xinghen.work/server/pptx-notes.ts

import { readFile } from 'node:fs/promises'
import JSZip from 'jszip'
import { XMLParser } from 'fast-xml-parser'

const parser = new XMLParser({ ignoreAttributes: false, removeNSPrefix: true })
const list = <T>(value: T | T[] | undefined): T[] => value === undefined ? [] : Array.isArray(value) ? value : [value]

function textNodes(value: unknown): string[] {
  if (!value || typeof value !== 'object') return []
  return Object.entries(value).flatMap(([key, child]) => key === 't'
    ? [typeof child === 'string' ? child : String(child ?? '')]
    : Array.isArray(child) ? child.flatMap(textNodes) : textNodes(child))
}

export async function extractPptxSpeakerNotes(path: string): Promise<string[]> {
  const zip = await JSZip.loadAsync(await readFile(path), { checkCRC32: true })
  const files = Object.keys(zip.files).filter(name => /^ppt\/notesSlides\/notesSlide\d+\.xml$/.test(name))
    .sort((a, b) => Number(a.match(/\d+(?=\.xml$)/)![0]) - Number(b.match(/\d+(?=\.xml$)/)![0]))
  if (files.length > 500) throw new Error('PPTX 页数超出备注解析限制')
  const notes: string[] = []
  for (const name of files) {
    const xml = await zip.file(name)!.async('string')
    const shapes = list(parser.parse(xml)?.notes?.cSld?.spTree?.sp)
    const body = shapes.find((shape: any) => shape?.nvSpPr?.nvPr?.ph?.['@_type'] === 'body') as any
    const paragraphs = list(body?.txBody?.p)
      .map((paragraph: unknown) => textNodes(paragraph).join('').trim())
      .filter(Boolean)
    notes.push(paragraphs.join('\n'))
  }
  return notes
}