fix: preserve ppt text styles during replacement

2026-06-26 17:16:46 +08:00
parent 0b2a8b0190
commit 203f3b9ec1
1 changed files with 206 additions and 38 deletions
@@ -55,7 +55,6 @@ type ReportStudentData = {
  comment: string
  comments: string
  teacherName: string
-  teacher_name: string
 }

 function getReportStoragePath(): string {
@@ -136,6 +135,28 @@ function getFormatFromExtension(extension: string): ReportFormat {
  return 'ppt'
 }

+function escapeXml(value: string): string {
+  return value
+    .replace(/&/g, '&amp;')
+    .replace(/</g, '&lt;')
+    .replace(/>/g, '&gt;')
+    .replace(/"/g, '&quot;')
+    .replace(/'/g, '&apos;')
+}
+
+function decodeXml(value: string): string {
+  return value
+    .replace(/&lt;/g, '<')
+    .replace(/&gt;/g, '>')
+    .replace(/&quot;/g, '"')
+    .replace(/&apos;/g, "'")
+    .replace(/&amp;/g, '&')
+}
+
+function escapeRegExp(value: string): string {
+  return value.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
+}
+
 function getStudentData(student: StudentProfileEntity, classItem: ClassEntity): ReportStudentData {
  const friends = parseJsonList(student.friends).join('、') || ' '
  const hobbies = parseJsonList(student.hobbies).join('、') || ' '
@@ -168,8 +189,7 @@ function getStudentData(student: StudentProfileEntity, classItem: ClassEntity):
    traits: student.traits || ' ',
    comment: student.comment || '暂无评语',
    comments: student.comment || '暂无评语',
-    teacherName,
-    teacher_name: teacherName
+    teacherName
  }
 }

@@ -198,7 +218,7 @@ function buildTextReplacements(studentData: ReportStudentData, classItem: ClassE
    comments: studentData.comments,
    comment: studentData.comment,
    teacherName: studentData.teacherName,
-    teacher_name: studentData.teacher_name,
+    teacher_name: studentData.teacherName,
    englishName: studentData.englishName,
    english_name: studentData.englishName,
    sex: studentData.sex,
@@ -400,6 +420,187 @@ function logSlideObjects(ppt: PPTXTemplater, traceId: string): void {
  }
 }

+function getSlidePaths(zip: JSZip): string[] {
+  return Object.keys(zip.files)
+    .filter((path) => /^ppt\/slides\/slide\d+\.xml$/.test(path))
+    .sort((first, second) => {
+      const firstNumber = Number(first.match(/slide(\d+)\.xml$/)?.[1] ?? 0)
+      const secondNumber = Number(second.match(/slide(\d+)\.xml$/)?.[1] ?? 0)
+
+      return firstNumber - secondNumber
+    })
+}
+
+function hasShapeName(shapeXml: string, shapeName: string): boolean {
+  const normalizedShapeName = normalizePptObjectName(shapeName)
+  const shapeProperties = shapeXml.match(/<p:cNvPr[^>]*>/)?.[0] ?? ''
+  const candidateNames = ['name', 'descr', 'title']
+    .map((attributeName) => getXmlAttributeValue(shapeProperties, attributeName))
+    .filter((value): value is string => Boolean(value))
+
+  return candidateNames.some(
+    (candidateName) => normalizePptObjectName(decodeXml(candidateName)) === normalizedShapeName
+  )
+}
+
+function getXmlAttributeValue(xml: string, attributeName: string): string | null {
+  const match = xml.match(new RegExp(`\\s${attributeName}="([^"]*)"`))
+
+  return match?.[1] ?? null
+}
+
+function replaceTextNodeValue(
+  textNode: string,
+  replacer: (value: string) => string | null
+): string {
+  const match = textNode.match(/^(<a:t(?:\s[^>]*)?>)([\s\S]*?)(<\/a:t>)$/)
+
+  if (!match) {
+    return textNode
+  }
+
+  const nextValue = replacer(match[2])
+
+  if (nextValue === null) {
+    return textNode
+  }
+
+  return `${match[1]}${nextValue}${match[3]}`
+}
+
+function replaceShapeTextPreservingStyle(slideXml: string, shapeName: string, text: string): string {
+  return slideXml.replace(/<p:sp[\s\S]*?<\/p:sp>/g, (shapeXml: string) => {
+    if (!hasShapeName(shapeXml, shapeName)) {
+      return shapeXml
+    }
+
+    const escapedText = escapeXml(text)
+    const textNodes = [...shapeXml.matchAll(/<a:t(?:\s[^>]*)?>[\s\S]*?<\/a:t>/g)]
+
+    if (textNodes.length === 0) {
+      return shapeXml
+    }
+
+    let replacedFirstText = false
+
+    return shapeXml.replace(/<a:t(?:\s[^>]*)?>[\s\S]*?<\/a:t>/g, (textNode: string) => {
+      if (replacedFirstText) {
+        return replaceTextNodeValue(textNode, () => '')
+      }
+
+      replacedFirstText = true
+      return replaceTextNodeValue(textNode, () => escapedText)
+    })
+  })
+}
+
+function replaceInlineTextPlaceholderPreservingStyle(
+  slideXml: string,
+  placeholder: string,
+  text: string
+): string {
+  const placeholderPattern = new RegExp(`\\{\\{\\s*${escapeRegExp(placeholder)}\\s*\\}\\}`, 'g')
+
+  return slideXml.replace(/<a:t(?:\s[^>]*)?>[\s\S]*?<\/a:t>/g, (textNode: string) =>
+    replaceTextNodeValue(textNode, (value) => {
+      placeholderPattern.lastIndex = 0
+
+      if (!placeholderPattern.test(value)) {
+        return null
+      }
+
+      placeholderPattern.lastIndex = 0
+      return value.replace(placeholderPattern, escapeXml(text))
+    })
+  )
+}
+
+function getTextNodeValue(textNode: string): string {
+  return textNode.match(/^<a:t(?:\s[^>]*)?>([\s\S]*?)<\/a:t>$/)?.[1] ?? ''
+}
+
+function replaceShapeInlinePlaceholdersPreservingStyle(
+  slideXml: string,
+  replacements: Record<string, string>
+): string {
+  return slideXml.replace(/<p:sp[\s\S]*?<\/p:sp>/g, (shapeXml: string) => {
+    const textNodes = [...shapeXml.matchAll(/<a:t(?:\s[^>]*)?>[\s\S]*?<\/a:t>/g)]
+
+    if (textNodes.length === 0) {
+      return shapeXml
+    }
+
+    const combinedText = textNodes.map((match) => getTextNodeValue(match[0])).join('')
+    let nextCombinedText = combinedText
+
+    for (const [placeholder, text] of Object.entries(replacements)) {
+      const placeholderPattern = new RegExp(
+        `\\{\\{\\s*${escapeRegExp(placeholder)}\\s*\\}\\}`,
+        'g'
+      )
+      nextCombinedText = nextCombinedText.replace(placeholderPattern, escapeXml(text))
+    }
+
+    if (nextCombinedText === combinedText) {
+      return shapeXml
+    }
+
+    let replacedFirstText = false
+
+    return shapeXml.replace(/<a:t(?:\s[^>]*)?>[\s\S]*?<\/a:t>/g, (textNode: string) => {
+      if (replacedFirstText) {
+        return replaceTextNodeValue(textNode, () => '')
+      }
+
+      replacedFirstText = true
+      return replaceTextNodeValue(textNode, () => nextCombinedText)
+    })
+  })
+}
+
+async function replacePptxTextPreservingStyle(
+  filePath: string,
+  replacements: Record<string, string>,
+  traceId: string
+): Promise<void> {
+  const fileBuffer = await readFile(filePath)
+  const zip = await JSZip.loadAsync(fileBuffer)
+  let replacementCount = 0
+
+  for (const slidePath of getSlidePaths(zip)) {
+    const slideFile = zip.file(slidePath)
+
+    if (!slideFile) {
+      continue
+    }
+
+    const originalSlideXml = await slideFile.async('text')
+    let slideXml = originalSlideXml
+
+    for (const [placeholder, text] of Object.entries(replacements)) {
+      slideXml = replaceInlineTextPlaceholderPreservingStyle(slideXml, placeholder, text)
+      slideXml = replaceShapeTextPreservingStyle(slideXml, placeholder, text)
+    }
+    slideXml = replaceShapeInlinePlaceholdersPreservingStyle(slideXml, replacements)
+
+    if (slideXml !== originalSlideXml) {
+      replacementCount += 1
+      zip.file(slidePath, slideXml)
+    }
+  }
+
+  const outputBuffer = await zip.generateAsync({
+    type: 'nodebuffer',
+    compression: 'DEFLATE'
+  })
+  await writeFile(filePath, outputBuffer)
+
+  logger.debug(`[${traceId}] 保留样式的 PPTX 文本替换完成`, {
+    touchedSlideCount: replacementCount,
+    replacementKeys: Object.keys(replacements)
+  })
+}
+
 function findImageReplacementTargets(
  ppt: PPTXTemplater,
  slideNumber: number,
@@ -478,24 +679,6 @@ function findImageReplacementTargets(
  return fallbackTargets
 }

-async function updateShapeTextIfExists(
-  ppt: PPTXTemplater,
-  shapeName: string,
-  text: string
-): Promise<void> {
-  for (let slideNumber = 1; slideNumber <= ppt.slideCount; slideNumber += 1) {
-    try {
-      await ppt.useSlide(slideNumber).updateShape(shapeName, { text })
-    } catch (error) {
-      const message = error instanceof Error ? error.message : ''
-
-      if (!message.includes('not found')) {
-        throw error
-      }
-    }
-  }
-}
-
 async function replaceImageIfExists(
  ppt: PPTXTemplater,
  replacementPlan: PptImageReplacementPlan,
@@ -698,22 +881,6 @@ async function buildPptxReport(
  })
  logSlideObjects(ppt, traceId)

-  // 文本替换分两步：
-  // 1. replaceText({ '{{name}}': '张三' }) 处理文本框里的 Mustache 风格占位符；
-  // 2. updateShape('name', { text }) 处理“形状对象名就是字段名”的模板。
-  ppt.replaceText(
-    Object.fromEntries(
-      Object.entries(textReplacements).map(([placeholder, value]) => [`{{${placeholder}}}`, value])
-    )
-  )
-
-  for (const [shapeName, text] of Object.entries(textReplacements)) {
-    await updateShapeTextIfExists(ppt, shapeName, text)
-  }
-  logger.debug(`[${traceId}] 文本替换完成`, {
-    replacementCount: Object.keys(textReplacements).length
-  })
-
  const imageReplacementPlans = buildImageReplacementPlans(
    ppt,
    buildImageReplacements(),
@@ -729,6 +896,7 @@ async function buildPptxReport(
  }

  await ppt.saveToFile(outputPath)
+  await replacePptxTextPreservingStyle(outputPath, textReplacements, traceId)
  logger.info(`[${traceId}] PPTX 报告保存完成`, {
    outputPath,
    replacedImageCount: replacedImageKeys.size