;
-    const pageHeights = [];
-    
-    for (const pageBuffer of pages) {
-      const metadata = await sharp(pageBuffer).metadata();
-      maxWidth = Math.max(maxWidth, metadata.width);
-      totalHeight += metadata.height;
-      pageHeights.push(metadata.height);
-    }
-
-    // 4. 构建 sharp 的输入：将图片 Buffer 数组叠加，并设置垂直偏移
-    const sharpInput = [];
-    let currentHeight = 0;
-    
-    for (let i = 0; i < pages.length; i++) {
-      const pageBuffer = pages[i];
-      const pageHeight = pageHeights[i];
-      
-      // 获取当前页面的实际宽度
-      const pageMetadata = await sharp(pageBuffer).metadata();
-      const pageWidth = pageMetadata.width;
-      
-      // 计算水平居中的偏移量
-      const leftOffset = Math.floor((maxWidth - pageWidth) / 2);
-      
-      sharpInput.push({
-        input: pageBuffer,
-        top: currentHeight, // 每页依次向下排列
-        left: leftOffset, // 水平居中
-      });
-      
-      currentHeight += pageHeight;
-    }
-
-    // 5. 使用 sharp 合成长图
-    await sharp({
-      create: {
-        width: maxWidth,
-        height: totalHeight, // 总高度 = 所有页面高度之和
-        channels: 4, // RGBA
-        background: { r: 255, g: 255, b: 255, alpha: 1 }, // 白色背景
-      },
-    })
-      .composite(sharpInput)
-      .png() // 输出为 PNG 格式
-      .toFile(outputImagePath)
-
-    return ok({ pdfPath, outputImagePath, pageCount: pages.length, width: maxWidth, height: totalHeight })
-  } catch (error) {
-    console.error(error)
-    return fail(error, { pdfPath, outputImagePath })
-  }
-}
-
-// ─── 工具描述 ─────────────────────────────────────────────────────────────────
-
-const descriptions = [
-  {
-    type: 'function',
-    function: {
-      name: 'readPdfText',
-      description: '读取 PDF 文件并提取全部文本内容。参数：filePath 为 PDF 路径。返回值：对象，包含 success、data（含 filePath、text、pageCount、wordCount、charCount）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: 'PDF 文件路径。' },
-        },
-        required: ['filePath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'getPdfInfo',
-      description: '获取 PDF 文档元信息（页数、标题、作者、创建时间、文件大小等）。参数：filePath 为 PDF 路径。返回值：对象，包含 success、data（含 pageCount、size、title、author、subject、keywords、creator、producer、creationDate、modificationDate）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: 'PDF 文件路径。' },
-        },
-        required: ['filePath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'getPdfPageCount',
-      description: '快速获取 PDF 总页数。参数：filePath 为 PDF 路径。返回值：对象，包含 success、data（含 filePath、pageCount）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: 'PDF 文件路径。' },
-        },
-        required: ['filePath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'searchPdfText',
-      description: '在 PDF 内容中搜索关键词，返回所有包含该词的行。参数：filePath 为 PDF 路径；keyword 为搜索关键词。返回值：对象，包含 success、data（含 keyword、matchCount、matches 数组）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: 'PDF 文件路径。' },
-          keyword: { type: 'string', description: '搜索关键词。' },
-        },
-        required: ['filePath', 'keyword'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'mergePdfs',
-      description: '将多个 PDF 文件合并为一个。参数：inputPaths 为 PDF 路径数组（按顺序合并）；outputPath 为输出路径。返回值：对象，包含 success、data（含 outputPath、pageCount、inputCount）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          inputPaths: { type: 'array', items: { type: 'string' }, description: '要合并的 PDF 文件路径数组。' },
-          outputPath: { type: 'string', description: '合并后输出的 PDF 文件路径。' },
-        },
-        required: ['inputPaths', 'outputPath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'splitPdf',
-      description: '将 PDF 按页拆分，每页生成独立 PDF 文件。参数：filePath 为源 PDF 路径；outputDir 为输出目录。返回值：对象，包含 success、data（含 outputDir、pageCount、files 数组）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '要拆分的 PDF 文件路径。' },
-          outputDir: { type: 'string', description: '拆分后各页 PDF 的输出目录。' },
-        },
-        required: ['filePath', 'outputDir'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'extractPdfPages',
-      description: '从 PDF 中提取指定页码（1-based）生成新 PDF。参数：filePath 为源 PDF 路径；outputPath 为输出路径；pages 为页码数组（如 [1, 3, 5]）。返回值：对象，包含 success、data（含 outputPath、extractedPages）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          pages: { type: 'array', items: { type: 'number' }, description: '要提取的页码数组（从 1 开始），如 [1, 2, 5]。' },
-        },
-        required: ['filePath', 'outputPath', 'pages'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'rotatePdfPages',
-      description: '旋转 PDF 页面。参数：filePath 为源 PDF 路径；outputPath 为输出路径；rotateDegrees 为旋转角度（90/180/270，默认 90）；pageNumbers 为要旋转的页码数组（1-based，省略则旋转全部页）。返回值：对象，包含 success、data（含 outputPath、degrees、rotatedPages）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          rotateDegrees: { type: 'number', description: '旋转角度（90/180/270），默认 90。' },
-          pageNumbers: { type: 'array', items: { type: 'number' }, description: '要旋转的页码数组（从 1 开始），省略则旋转全部页。' },
-        },
-        required: ['filePath', 'outputPath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'addWatermarkText',
-      description: '向 PDF 全部页面添加文字水印。参数：filePath 为源 PDF 路径；outputPath 为输出路径；text 为水印文字；options 可选（fontSize 默认 48，opacity 默认 0.2，angle 默认 45，color: {r,g,b}）。返回值：对象，包含 success、data（含 outputPath、text、pageCount）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          text: { type: 'string', description: '水印文字内容。' },
-          options: { type: 'object', description: '可选：{ fontSize, opacity, angle, color: {r,g,b} }。' },
-        },
-        required: ['filePath', 'outputPath', 'text'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'createPdf',
-      description: `创建新的 PDF 文档，支持多种内容类型。
-参数：outputPath 为输出路径；content 为内容数组，每项格式：
-- 文本：{ type:'text', text, fontSize, bold, color, align }
-- 标题：{ type:'heading', text, fontSize, color, align }
-- 分页：{ type:'pageBreak' }
-options 可选：{ pageSize: A4/A3/Letter, margin, title, author, subject }。
-返回值：对象，包含 success、data（含 outputPath、itemCount）、error。`,
-      parameters: {
-        type: 'object',
-        properties: {
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          content: { type: 'array', description: '内容块数组，支持 text/heading/pageBreak 类型。', items: { type: 'object' } },
-          options: { type: 'object', description: '可选：{ pageSize, margin, title, author, subject }。' },
-        },
-        required: ['outputPath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'textToPdf',
-      description: '将纯文本文件（.txt）转换为 PDF 文档。参数：inputPath 为 .txt 文件路径；outputPath 为输出 PDF 路径；options 可选（同 createPdf options）。返回值：对象，包含 success、data（含 outputPath）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          inputPath: { type: 'string', description: '源 .txt 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          options: { type: 'object', description: '可选：{ pageSize, margin, title, author }。' },
-        },
-        required: ['inputPath', 'outputPath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'appendBlankPages',
-      description: '在 PDF 末尾追加指定数量的空白页。参数：filePath 为源 PDF 路径；outputPath 为输出路径；count 为追加页数（默认 1）。返回值：对象，包含 success、data（含 outputPath、addedPages、totalPages）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          count: { type: 'number', description: '要追加的空白页数量，默认 1。' },
-        },
-        required: ['filePath', 'outputPath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'setPdfMetadata',
-      description: '设置 PDF 文档的元信息（标题、作者、主题、关键词）。参数：filePath 为源 PDF 路径；outputPath 为输出路径；metadata 为元信息对象（含 title/author/subject/keywords/creator 字段）。返回值：对象，包含 success、data（含 outputPath、metadata）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputPath: { type: 'string', description: '输出 PDF 文件路径。' },
-          metadata: {
-            type: 'object',
-            description: '元信息对象，如 { title, author, subject, keywords, creator }。',
-          },
-        },
-        required: ['filePath', 'outputPath', 'metadata'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'getPdfTextStats',
-      description: '统计 PDF 文档的文字数量（总字数、总字符数、总行数及每页平均值）。参数：filePath 为 PDF 路径。返回值：对象，包含 success、data（含 pageCount、totalChars、totalWords、totalLines、avgCharsPerPage、avgWordsPerPage）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          filePath: { type: 'string', description: 'PDF 文件路径。' },
-        },
-        required: ['filePath'],
-      },
-    },
-  },
-  {
-    type: 'function',
-    function: {
-      name: 'pdfToLongImage',
-      description: '将 PDF 所有页面依次拼接为一张竖向长图（PNG 格式）。参数：pdfPath 为 PDF 文件路径；outputImagePath 为输出 PNG 图片路径。返回值：对象，包含 success、data（含 pdfPath、outputImagePath、pageCount、width、height）、error。',
-      parameters: {
-        type: 'object',
-        properties: {
-          pdfPath: { type: 'string', description: '源 PDF 文件路径。' },
-          outputImagePath: { type: 'string', description: '输出长图的 PNG 文件路径。' },
-        },
-        required: ['pdfPath', 'outputImagePath'],
-      },
-    },
-  },
-]
-
-// ─── 导出 ──────────────────────────────────────────────────────────────────────
-
-const functions = {
-  readPdfText,
-  getPdfInfo,
-  getPdfPageCount,
-  searchPdfText,
-  mergePdfs,
-  splitPdf,
-  extractPdfPages,
-  rotatePdfPages,
-  addWatermarkText,
-  createPdf,
-  textToPdf,
-  appendBlankPages,
-  setPdfMetadata,
-  getPdfTextStats,
-  pdfToLongImage,
-}
-
-const PdfTool = {
-  name: 'PdfTool',
-  description: '提供 PDF 文档的文本提取、搜索、元信息读写、合并、拆分、提取页面、旋转、添加水印、创建、文本转 PDF、PDF 转长图等全面处理能力',
-  platform: 'all',
-  descriptions,
-  functions,
-  isSystem: true
-}
-
-module.exports = PdfTool