pptx2js 0.4.0

package/lib/table.js

@@ -0,0 +1,138 @@
+/**
+ * 表格提取（design.html §4.5）
+ */
+const { asArray, attr, child, textContent } = require('./xml-utils');
+const { getGraphicXfrm } = require('./graphic');
+const { boundsFromXfrm } = require('./utils/bounds');
+const { resolveFillColor } = require('./utils/color');
+const { emuToInch } = require('./utils/emu');
+
+/**
+ * @param {object} graphicFrame
+ * @param {object} ctx
+ * @returns {import('./extractor').SlideEntity|null}
+ */
+function extractTable(graphicFrame, ctx) {
+  const bounds = boundsFromXfrm(getGraphicXfrm(graphicFrame), ctx.offset);
+  const graphicData = child(child(graphicFrame, 'a:graphic'), 'a:graphicData');
+  const tbl = child(graphicData, 'a:tbl');
+
+  if (!tbl) {
+    return {
+      slideIndex: ctx.slideIndex,
+      slidePath: ctx.slidePath,
+      decision: 'SKIP',
+      kind: 'skip',
+      bounds,
+      skipReason: '表格数据无法解析（无内联 a:tbl）',
+    };
+  }
+
+  const rows = extractTableRows(tbl, ctx.scheme);
+  if (rows.length === 0) return null;
+
+  const colWidths = extractColWidths(tbl);
+
+  return {
+    slideIndex: ctx.slideIndex,
+    slidePath: ctx.slidePath,
+    decision: 'FULL',
+    kind: 'table',
+    bounds,
+    table: { rows, colWidths },
+  };
+}
+
+/**
+ * @param {object} tbl
+ * @param {Record<string, string>} scheme
+ */
+function extractTableRows(tbl, scheme) {
+  const rows = [];
+  for (const tr of asArray(tbl['a:tr'])) {
+    const cells = [];
+    for (const tc of asArray(tr['a:tc'])) {
+      const text = extractCellText(tc);
+      const tcPr = child(tc, 'a:tcPr');
+      const fill = tcPr ? resolveFillColor(tcPr, scheme).color : null;
+      const merge = {};
+      const rowSpan = attr(tcPr, 'rowSpan');
+      const gridSpan = attr(tcPr, 'gridSpan');
+      if (rowSpan && rowSpan !== '1') merge.rowspan = parseInt(rowSpan, 10);
+      if (gridSpan && gridSpan !== '1') merge.colspan = parseInt(gridSpan, 10);
+
+      const cell = { text, options: {} };
+      if (fill) cell.options.fill = { color: fill };
+      if (Object.keys(merge).length) Object.assign(cell.options, merge);
+
+      const border = extractCellBorder(tcPr, scheme);
+      if (border) cell.options.border = border;
+
+      cells.push(cell);
+    }
+    if (cells.length) rows.push(cells);
+  }
+  return rows;
+}
+
+/**
+ * @param {object|null|undefined} tcPr
+ * @param {Record<string, string>} scheme
+ */
+function extractCellBorder(tcPr, scheme) {
+  if (!tcPr) return null;
+  const border = {};
+  const sides = {
+    left: 'a:lnL',
+    right: 'a:lnR',
+    top: 'a:lnT',
+    bottom: 'a:lnB',
+  };
+  let hasAny = false;
+
+  for (const [side, key] of Object.entries(sides)) {
+    const ln = child(tcPr, key);
+    if (!ln) continue;
+    const { color } = resolveFillColor(ln, scheme);
+    const w = parseInt(attr(ln, 'w') ?? '0', 10);
+    if (color || w) {
+      border[side] = {
+        ...(color && { color }),
+        ...(w > 0 && { pt: Math.round(w / 12700) }),
+      };
+      hasAny = true;
+    }
+  }
+  return hasAny ? border : null;
+}
+
+/**
+ * @param {object} tc
+ */
+function extractCellText(tc) {
+  const txBody = child(tc, 'a:txBody');
+  if (!txBody) return '';
+  const parts = [];
+  for (const p of asArray(txBody['a:p'])) {
+    for (const r of asArray(p['a:r'])) {
+      parts.push(textContent(r['a:t']));
+    }
+  }
+  return parts.join('');
+}
+
+/**
+ * @param {object} tbl
+ * @returns {number[]|undefined}
+ */
+function extractColWidths(tbl) {
+  const grid = child(tbl, 'a:tblGrid');
+  const cols = asArray(child(grid, 'a:gridCol'));
+  if (!cols.length) return undefined;
+  return cols.map((col) => {
+    const w = parseInt(attr(col, 'w') ?? '0', 10);
+    return Math.round(emuToInch(w) * 1000) / 1000;
+  });
+}
+
+module.exports = { extractTable };

package/lib/unpacker.js

@@ -0,0 +1,40 @@
+/**
+ * ① 解压与索引
+ * JSZip 解压 PPTX，建立文件清单。
+ */
+const fs = require('fs');
+const JSZip = require('jszip');
+
+/**
+ * @typedef {object} PptxArchive
+ * @property {JSZip} zip
+ * @property {Map<string, string>} files - 路径 → 文本内容或 __binary__
+ */
+
+/**
+ * @param {string} filePath
+ * @returns {Promise<PptxArchive>}
+ */
+async function unpack(filePath) {
+  const buffer = fs.readFileSync(filePath);
+  const zip = await JSZip.loadAsync(buffer);
+
+  /** @type {Map<string, string>} */
+  const files = new Map();
+  for (const [entryPath, entry] of Object.entries(zip.files)) {
+    if (entry.dir) continue;
+    const isText =
+      entryPath.endsWith('.xml') ||
+      entryPath.endsWith('.rels') ||
+      entryPath.includes('[Content_Types]');
+    if (isText) {
+      files.set(entryPath, await entry.async('string'));
+    } else {
+      files.set(entryPath, '__binary__');
+    }
+  }
+
+  return { zip, files };
+}
+
+module.exports = { unpack };

package/lib/utils/bounds.js

@@ -0,0 +1,34 @@
+/**
+ * 形状坐标：a:xfrm → 英寸边界
+ */
+const { attr, child } = require('../xml-utils');
+const { emuToInch } = require('./emu');
+
+/**
+ * @param {object|null|undefined} xfrm
+ * @param {{ x: number, y: number }} offset EMU
+ * @returns {{ x: number, y: number, w: number, h: number }}
+ */
+function boundsFromXfrm(xfrm, offset = { x: 0, y: 0 }) {
+  const off = child(xfrm, 'a:off');
+  const ext = child(xfrm, 'a:ext');
+  const x = parseInt(attr(off, 'x') ?? '0', 10) + offset.x;
+  const y = parseInt(attr(off, 'y') ?? '0', 10) + offset.y;
+  const cx = parseInt(attr(ext, 'cx') ?? '0', 10);
+  const cy = parseInt(attr(ext, 'cy') ?? '0', 10);
+  return {
+    x: round3(emuToInch(x)),
+    y: round3(emuToInch(y)),
+    w: round3(emuToInch(cx)),
+    h: round3(emuToInch(cy)),
+  };
+}
+
+/**
+ * @param {number} n
+ */
+function round3(n) {
+  return Math.round(n * 1000) / 1000;
+}
+
+module.exports = { boundsFromXfrm };

package/lib/utils/color.js

@@ -0,0 +1,128 @@
+/**
+ * 颜色规范化（srgbClr、schemeClr、渐变首色标）
+ */
+const { asArray, attr, child } = require('../xml-utils');
+
+/** @type {Record<string, string>} */
+const DEFAULT_SCHEME = {
+  dk1: '000000',
+  lt1: 'FFFFFF',
+  dk2: '44546A',
+  lt2: 'E7E6E6',
+  accent1: '4472C4',
+  accent2: 'ED7D31',
+  accent3: 'A5A5A5',
+  accent4: 'FFC000',
+  accent5: '5B9BD5',
+  accent6: '70AD47',
+  hlink: '0563C1',
+  folHlink: '954F72',
+};
+
+/**
+ * @param {Record<string, object>|null} parsed
+ * @param {string|null} themePath
+ * @returns {Record<string, string>}
+ */
+function loadColorScheme(parsed, themePath) {
+  if (!themePath || !parsed[themePath]) return { ...DEFAULT_SCHEME };
+
+  const theme = child(parsed[themePath], 'a:theme');
+  const clrScheme = child(child(theme, 'a:themeElements'), 'a:clrScheme');
+  if (!clrScheme) return { ...DEFAULT_SCHEME };
+
+  const scheme = { ...DEFAULT_SCHEME };
+  for (const name of Object.keys(DEFAULT_SCHEME)) {
+    const slot = child(clrScheme, `a:${name}`);
+    const rgb = extractRgbFromColorNode(slot);
+    if (rgb) scheme[name] = rgb;
+  }
+  return scheme;
+}
+
+/**
+ * @param {object|null|undefined} colorNode a:srgbClr / a:schemeClr 的父级
+ * @returns {string|null}
+ */
+function extractRgbFromColorNode(colorNode) {
+  if (!colorNode) return null;
+
+  const srgb = child(colorNode, 'a:srgbClr');
+  if (srgb) {
+    const val = attr(srgb, 'val');
+    return val ? val.toUpperCase() : null;
+  }
+
+  const sys = child(colorNode, 'a:sysClr');
+  if (sys) {
+    const last = attr(sys, 'lastClr');
+    return last ? last.toUpperCase() : null;
+  }
+
+  return null;
+}
+
+/**
+ * @param {object|null|undefined} fillNode a:solidFill | a:gradFill 等
+ * @param {Record<string, string>} scheme
+ * @returns {{ color: string|null, degraded: boolean }}
+ */
+function resolveFillColor(fillNode, scheme) {
+  if (!fillNode) return { color: null, degraded: false };
+
+  const solid = child(fillNode, 'a:solidFill');
+  if (solid) {
+    return { color: resolveColorFromContainer(solid, scheme), degraded: false };
+  }
+
+  const grad = child(fillNode, 'a:gradFill');
+  if (grad) {
+    const gs = asArray(child(child(grad, 'a:gsLst'), 'a:gs'))[0];
+    return {
+      color: gs ? resolveColorFromContainer(gs, scheme) : null,
+      degraded: true,
+    };
+  }
+
+  return { color: null, degraded: false };
+}
+
+/**
+ * @param {object|null|undefined} container 含 a:srgbClr / a:schemeClr
+ * @param {Record<string, string>} scheme
+ * @returns {string|null}
+ */
+function resolveColorFromContainer(container, scheme) {
+  if (!container) return null;
+
+  const srgb = child(container, 'a:srgbClr');
+  if (srgb) {
+    const val = attr(srgb, 'val');
+    return val ? val.toUpperCase() : null;
+  }
+
+  const schemeClr = child(container, 'a:schemeClr');
+  if (schemeClr) {
+    const name = attr(schemeClr, 'val');
+    return name ? scheme[name] ?? DEFAULT_SCHEME[name] ?? null : null;
+  }
+
+  return extractRgbFromColorNode(container);
+}
+
+/**
+ * @param {Record<string, string>} scheme
+ * @param {object|null|undefined} clrNode
+ * @returns {string|null}
+ */
+function resolveColor(scheme, clrNode) {
+  return resolveColorFromContainer(clrNode, scheme);
+}
+
+module.exports = {
+  DEFAULT_SCHEME,
+  loadColorScheme,
+  resolveFillColor,
+  resolveColor,
+  resolveColorFromContainer,
+};

package/lib/utils/emu.js

@@ -0,0 +1,20 @@
+/** EMU（English Metric Units）与英寸换算，1 英寸 = 914400 EMU */
+const EMU_PER_INCH = 914400;
+
+/**
+ * @param {number} emu
+ * @returns {number}
+ */
+function emuToInch(emu) {
+  return emu / EMU_PER_INCH;
+}
+
+/**
+ * @param {number} inch
+ * @returns {number}
+ */
+function inchToEmu(inch) {
+  return inch * EMU_PER_INCH;
+}
+
+module.exports = { EMU_PER_INCH, emuToInch, inchToEmu };

package/lib/xml-parser.js

@@ -0,0 +1,25 @@
+/**
+ * 项目唯一导出的 XML 解析器实例。
+ * 全项目必须使用此入口，禁止在模块内自行实例化解析器。
+ * @see design.html §4.6
+ */
+const xml2js = require('xml2js');
+
+const PARSER_OPTIONS = {
+  explicitArray: false,
+  mergeAttrs: false,
+  explicitCharkey: false,
+  tagNameProcessors: [],
+  attrNameProcessors: [],
+  xmlns: false,
+};
+
+/**
+ * @param {string} str
+ * @returns {Promise<object>}
+ */
+function parseXml(str) {
+  return xml2js.parseStringPromise(str, PARSER_OPTIONS);
+}
+
+module.exports = { parseXml, PARSER_OPTIONS };

package/lib/xml-utils.js

@@ -0,0 +1,68 @@
+/**
+ * xml2js 对象树访问工具（配合 lib/xml-parser.js 统一配置）
+ */
+
+/**
+ * @param {*} value
+ * @returns {Array}
+ */
+function asArray(value) {
+  if (value == null) return [];
+  return Array.isArray(value) ? value : [value];
+}
+
+/**
+ * @param {object|null|undefined} node
+ * @returns {Record<string, string>}
+ */
+function attrs(node) {
+  if (node == null) return {};
+  if (node.$) return node.$;
+  return node;
+}
+
+/**
+ * @param {object|null|undefined} node
+ * @param {string} name
+ * @returns {string|undefined}
+ */
+function attr(node, name) {
+  return attrs(node)[name];
+}
+
+/**
+ * @param {object|null|undefined} node
+ * @param {string} key
+ * @returns {*}
+ */
+function child(node, key) {
+  if (node == null) return undefined;
+  return node[key];
+}
+
+/**
+ * @param {object|null|undefined} node
+ * @param {string[]} keys
+ * @returns {*}
+ */
+function firstChild(node, keys) {
+  for (const key of keys) {
+    const v = child(node, key);
+    if (v != null) return v;
+  }
+  return undefined;
+}
+
+/**
+ * @param {object|null|undefined} node
+ * @returns {string}
+ */
+function textContent(node) {
+  if (node == null) return '';
+  if (typeof node === 'string') return node;
+  if (typeof node === 'number') return String(node);
+  if (node._ != null) return String(node._);
+  return '';
+}
+
+module.exports = { asArray, attrs, attr, child, firstChild, textContent };

package/package.json

@@ -0,0 +1,41 @@
+{
+  "name": "pptx2js",
+  "version": "0.4.0",
+  "description": "将 .pptx 文件转换为可运行的 PptxGenJS 生成脚本",
+  "main": "lib/index.js",
+  "bin": {
+    "pptx2js": "bin/pptx2js.js"
+  },
+  "scripts": {
+    "test": "jest",
+    "test:watch": "jest --watch"
+  },
+  "keywords": [
+    "pptx",
+    "pptxgenjs",
+    "converter",
+    "powerpoint",
+    "presentation",
+    "nodejs",
+    "cli"
+  ],
+  "author": "yuese12333",
+  "license": "MIT",
+  "engines": {
+    "node": ">=18"
+  },
+  "dependencies": {
+    "chalk": "^4.1.2",
+    "commander": "^12.1.0",
+    "jszip": "^3.10.1",
+    "xml2js": "^0.6.2"
+  },
+  "devDependencies": {
+    "jest": "^29.7.0"
+  },
+  "files": [
+    "lib",
+    "bin",
+    "README.md"
+  ]
+}