pptx2js 0.4.0 → 0.4.3

package/lib/text-utils.js

@@ -0,0 +1,235 @@
+/**
+ * 文本 run 提取（从 extractor 拆出，避免 extractor ↔ table 循环依赖）
+ */
+const { attr, child, childNodes, children, textContent } = require('./xml-utils');
+const {
+  resolveFillColor,
+  resolveColorFromContainer,
+} = require('./color');
+const { compressTextRuns } = require('./run-utils');
+const {
+  getLstStyleDefRPrByLevel,
+  mergeRPrAttrs,
+} = require('./placeholder');
+
+const PARA_ALIGN_MAP = {
+  l: 'left',
+  ctr: 'center',
+  r: 'right',
+  just: 'justify',
+  dist: 'justify',
+};
+
+/**
+ * @param {object} txBody a:txBody
+ * @param {Record<string, string>} scheme
+ * @param {{ relIndex?: import('./rels').RelationIndex, slidePath?: string }|null} linkCtx
+ */
+function extractTextRuns(txBody, scheme, linkCtx) {
+  const runs = [];
+  const bodyParaOpts = extractBodyPrParaOptions(txBody);
+  const defRPrByLevel = getLstStyleDefRPrByLevel(txBody);
+  const defaultFontSize =
+    bodyParaOpts._defaultFontSize ??
+    runFontSizeFromDefRPr(defRPrByLevel[0]);
+  const paragraphs = children(txBody, 'a:p');
+
+  for (let pi = 0; pi < paragraphs.length; pi++) {
+    const p = paragraphs[pi];
+    let currentPPr = null;
+    let firstPPr = null;
+
+    for (const node of childNodes(p)) {
+      if (node.tag === 'a:pPr') {
+        currentPPr = node;
+        if (!firstPPr) firstPPr = node;
+        continue;
+      }
+
+      const pPr = currentPPr ?? firstPPr;
+      const paraOpts = {
+        ...bodyParaOpts,
+        ...extractParaOptions(pPr, defaultFontSize),
+      };
+      delete paraOpts._defaultFontSize;
+      const bullet0 = extractBullet(firstPPr);
+      const bullet = extractBullet(pPr) ?? bullet0;
+
+      const lvl = paraOpts.indentLevel ?? 0;
+      const defRPr =
+        defRPrByLevel[lvl] ?? defRPrByLevel[0] ?? null;
+
+      if (node.tag === 'a:r') {
+        const text = textContent(child(node, 'a:t'));
+        if (!text) continue;
+        const rPr = mergeRPrAttrs(child(node, 'a:rPr'), defRPr);
+        const options = extractRunOptions(rPr, scheme, linkCtx);
+        Object.assign(options, paraOpts);
+        if (bullet) Object.assign(options, bullet);
+        runs.push({ text, options, degraded: options._degraded });
+      } else if (node.tag === 'a:fld') {
+        const text = textContent(child(node, 'a:t'));
+        if (!text) continue;
+        const rPr = mergeRPrAttrs(child(node, 'a:rPr'), defRPr);
+        const options = extractRunOptions(rPr, scheme, linkCtx);
+        Object.assign(options, paraOpts);
+        runs.push({ text, options, degraded: options._degraded });
+      } else if (node.tag === 'a:br') {
+        runs.push({ text: '\n', options: { ...paraOpts }, degraded: false });
+      }
+    }
+
+    if (pi < paragraphs.length - 1) {
+      const last = runs[runs.length - 1];
+      if (!last || !last.text.endsWith('\n')) {
+        const paraOpts = {
+          ...bodyParaOpts,
+          ...extractParaOptions(firstPPr, defaultFontSize),
+        };
+        delete paraOpts._defaultFontSize;
+        if (extractBullet(firstPPr)) delete paraOpts.bullet;
+        runs.push({ text: '\n', options: { ...paraOpts }, degraded: false });
+      }
+    }
+  }
+  return compressTextRuns(runs);
+}
+
+/**
+ * @param {import('./xml-parser').OONode|null|undefined} defRPr
+ * @returns {number|undefined}
+ */
+function runFontSizeFromDefRPr(defRPr) {
+  if (!defRPr) return undefined;
+  const sz = attr(defRPr, 'sz');
+  return sz ? parseInt(sz, 10) / 100 : undefined;
+}
+
+/**
+ * @param {object|null|undefined} txBody
+ */
+function extractBodyPrParaOptions(txBody) {
+  if (!txBody) return {};
+  const bodyPr = child(txBody, 'a:bodyPr');
+  const lstStyle = child(txBody, 'a:lstStyle');
+  const defPPr = lstStyle && child(lstStyle, 'a:defPPr');
+  const defRPr = defPPr && child(defPPr, 'a:defRPr');
+  const fontSize = runFontSizeFromDefRPr(defRPr);
+  const opts = defPPr ? extractParaOptions(defPPr, fontSize) : {};
+  if (fontSize != null) opts._defaultFontSize = fontSize;
+  if (bodyPr && attr(bodyPr, 'anchor')) {
+    // bodyPr anchor 不映射到 run
+  }
+  return opts;
+}
+
+/**
+ * @param {object|null|undefined} spcParent a:spcBef | a:spcAft | a:lnSpc
+ * @param {number|undefined} fontSizePt
+ * @returns {number|undefined} 点数
+ */
+function readSpacingPt(spcParent, fontSizePt) {
+  if (!spcParent) return undefined;
+  const spcPts = child(spcParent, 'a:spcPts');
+  if (spcPts) {
+    const val = attr(spcPts, 'val');
+    return val ? parseInt(val, 10) / 100 : undefined;
+  }
+  const spcPct = child(spcParent, 'a:spcPct');
+  if (spcPct && fontSizePt) {
+    const val = attr(spcPct, 'val');
+    if (!val) return undefined;
+    return (fontSizePt * parseInt(val, 10)) / 100000;
+  }
+  return undefined;
+}
+
+/**
+ * @param {object|null|undefined} pPr
+ * @param {number|undefined} [defaultFontSizePt]
+ */
+function extractParaOptions(pPr, defaultFontSizePt) {
+  if (!pPr) return {};
+  const node = pPr;
+  const opts = {};
+
+  const algn = attr(node, 'algn');
+  if (algn && PARA_ALIGN_MAP[algn]) opts.align = PARA_ALIGN_MAP[algn];
+
+  const lvl = attr(node, 'lvl');
+  if (lvl) opts.indentLevel = parseInt(lvl, 10);
+
+  const defRPr = child(node, 'a:defRPr');
+  const fontSize = runFontSizeFromDefRPr(defRPr) ?? defaultFontSizePt;
+
+  const spcBef = readSpacingPt(child(node, 'a:spcBef'), fontSize);
+  if (spcBef != null) opts.paraSpaceBefore = spcBef;
+
+  const spcAft = readSpacingPt(child(node, 'a:spcAft'), fontSize);
+  if (spcAft != null) opts.paraSpaceAfter = spcAft;
+
+  const lnSpc = readSpacingPt(child(node, 'a:lnSpc'), fontSize);
+  if (lnSpc != null) opts.lineSpacing = lnSpc;
+
+  return opts;
+}
+
+/**
+ * @param {object|null|undefined} pPr
+ */
+function extractBullet(pPr) {
+  if (!pPr) return null;
+  if (child(pPr, 'a:buChar')) return { bullet: true };
+  if (child(pPr, 'a:buAutoNum')) return { bullet: { type: 'number' } };
+  return null;
+}
+
+/**
+ * @param {object|null|undefined} rPr
+ * @param {Record<string, string>} scheme
+ * @param {{ relIndex?: import('./rels').RelationIndex, slidePath?: string }|null} linkCtx
+ */
+function extractRunOptions(rPr, scheme, linkCtx) {
+  const options = {};
+  let degraded = false;
+  if (!rPr) return options;
+
+  const sz = attr(rPr, 'sz');
+  if (sz) options.fontSize = parseInt(sz, 10) / 100;
+  if (attr(rPr, 'b') === '1') options.bold = true;
+  if (attr(rPr, 'i') === '1') options.italic = true;
+
+  const u = attr(rPr, 'u');
+  if (u && u !== 'none') options.underline = { style: 'sng' };
+
+  const face = attr(child(rPr, 'a:latin'), 'typeface');
+  if (face) options.fontFace = face;
+
+  const solid = child(rPr, 'a:solidFill');
+  const grad = child(rPr, 'a:gradFill');
+  if (grad) {
+    degraded = true;
+    const g = resolveFillColor(grad, scheme);
+    if (g.color) options.color = g.color;
+  } else if (solid) {
+    const c = resolveColorFromContainer(solid, scheme);
+    if (c) options.color = c;
+  }
+
+  const hlink = child(rPr, 'a:hlinkClick');
+  if (hlink && linkCtx) {
+    const relId = attr(hlink, 'r:id');
+    const url = relId && linkCtx.relIndex.resolve(linkCtx.slidePath, relId);
+    if (url) options.hyperlink = { url };
+  }
+
+  if (degraded) options._degraded = true;
+  return options;
+}
+
+module.exports = {
+  extractTextRuns,
+  extractParaOptions,
+  extractRunOptions,
+  readSpacingPt,
+};

package/lib/xml-parser.js

@@ -1,25 +1,201 @@
 /**
- * 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
+ * OOXML 专用 XML 解析器（无外部依赖）
+ *
+ * 节点结构：
+ *   { tag: 'a:r', attrs: { 'r:id': 'rId1', w: '25400' }, children: [...], text: '' }
+ *
+ * 特性：
+ *   - children 保留原始 XML 子节点顺序
+ *   - children() 始终返回数组（消除 xml2js 单/多不一致）
+ *   - 空节点（如 <a:buNone/>）返回节点对象而非 '' / null（修复 xml2js 的 falsy 陷阱）
+ *   - 命名空间前缀保留，与 OOXML 原文一致
  */
-const xml2js = require('xml2js');
 
-const PARSER_OPTIONS = {
-  explicitArray: false,
-  mergeAttrs: false,
-  explicitCharkey: false,
-  tagNameProcessors: [],
-  attrNameProcessors: [],
-  xmlns: false,
-};
+/**
+ * @typedef {object} OONode
+ * @property {string} tag
+ * @property {Record<string,string>} attrs
+ * @property {OONode[]} children
+ * @property {string} text
+ */
 
 /**
  * @param {string} str
- * @returns {Promise<object>}
+ * @returns {OONode}
  */
 function parseXml(str) {
-  return xml2js.parseStringPromise(str, PARSER_OPTIONS);
+  return new Parser(str).parse();
+}
+
+class Parser {
+  constructor(input) {
+    this.s = input;
+    this.pos = 0;
+  }
+
+  parse() {
+    this.skipProlog();
+    return this.parseElement();
+  }
+
+  skipProlog() {
+    while (this.pos < this.s.length) {
+      this.skipSpaces();
+      if (this.peek('<?')) {
+        const end = this.s.indexOf('?>', this.pos);
+        this.pos = end === -1 ? this.s.length : end + 2;
+      } else if (this.peek('<!--')) {
+        const end = this.s.indexOf('-->', this.pos);
+        this.pos = end === -1 ? this.s.length : end + 3;
+      } else {
+        break;
+      }
+    }
+  }
+
+  parseElement() {
+    this.skipSpaces();
+    this.eat('<');
+
+    const tag = this.readName();
+    const rawAttrs = this.readAttributes();
+    this.skipSpaces();
+
+    // 过滤掉 xmlns 声明，暴露给调用方的 attrs 只有业务属性
+    const attrs = {};
+    for (const [k, v] of Object.entries(rawAttrs)) {
+      if (k !== 'xmlns' && !k.startsWith('xmlns:')) {
+        attrs[k] = v;
+      }
+    }
+
+    /** @type {OONode} */
+    const node = { tag, attrs, children: [], text: '' };
+
+    if (this.s[this.pos] === '/') {
+      this.pos += 2; // />
+      return node;
+    }
+    this.eat('>');
+
+    // 收集子内容
+    let textBuf = '';
+
+    const flushText = () => {
+      if (textBuf) {
+        // 只在没有子元素时，或文本非纯空白时记录
+        // 这样格式化缩进不会污染有子节点的父元素
+        const decoded = decodeEntities(textBuf);
+        if (node.children.length === 0) {
+          // 叶节点：保留全部文本（包括 \n，因为 <a:t>\n</a:t> 是有意义的）
+          node.text += decoded;
+        } else {
+          // 有子节点的父元素：只保留非纯空白文本
+          if (decoded.trim()) {
+            node.text += decoded;
+          }
+        }
+        textBuf = '';
+      }
+    };
+
+    while (this.pos < this.s.length) {
+      if (this.peek('</')) {
+        flushText();
+        this.pos += 2;
+        this.readName();
+        this.skipSpaces();
+        this.eat('>');
+        break;
+      } else if (this.peek('<!--')) {
+        const end = this.s.indexOf('-->', this.pos);
+        this.pos = end === -1 ? this.s.length : end + 3;
+      } else if (this.peek('<![CDATA[')) {
+        const end = this.s.indexOf(']]>', this.pos);
+        if (end === -1) break;
+        textBuf += this.s.slice(this.pos + 9, end);
+        this.pos = end + 3;
+      } else if (this.s[this.pos] === '<') {
+        flushText();
+        node.children.push(this.parseElement());
+      } else {
+        // 文本字符
+        textBuf += this.s[this.pos++];
+      }
+    }
+
+    return node;
+  }
+
+  readName() {
+    const start = this.pos;
+    while (this.pos < this.s.length && /[\w\-\.:]/.test(this.s[this.pos])) {
+      this.pos++;
+    }
+    return this.s.slice(start, this.pos);
+  }
+
+  readAttributes() {
+    const attrs = {};
+    while (this.pos < this.s.length) {
+      this.skipSpaces();
+      const c = this.s[this.pos];
+      if (c === '>' || c === '/' || c === '?') break;
+
+      const name = this.readName();
+      if (!name) { this.pos++; continue; }
+
+      this.skipSpaces();
+      if (this.s[this.pos] !== '=') {
+        attrs[name] = '';
+        continue;
+      }
+      this.pos++; // =
+      this.skipSpaces();
+
+      const quote = this.s[this.pos];
+      if (quote !== '"' && quote !== "'") {
+        this.pos++;
+        const start = this.pos;
+        while (this.pos < this.s.length && !/[\s>]/.test(this.s[this.pos])) this.pos++;
+        attrs[name] = decodeEntities(this.s.slice(start, this.pos));
+        continue;
+      }
+      this.pos++;
+      const start = this.pos;
+      while (this.pos < this.s.length && this.s[this.pos] !== quote) this.pos++;
+      attrs[name] = decodeEntities(this.s.slice(start, this.pos));
+      this.pos++;
+    }
+    return attrs;
+  }
+
+  skipSpaces() {
+    while (this.pos < this.s.length && /[ \t\r\n]/.test(this.s[this.pos])) this.pos++;
+  }
+
+  peek(s) {
+    return this.s.startsWith(s, this.pos);
+  }
+
+  eat(ch) {
+    if (this.s[this.pos] !== ch) {
+      const ctx = this.s.slice(Math.max(0, this.pos - 30), this.pos + 30);
+      throw new Error(`XML parse error: expected '${ch}' at pos ${this.pos}, context: ...${ctx}...`);
+    }
+    this.pos++;
+  }
+}
+
+function decodeEntities(s) {
+  return s
+    .replace(/&amp;/g,  '&')
+    .replace(/&lt;/g,   '<')
+    .replace(/&gt;/g,   '>')
+    .replace(/&quot;/g, '"')
+    .replace(/&apos;/g, "'")
+    .replace(/&#(\d+);/g,   (_, n) => String.fromCharCode(parseInt(n, 10)))
+    .replace(/&#x([0-9a-fA-F]+);/g, (_, n) => String.fromCharCode(parseInt(n, 16)));
 }
 
-module.exports = { parseXml, PARSER_OPTIONS };
+module.exports = { parseXml };

package/lib/xml-utils.js

@@ -1,68 +1,114 @@
 /**
- * xml2js 对象树访问工具（配合 lib/xml-parser.js 统一配置）
+ * OOXML 节点访问工具
+ * 配合自定义 xml-parser.js 使用
+ *
+ * 节点结构：{ tag, attrs, children: OONode[], text }
  */
 
 /**
- * @param {*} value
- * @returns {Array}
+ * 取第一个匹配 tag 的子节点
+ * @param {import('./xml-parser').OONode|null|undefined} node
+ * @param {string} tag
+ * @returns {import('./xml-parser').OONode|undefined}
  */
-function asArray(value) {
-  if (value == null) return [];
-  return Array.isArray(value) ? value : [value];
+function child(node, tag) {
+  if (!node || !node.children) return undefined;
+  return node.children.find(c => c.tag === tag);
 }
 
 /**
- * @param {object|null|undefined} node
- * @returns {Record<string, string>}
+ * 取所有匹配 tag 的子节点（始终返回数组，消除 xml2js 单/多不一致）
+ * @param {import('./xml-parser').OONode|null|undefined} node
+ * @param {string} tag
+ * @returns {import('./xml-parser').OONode[]}
  */
-function attrs(node) {
-  if (node == null) return {};
-  if (node.$) return node.$;
-  return node;
+function children(node, tag) {
+  if (!node || !node.children) return [];
+  return node.children.filter(c => c.tag === tag);
 }
 
 /**
- * @param {object|null|undefined} node
+ * 取节点的有序子节点列表（保留所有 tag，用于顺序敏感场景如 a:p）
+ * @param {import('./xml-parser').OONode|null|undefined} node
+ * @returns {import('./xml-parser').OONode[]}
+ */
+function childNodes(node) {
+  if (!node || !node.children) return [];
+  return node.children;
+}
+
+/**
+ * 取属性值
+ * @param {import('./xml-parser').OONode|null|undefined} node
  * @param {string} name
  * @returns {string|undefined}
  */
 function attr(node, name) {
-  return attrs(node)[name];
+  if (!node || !node.attrs) return undefined;
+  return node.attrs[name];
+}
+
+/**
+ * 取文本内容
+ * @param {import('./xml-parser').OONode|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);
+  return node.text ?? '';
 }
 
 /**
- * @param {object|null|undefined} node
- * @param {string} key
- * @returns {*}
+ * 兼容旧版 asArray（对迁移中未改完的调用点提供临时支持）
+ * 新代码应直接用 children(node, tag) 代替 asArray(node[tag])
+ * @param {*} value
+ * @returns {Array}
  */
-function child(node, key) {
-  if (node == null) return undefined;
-  return node[key];
+function asArray(value) {
+  if (value == null) return [];
+  return Array.isArray(value) ? value : [value];
 }
 
 /**
- * @param {object|null|undefined} node
- * @param {string[]} keys
- * @returns {*}
+ * 取第一个匹配多个 tag 之一的子节点
+ * @param {import('./xml-parser').OONode|null|undefined} node
+ * @param {string[]} tags
+ * @returns {import('./xml-parser').OONode|undefined}
  */
-function firstChild(node, keys) {
-  for (const key of keys) {
-    const v = child(node, key);
-    if (v != null) return v;
+function firstChild(node, tags) {
+  if (!node || !node.children) return undefined;
+  for (const tag of tags) {
+    const found = node.children.find(c => c.tag === tag);
+    if (found != null) return found;
   }
   return undefined;
 }
 
 /**
- * @param {object|null|undefined} node
- * @returns {string}
+ * parseXml 返回文档根 OONode；若已是目标 tag 则直接返回，否则取同名子节点
+ * @param {import('./xml-parser').OONode|null|undefined} doc
+ * @param {...string} tags
+ * @returns {import('./xml-parser').OONode|undefined}
  */
-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 '';
+function documentRoot(doc, ...tags) {
+  if (!doc) return undefined;
+  if (tags.includes(doc.tag)) return doc;
+  for (const tag of tags) {
+    const found = child(doc, tag);
+    if (found) return found;
+  }
+  return undefined;
 }
 
-module.exports = { asArray, attrs, attr, child, firstChild, textContent };
+module.exports = {
+  child,
+  children,
+  childNodes,
+  attr,
+  textContent,
+  asArray,
+  firstChild,
+  documentRoot,
+};

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pptx2js",
-  "version": "0.4.0",
+  "version": "0.4.3",
   "description": "将 .pptx 文件转换为可运行的 PptxGenJS 生成脚本",
   "main": "lib/index.js",
   "bin": {
@@ -8,7 +8,8 @@
   },
   "scripts": {
     "test": "jest",
-    "test:watch": "jest --watch"
+    "test:watch": "jest --watch",
+    "test:sample": "npm run generate --prefix testPPT && node bin/pptx2js.js testppt/sample.pptx -o testppt/pptx2js-output"
   },
   "keywords": [
     "pptx",
@@ -27,8 +28,7 @@
   "dependencies": {
     "chalk": "^4.1.2",
     "commander": "^12.1.0",
-    "jszip": "^3.10.1",
-    "xml2js": "^0.6.2"
+    "jszip": "^3.10.1"
   },
   "devDependencies": {
     "jest": "^29.7.0"