hanhan
6e1d2c6728
- 新增 /editor135 子页面：上传幼儿园食谱docx → 解析并填充135模板 → 预览/源代码/解析结果三视图 - 新增 /api/editor135/render 解析接口（jsdom 解析填充，逐条回读校验 mismatch） - 新增 /api/editor135/img 微信CDN图片代理（解决预览防盗链，仅允许 *.qpic.cn） - 缺失星期自动删除：docx 缺某天（含周末/间隔天）时整段移除该天内容，尾部三审三校/温馨提示保留 - 逐节点精确区间删除算法：处理相邻天共用容器、嵌套结构，不误删保留天 - 新增依赖 jsdom + @types/jsdom；附 6 场景回归测试脚本 scripts/test-remove-days.ts - 首页 header 增加「135编辑器转换」导航入口
614 lines
27 KiB
TypeScript
614 lines
27 KiB
TypeScript
/**
|
||
* recipe-core.ts — 尚城幼儿园食谱转换核心库
|
||
*
|
||
* 职责:
|
||
* 1. parseDocxToRecipe: docx(mammoth) → 结构化食谱 JSON(支持常规宽表/转置宽表/长表/分段文本)
|
||
* 2. fillTemplate: 食谱 JSON + 135编辑器模板HTML → 按原样式填充后的完整 HTML
|
||
* (只改文本,不改任何 style/图片/结构;条目不足克隆模板条目,多余删除)
|
||
*
|
||
* 逻辑移植自已验证的 tools/parse_recipe.mjs 与 tools/fill_135editor.mjs。
|
||
*/
|
||
import { JSDOM } from "jsdom";
|
||
|
||
// ---------- 类型 ----------
|
||
export interface RecipeItem { zh: string; en: string }
|
||
export interface RecipeMeal { meal: string; items: RecipeItem[] }
|
||
export interface RecipeDay { day: string; meals: RecipeMeal[] }
|
||
export interface Recipe { days: RecipeDay[] }
|
||
|
||
export interface MealFillInfo {
|
||
day: string; meal: string;
|
||
replaced: number; added: number; removed: number;
|
||
oldCount: number; newCount: number;
|
||
}
|
||
export interface FillReport {
|
||
filled: MealFillInfo[];
|
||
warnings: string[];
|
||
repaired: { day: string; meal: string; text: string }[];
|
||
mismatch: string[];
|
||
/** 模板中存在、但 docx 中没有的星期 → 已从输出中整段移除 */
|
||
removedDays: string[];
|
||
verification: {
|
||
stylesAdded: number; stylesRemoved: number;
|
||
styleSampleAdded: [string, number][]; styleSampleRemoved: [string, number][];
|
||
imgsUnchanged: boolean; sections: string; elements: string;
|
||
};
|
||
}
|
||
|
||
// ---------- docx 解析 ----------
|
||
const DAY_RE = /^星期[一二三四五六日]$/;
|
||
const DAY_ALIASES: Record<string, string> = {
|
||
周一: "星期一", 周二: "星期二", 周三: "星期三", 周四: "星期四",
|
||
周五: "星期五", 周六: "星期六", 周日: "星期日", 周天: "星期日",
|
||
礼拜一: "星期一", 礼拜二: "星期二", 礼拜三: "星期三", 礼拜四: "星期四",
|
||
礼拜五: "星期五", 礼拜六: "星期六", 礼拜日: "星期日", 礼拜天: "星期日",
|
||
};
|
||
const MEAL_ALIASES: Record<string, string> = {
|
||
早餐: "早餐", 早点: "早点", 水果: "早点", 果点: "早点",
|
||
午餐: "午餐", 中餐: "午餐", 午饭: "午餐",
|
||
体弱儿餐: "体弱儿餐", 体弱儿: "体弱儿餐", 特殊餐: "体弱儿餐",
|
||
午点: "午点", 午睡后: "午点", 点心: "午点", 下午点心: "午点", 下午茶: "午点",
|
||
晚餐: "晚餐", 晚点: "晚点",
|
||
};
|
||
const SEP_RE = /[;；\n、]+/;
|
||
const CJK_RE = /[\u4e00-\u9fff]/;
|
||
const LATIN_RE = /[A-Za-z]/;
|
||
|
||
const norm = (s: unknown) => String(s ?? "").replace(/\u00a0/g, " ").replace(/\s+/g, " ").trim();
|
||
const normKey = (s: unknown) => String(s ?? "").replace(/\s+/g, "");
|
||
|
||
function canonDay(s: unknown): string | null {
|
||
const t = normKey(s);
|
||
const m = t.match(/星期[一二三四五六日]/);
|
||
if (m) return m[0];
|
||
if (DAY_RE.test(t)) return t;
|
||
const w = t.match(/周[一二三四五六日天]/);
|
||
if (w) return w[0] === "周天" ? "星期日" : w[0].replace("周", "星期");
|
||
return DAY_ALIASES[t] ?? null;
|
||
}
|
||
function canonMeal(s: unknown): string | null {
|
||
const t = normKey(s);
|
||
if (MEAL_ALIASES[t]) return MEAL_ALIASES[t];
|
||
// 关键词按长短优先, 避免 "早点/水果餐" 被 "早" 误判为早餐
|
||
if (t.includes("体弱")) return "体弱儿餐";
|
||
if (t.includes("早点") || t.includes("水果")) return "早点";
|
||
if (t.includes("早餐")) return "早餐";
|
||
if (t.includes("午点") || t.includes("点心")) return "午点";
|
||
if (t.includes("午餐") || t.includes("中餐")) return "午餐";
|
||
if (t.includes("晚点")) return "晚点";
|
||
if (t.includes("晚餐")) return "晚餐";
|
||
if (t.includes("早")) return "早餐";
|
||
if (t.includes("午")) return "午餐";
|
||
if (t.includes("晚")) return "晚餐";
|
||
return null;
|
||
}
|
||
|
||
function ensureDay(rec: Recipe, day: string): RecipeDay {
|
||
let d = rec.days.find((x) => x.day === day);
|
||
if (!d) { d = { day, meals: [] }; rec.days.push(d); }
|
||
return d;
|
||
}
|
||
function ensureMeal(d: RecipeDay, meal: string): RecipeMeal {
|
||
let m = d.meals.find((x) => x.meal === meal);
|
||
if (!m) { m = { meal, items: [] }; d.meals.push(m); }
|
||
return m;
|
||
}
|
||
function addItem(m: RecipeMeal, zh: string, en: string) {
|
||
const z = norm(zh), e = norm(en);
|
||
if (!z && !e) return;
|
||
m.items.push({ zh: z, en: e });
|
||
}
|
||
/** 格子→菜品列表(中英交替行: 中文行=菜名, 紧跟英文行=其英文名)。必须接收原始文本(含换行)。 */
|
||
function splitCell(cell: string): { zhList: string[]; enList: string[] } {
|
||
const raw = String(cell ?? "").replace(/\r/g, "");
|
||
const parts = raw.split(SEP_RE).map(norm).filter(Boolean);
|
||
const items: { zh: string; en: string }[] = [];
|
||
for (const p of parts) {
|
||
if (CJK_RE.test(p)) items.push({ zh: p, en: "" });
|
||
else if (LATIN_RE.test(p)) {
|
||
if (items.length) items[items.length - 1].en = p;
|
||
else items.push({ zh: "", en: p });
|
||
} else items.push({ zh: p, en: "" });
|
||
}
|
||
if (!items.some((x) => x.zh)) return { zhList: [], enList: items.map((x) => x.en) };
|
||
return { zhList: items.map((x) => x.zh).filter(Boolean), enList: items.map((x) => x.en).filter(Boolean) };
|
||
}
|
||
|
||
/** B1 常规宽表: 表头含≥2个星期列, 第一列是餐次 */
|
||
function parseWideB1(rows: string[][]): Recipe | null {
|
||
const rec: Recipe = { days: [] };
|
||
let header: string[] | null = null;
|
||
let dayCols: (string | null)[] = [];
|
||
for (const row of rows) {
|
||
const days = row.map(normKey).map(canonDay);
|
||
if (days.filter(Boolean).length >= 2) { header = row; dayCols = days; break; }
|
||
}
|
||
if (!header) return null;
|
||
let curMeal: string | null = null;
|
||
for (const row of rows) {
|
||
const mealKey = canonMeal(row[0] ?? "");
|
||
if (mealKey) curMeal = mealKey;
|
||
if (!curMeal) continue;
|
||
for (let ci = 0; ci < header.length; ci++) {
|
||
const day = dayCols[ci];
|
||
if (!day || !(row[ci] ?? "").trim()) continue;
|
||
const { zhList, enList } = splitCell(row[ci]);
|
||
if (!zhList.length && !enList.length) continue;
|
||
const m = ensureMeal(ensureDay(rec, day), curMeal);
|
||
const n = Math.max(zhList.length, enList.length);
|
||
for (let i = 0; i < n; i++) addItem(m, zhList[i] ?? "", enList[i] ?? "");
|
||
}
|
||
}
|
||
return rec.days.length ? rec : null;
|
||
}
|
||
|
||
/** B2 转置宽表: 星期在第一列, 餐次在表头行(用户真实文件即此格式) */
|
||
function parseWideB2(rows: string[][]): Recipe | null {
|
||
const rec: Recipe = { days: [] };
|
||
for (let ri = 0; ri < rows.length; ri++) {
|
||
const cells = rows[ri].map(normKey);
|
||
const mealCols = cells
|
||
.map((c, i) => (canonMeal(c) && i > 0 ? { meal: canonMeal(c)!, col: i } : null))
|
||
.filter((x): x is { meal: string; col: number } => !!x);
|
||
if (mealCols.length < 2) continue;
|
||
for (let rj = ri + 1; rj < rows.length; rj++) {
|
||
const day = canonDay(rows[rj][0] ?? "");
|
||
if (!day) continue;
|
||
for (const { meal, col } of mealCols) {
|
||
if (!(rows[rj][col] ?? "").trim()) continue;
|
||
const { zhList, enList } = splitCell(rows[rj][col]);
|
||
if (!zhList.length && !enList.length) continue;
|
||
const m = ensureMeal(ensureDay(rec, day), meal);
|
||
const n = Math.max(zhList.length, enList.length);
|
||
for (let i = 0; i < n; i++) addItem(m, zhList[i] ?? "", enList[i] ?? "");
|
||
}
|
||
}
|
||
if (rec.days.length) return rec;
|
||
}
|
||
return null;
|
||
}
|
||
|
||
/** 长表: 列 = 星期|餐次|菜品|英文, 空星期/餐次沿用上一行(兼容合并单元格) */
|
||
function parseLongRows(rows: string[][]): Recipe {
|
||
const rec: Recipe = { days: [] };
|
||
let curDay: string | null = null, curMeal: string | null = null;
|
||
for (const row of rows) {
|
||
if (!row.some((c) => norm(c))) continue;
|
||
let dayIdx = -1, mealIdx = -1;
|
||
for (let i = 0; i < row.length; i++) {
|
||
if (canonDay(row[i]) && dayIdx < 0) dayIdx = i;
|
||
if (canonMeal(row[i]) && mealIdx < 0 && i !== dayIdx) mealIdx = i;
|
||
}
|
||
if (dayIdx >= 0) curDay = canonDay(row[dayIdx]);
|
||
if (mealIdx >= 0) curMeal = canonMeal(row[mealIdx]);
|
||
if (!curDay || !curMeal) continue;
|
||
const zhCells: string[] = [], enCells: string[] = [];
|
||
for (let i = 0; i < row.length; i++) {
|
||
if (i === dayIdx || i === mealIdx || !norm(row[i])) continue;
|
||
(CJK_RE.test(row[i]) ? zhCells : enCells).push(row[i]);
|
||
}
|
||
const m = ensureMeal(ensureDay(rec, curDay), curMeal);
|
||
const zhParts = zhCells.flatMap((c) => splitCell(c).zhList);
|
||
let enParts = enCells.map(norm).filter(Boolean);
|
||
if (!enParts.length) enParts = zhCells.flatMap((c) => splitCell(c).enList);
|
||
const n = Math.max(zhParts.length, enParts.length);
|
||
for (let i = 0; i < n; i++) addItem(m, zhParts[i] ?? "", enParts[i] ?? "");
|
||
}
|
||
return rec;
|
||
}
|
||
|
||
/** 分段文本: 【星期一】/ 早餐:菜;菜 */
|
||
function parseTextLines(lines: string[]): Recipe {
|
||
const rec: Recipe = { days: [] };
|
||
let curDay: string | null = null, curMeal: string | null = null;
|
||
for (const raw0 of lines) {
|
||
const raw = raw0.replace(/\r/g, "").trim();
|
||
if (!raw) continue;
|
||
const dayM = raw.match(/^[【[]?\s*(星期[一二三四五六日])\s*[】\]]?$/);
|
||
if (dayM) { curDay = dayM[1]; curMeal = null; continue; }
|
||
const mealM = raw.match(/^(早餐|早点|水果|午餐|中餐|体弱儿餐|午点|点心|晚餐|晚点)\s*[:：]\s*(.*)$/);
|
||
if (mealM) {
|
||
curMeal = canonMeal(mealM[1]);
|
||
if (curDay && curMeal) {
|
||
const m = ensureMeal(ensureDay(rec, curDay), curMeal);
|
||
const { zhList, enList } = splitCell(mealM[2]);
|
||
const n = Math.max(zhList.length, enList.length);
|
||
for (let i = 0; i < n; i++) addItem(m, zhList[i] ?? "", enList[i] ?? "");
|
||
}
|
||
continue;
|
||
}
|
||
if (curDay && curMeal && (CJK_RE.test(raw) || LATIN_RE.test(raw))) {
|
||
const m = ensureMeal(ensureDay(rec, curDay), curMeal);
|
||
const { zhList, enList } = splitCell(raw);
|
||
const n = Math.max(zhList.length, enList.length);
|
||
for (let i = 0; i < n; i++) addItem(m, zhList[i] ?? "", enList[i] ?? "");
|
||
}
|
||
}
|
||
return rec;
|
||
}
|
||
|
||
/** mammoth HTML → 表格行(单元格内多个<p>按行还原) */
|
||
function htmlToRows(html: string): string[][] {
|
||
const rows: string[][] = [];
|
||
const tableRe = /<tr[^>]*>([\s\S]*?)<\/tr>/g;
|
||
let mt: RegExpExecArray | null;
|
||
while ((mt = tableRe.exec(html))) {
|
||
const cells = [...mt[1].matchAll(/<t[dh][^>]*>([\s\S]*?)<\/t[dh]>/g)].map((c) =>
|
||
(c[1] ?? "")
|
||
.replace(/<br\s*\/?>/gi, "\n")
|
||
.replace(/<\/p>/gi, "\n")
|
||
.replace(/<[^>]+>/g, "")
|
||
.replace(/&/g, "&").replace(/</g, "<").replace(/>/g, ">").replace(/ /g, " "),
|
||
);
|
||
rows.push(cells);
|
||
}
|
||
return rows;
|
||
}
|
||
|
||
export async function parseDocxToRecipe(buf: Buffer): Promise<Recipe> {
|
||
const mammoth = await import("mammoth");
|
||
const { value: html } = await mammoth.convertToHtml({ buffer: buf });
|
||
const rows = htmlToRows(html);
|
||
if (rows.length) {
|
||
const rec = parseWideB1(rows) ?? parseWideB2(rows) ?? parseLongRows(rows);
|
||
if (rec.days.length) return rec;
|
||
}
|
||
return parseTextLines(html.replace(/<[^>]+>/g, "\n").split("\n"));
|
||
}
|
||
|
||
// ---------- 模板填充(逻辑与 fill_135editor.mjs 页面脚本一致, DOM 换成 jsdom) ----------
|
||
const MEAL_KEY: Record<string, string> = {
|
||
早餐: "breakfast", 早点: "fruit", 水果: "fruit",
|
||
午餐: "lunch", 中餐: "lunch",
|
||
体弱儿餐: "special",
|
||
午点: "snack", 点心: "snack",
|
||
晚餐: "dinner", 晚点: "dinner",
|
||
};
|
||
const MEAL_KEY_TO_LABEL: Record<string, string> = {
|
||
breakfast: "早餐", fruit: "早点", lunch: "午餐", special: "体弱儿餐", snack: "午点", dinner: "晚餐",
|
||
};
|
||
const DAY_SET = new Set(["星期一", "星期二", "星期三", "星期四", "星期五", "星期六", "星期日"]);
|
||
|
||
interface El {
|
||
tagName: string;
|
||
children: HTMLElement[];
|
||
childNodes: ChildNode[];
|
||
textContent: string;
|
||
parentElement: HTMLElement | null;
|
||
nextElementSibling: HTMLElement | null;
|
||
getAttribute(n: string): string | null;
|
||
cloneNode(deep: boolean): any;
|
||
querySelectorAll(s: string): any[];
|
||
remove(): void;
|
||
insertBefore(n: Node, r: Node | null): Node;
|
||
parentNode: HTMLElement | null;
|
||
}
|
||
|
||
interface Item {
|
||
zhEl: HTMLElement; zhCont: HTMLElement[];
|
||
enEl: HTMLElement | null; enCont: HTMLElement[];
|
||
zhBlock: HTMLElement | null; enBlock: HTMLElement | null;
|
||
ca: HTMLElement | null; valid: boolean;
|
||
}
|
||
interface MealCol { key: string; items: Item[] }
|
||
interface DayCol { day: string; meals: MealCol[]; titleEl: HTMLElement | null }
|
||
|
||
const leafText = (el: HTMLElement) => (el.textContent ?? "").replace(/\u00a0/g, " ").trim();
|
||
|
||
function collect(body: HTMLElement): DayCol[] {
|
||
const days: DayCol[] = [];
|
||
let curDay: DayCol | null = null, curMeal: MealCol | null = null, curItem: Item | null = null, lastKind: string | null = null;
|
||
const getDay = (t: string, el?: HTMLElement) => { let d = days.find((x) => x.day === t); if (!d) { d = { day: t, meals: [], titleEl: el ?? null }; days.push(d); } if (el && !d.titleEl) d.titleEl = el; return d; };
|
||
const getMeal = (d: DayCol, key: string) => { let m = d.meals.find((x) => x.key === key); if (!m) { m = { key, items: [] }; d.meals.push(m); } return m; };
|
||
|
||
for (const el of body.querySelectorAll("*") as unknown as HTMLElement[]) {
|
||
if (el.children.length) continue;
|
||
const t = leafText(el);
|
||
if (!t) continue;
|
||
if (DAY_SET.has(t)) { curDay = getDay(t, el); curMeal = null; curItem = null; lastKind = null; continue; }
|
||
const mk = MEAL_KEY[t];
|
||
if (mk) {
|
||
if (!curDay) continue;
|
||
curMeal = getMeal(curDay, mk); curItem = null; lastKind = null; continue;
|
||
}
|
||
if (!curDay || !curMeal) continue;
|
||
if (CJK_RE.test(t)) {
|
||
if (lastKind === "zh" && curItem) curItem.zhCont.push(el);
|
||
else { curItem = { zhEl: el, zhCont: [], enEl: null, enCont: [], zhBlock: null, enBlock: null, ca: null, valid: false }; curMeal.items.push(curItem); lastKind = "zh"; }
|
||
} else if (LATIN_RE.test(t)) {
|
||
if (curItem && lastKind === "zh" && !curItem.enEl) { curItem.enEl = el; lastKind = "en"; }
|
||
else if (curItem && lastKind === "en") curItem.enCont.push(el);
|
||
}
|
||
}
|
||
// 配对: 中英块的最近公共祖先下必须是相邻兄弟
|
||
const pathTo = (el: HTMLElement): HTMLElement[] => { const p: HTMLElement[] = []; let n: HTMLElement | null = el; while (n && n !== body) { p.unshift(n); n = n.parentElement; } return p; };
|
||
for (const d of days) for (const m of d.meals) for (const it of m.items) {
|
||
it.valid = false;
|
||
if (!it.enEl) continue;
|
||
const zp = pathTo(it.zhEl), ep = pathTo(it.enEl);
|
||
let ca: HTMLElement | null = null;
|
||
for (let i = 0; i < Math.min(zp.length, ep.length); i++) { if (zp[i] === ep[i]) ca = zp[i]; else break; }
|
||
if (!ca) continue;
|
||
const zb = zp[zp.indexOf(ca) + 1], eb = ep[ep.indexOf(ca) + 1];
|
||
if (!zb || !eb) continue;
|
||
if (zb.nextElementSibling !== eb) continue;
|
||
it.zhBlock = zb; it.enBlock = eb; it.ca = ca; it.valid = true;
|
||
}
|
||
return days;
|
||
}
|
||
|
||
function daysToRecipe(days: DayCol[]): Recipe {
|
||
return {
|
||
days: days.map((d) => ({
|
||
day: d.day,
|
||
meals: d.meals.map((m) => ({
|
||
meal: MEAL_KEY_TO_LABEL[m.key],
|
||
items: m.items.filter((it) => it.valid).map((it) => ({
|
||
zh: [it.zhEl, ...it.zhCont].map(leafText).join(""),
|
||
en: it.enEl && it.enCont.length ? [it.enEl, ...it.enCont].map(leafText).join("") : (it.enEl ? leafText(it.enEl) : ""),
|
||
})),
|
||
})),
|
||
})),
|
||
};
|
||
}
|
||
|
||
interface Snapshot { styles: Map<string, number>; imgs: string[]; sections: number; elements: number }
|
||
function snapshot(body: HTMLElement): Snapshot {
|
||
const styles = new Map<string, number>();
|
||
let sections = 0, elements = 0;
|
||
for (const el of body.querySelectorAll("*") as unknown as HTMLElement[]) {
|
||
elements++;
|
||
if (el.tagName === "SECTION") sections++;
|
||
const st = el.getAttribute("style") || "";
|
||
if (st) styles.set(st, (styles.get(st) ?? 0) + 1);
|
||
}
|
||
const imgs = [...body.querySelectorAll("img")].map((im: any) => im.getAttribute("src") || "");
|
||
return { styles, imgs, sections, elements };
|
||
}
|
||
function diffSnapshot(a: Snapshot, b: Snapshot) {
|
||
const extras: [string, number][] = [], missing: [string, number][] = [];
|
||
for (const [k, v] of b.styles) { const dv = v - (a.styles.get(k) ?? 0); if (dv > 0) extras.push([k, dv]); }
|
||
for (const [k, v] of a.styles) { const dv = (b.styles.get(k) ?? 0) - v; if (dv > 0) missing.push([k, dv]); }
|
||
const ia = [...a.imgs].sort().join("|"), ib = [...b.imgs].sort().join("|");
|
||
return {
|
||
stylesAdded: extras.length, stylesRemoved: missing.length,
|
||
styleSampleAdded: extras.slice(0, 3), styleSampleRemoved: missing.slice(0, 3),
|
||
imgsUnchanged: ia === ib,
|
||
sections: `${a.sections} -> ${b.sections}`, elements: `${a.elements} -> ${b.elements}`,
|
||
};
|
||
}
|
||
|
||
/** 块内文本统一设置: 兼容 元素叶子 / 块直接文本节点 / 空块;不触碰style与结构 */
|
||
function setBlockText(block: HTMLElement, txt: string, wantCJK: boolean) {
|
||
for (const n of [...block.childNodes]) if (n.nodeType === 3) n.textContent = "";
|
||
const leaves = [...block.querySelectorAll("*")].filter((el: any) => el.children.length === 0) as HTMLElement[];
|
||
if (leaves.length) {
|
||
const main = leaves.find((el) => (wantCJK ? CJK_RE.test(leafText(el)) : LATIN_RE.test(leafText(el)))) ?? leaves[0];
|
||
for (const el of leaves) if (el !== main) (el as HTMLElement).textContent = "";
|
||
main.textContent = txt;
|
||
return;
|
||
}
|
||
const dts = [...block.childNodes].filter((n) => n.nodeType === 3);
|
||
if (dts.length) { dts[0].textContent = txt; return; }
|
||
block.textContent = txt;
|
||
}
|
||
|
||
/** 归一化修复: 列表容器里的裸文本菜名包进模板自身样式块(视觉不变) */
|
||
function normalizeBareText(days: DayCol[]): { day: string; meal: string; text: string }[] {
|
||
const repaired: { day: string; meal: string; text: string }[] = [];
|
||
for (const d of days) for (const m of d.meals) {
|
||
const valid = m.items.filter((it) => it.valid);
|
||
if (!valid.length) continue;
|
||
const model = valid[valid.length - 1];
|
||
const containers = new Set(valid.map((it) => it.ca).filter(Boolean) as HTMLElement[]);
|
||
for (const ca of containers) {
|
||
for (const node of [...ca.childNodes]) {
|
||
if (node.nodeType !== 3) continue;
|
||
const t = (node.textContent ?? "").replace(/\u00a0/g, " ").trim();
|
||
if (!t || t.length > 60) continue;
|
||
const isZh = CJK_RE.test(t), isEn = LATIN_RE.test(t);
|
||
if (!isZh && !isEn) continue;
|
||
if (isZh && model.zhBlock && model.enBlock) {
|
||
const zc = model.zhBlock.cloneNode(true) as HTMLElement;
|
||
const ec = model.enBlock.cloneNode(true) as HTMLElement;
|
||
setBlockText(zc, t, true);
|
||
setBlockText(ec, "", false);
|
||
ca.insertBefore(zc, node);
|
||
(zc.parentNode as HTMLElement).insertBefore(ec, zc.nextSibling);
|
||
} else if (model.enBlock) {
|
||
const ec = model.enBlock.cloneNode(true) as HTMLElement;
|
||
setBlockText(ec, t, false);
|
||
ca.insertBefore(ec, node);
|
||
}
|
||
ca.removeChild(node);
|
||
repaired.push({ day: d.day, meal: MEAL_KEY_TO_LABEL[m.key], text: t });
|
||
}
|
||
}
|
||
}
|
||
return repaired;
|
||
}
|
||
|
||
/**
|
||
* 移除模板中存在、但 docx 食谱里没有的"天"区块(标题块+全部餐次块,含中间隔断)。
|
||
* 算法: 对相邻两天标题叶子求最近公共祖先(LCA),在该层级删除 [当天子块, 下一天子块) 区间;
|
||
* 对最后一个模板天, 以该天最后一个有效条目的英文块为终点(含端点)删除。
|
||
* 只删区块节点, 不触碰任何保留内容的 style 与结构。
|
||
*/
|
||
function removeMissingDays(body: HTMLElement, dayCols: DayCol[], recipe: Recipe): string[] {
|
||
const wanted = new Set((recipe.days ?? []).map((d) => d.day));
|
||
const removed: string[] = [];
|
||
const leaves = Array.from(body.querySelectorAll<HTMLElement>("*")).filter(
|
||
(el) => el.children.length === 0 && el.textContent?.trim(),
|
||
);
|
||
|
||
/**
|
||
* 逐节点精确删除文档序区间 [a, b) (includeB 时含 b)。
|
||
* 规则: 节点属于 a 的祖先链时不可整体删除(内部可能含 a 之前需保留的内容, 如相邻天共用容器),
|
||
* 只前进到其 nextSibling(必然位于 a 之后, 可删)或爬升;
|
||
* 其余节点整体位于 a 之后, 直接删除, 无 nextSibling 时爬升(父节点按祖先链判断再决定);
|
||
* 前进进入的节点若包含 b 则下钻。保证不误删保留天。
|
||
*/
|
||
const removeBetween = (a: HTMLElement, b: HTMLElement | null, includeB: boolean) => {
|
||
if (!a || !a.parentNode || (b && a.contains(b))) return;
|
||
const ancestors = new Set<Node>();
|
||
for (let n: Node | null = a.parentElement; n; n = n.parentElement) ancestors.add(n);
|
||
let cur: Node | null = a;
|
||
let guard = 0;
|
||
while (cur && guard++ < 500000) {
|
||
if (cur === body) return;
|
||
if (b && cur === b) {
|
||
if (includeB) cur.parentNode?.removeChild(cur);
|
||
return;
|
||
}
|
||
const el = cur as HTMLElement;
|
||
const isAncestor = ancestors.has(cur);
|
||
if (!isAncestor && b && typeof el.contains === "function" && el.contains(b)) {
|
||
cur = el.firstChild; // 整体位于 a 之后且含 b: 下钻
|
||
continue;
|
||
}
|
||
if (isAncestor) {
|
||
cur = cur.nextSibling ?? cur.parentNode; // 祖先不删: 前进(在 a 之后)或爬升
|
||
continue;
|
||
}
|
||
const nx = cur.nextSibling;
|
||
const parent = cur.parentNode;
|
||
cur.parentNode?.removeChild(cur);
|
||
cur = nx ?? parent;
|
||
}
|
||
};
|
||
|
||
for (let i = 0; i < dayCols.length; i++) {
|
||
const d = dayCols[i];
|
||
if (wanted.has(d.day)) continue;
|
||
const titleEl = d.titleEl;
|
||
if (!titleEl || !body.contains(titleEl)) continue;
|
||
|
||
// 区间终点: 优先用下一个模板天的标题(排除); 最后一个天用尾部标记(排除), 兜底用最后条目的英文块(含端点)
|
||
const nextTitle = dayCols[i + 1]?.titleEl ?? null;
|
||
let endLeaf: HTMLElement | null = nextTitle && nextTitle !== titleEl && body.contains(nextTitle) ? nextTitle : null;
|
||
let includeEnd = false;
|
||
if (!endLeaf) {
|
||
const tail = leaves.find((el) => {
|
||
const t = el.textContent!.trim();
|
||
return (t === "三审三校" || t.startsWith("温馨提示")) && el !== titleEl && !titleEl.contains(el);
|
||
});
|
||
if (tail && body.contains(tail)) {
|
||
endLeaf = tail;
|
||
includeEnd = false;
|
||
} else {
|
||
const items = d.meals.flatMap((m) => m.items.filter((it) => it.valid));
|
||
endLeaf = items.length ? (items[items.length - 1].enBlock ?? null) : null;
|
||
includeEnd = !!endLeaf;
|
||
}
|
||
}
|
||
removeBetween(titleEl, endLeaf, includeEnd);
|
||
removed.push(d.day);
|
||
}
|
||
return removed;
|
||
}
|
||
|
||
export function fillTemplate(templateHtml: string, recipe: Recipe): { html: string; report: FillReport } {
|
||
const dom = new JSDOM(`<html><body>${templateHtml}</body></html>`, { url: "https://www.135editor.com/" });
|
||
const body = dom.window.document.body as unknown as HTMLElement;
|
||
|
||
const days = collect(body);
|
||
const repaired = normalizeBareText(days);
|
||
const dayCols = repaired.length ? collect(body) : days;
|
||
// 先移除 docx 中缺失的星期(如周五缺席则整段删除周五), 再重新收集(删除后节点引用会失效)并做基线快照与填充
|
||
const removedDays = removeMissingDays(body, dayCols, recipe);
|
||
const templateDayNames = new Set(dayCols.map((d) => d.day));
|
||
const fillCols = removedDays.length ? collect(body) : dayCols;
|
||
const baseline = snapshot(body);
|
||
|
||
const warnings: string[] = [];
|
||
const filled: MealFillInfo[] = [];
|
||
const joinLeaves = (els: HTMLElement[]) => els.map(leafText).join("");
|
||
|
||
for (const rd of recipe.days ?? []) {
|
||
const d = fillCols.find((x) => x.day === rd.day);
|
||
if (!d) { warnings.push(`编辑器模板中找不到 ${rd.day}`); continue; }
|
||
for (const rm of rd.meals ?? []) {
|
||
const mk = MEAL_KEY[rm.meal];
|
||
if (!mk) { warnings.push(`未知餐次: ${rm.meal}`); continue; }
|
||
const m = d.meals.find((x) => x.key === mk);
|
||
if (!m) { warnings.push(`${rd.day} 模板中无 ${rm.meal}`); continue; }
|
||
const oldItems = m.items.filter((it) => it.valid);
|
||
if (!oldItems.length) { warnings.push(`${rd.day} ${rm.meal} 模板无可克隆条目`); continue; }
|
||
const newItems = rm.items ?? [];
|
||
let added = 0, removed = 0, replaced = 0;
|
||
// 1) 就地替换(文本未变不碰叶子, 保持字节分布)
|
||
for (let i = 0; i < Math.min(oldItems.length, newItems.length); i++) {
|
||
const it = oldItems[i], ni = newItems[i];
|
||
if (joinLeaves([it.zhEl, ...it.zhCont]) !== (ni.zh ?? "")) {
|
||
it.zhEl.textContent = ni.zh ?? "";
|
||
for (const el of it.zhCont) el.textContent = "";
|
||
}
|
||
const enNow = it.enEl ? joinLeaves([it.enEl, ...it.enCont]) : "";
|
||
if (enNow !== (ni.en ?? "")) {
|
||
if (it.enEl) it.enEl.textContent = ni.en ?? "";
|
||
for (const el of it.enCont) el.textContent = "";
|
||
}
|
||
replaced++;
|
||
}
|
||
// 2) 锚点递进克隆补足
|
||
const model = oldItems[oldItems.length - 1];
|
||
let anchorEn = model.enBlock!;
|
||
for (let i = oldItems.length; i < newItems.length; i++) {
|
||
const zc = model.zhBlock!.cloneNode(true) as HTMLElement;
|
||
const ec = model.enBlock!.cloneNode(true) as HTMLElement;
|
||
const parent = anchorEn.parentNode!;
|
||
parent.insertBefore(zc, anchorEn.nextSibling);
|
||
parent.insertBefore(ec, zc.nextSibling);
|
||
setBlockText(zc, newItems[i].zh ?? "", true);
|
||
setBlockText(ec, newItems[i].en ?? "", false);
|
||
anchorEn = ec;
|
||
added++;
|
||
}
|
||
// 3) 从后往前删除多余
|
||
for (let i = oldItems.length - 1; i >= newItems.length; i--) {
|
||
const it = oldItems[i];
|
||
it.zhBlock?.remove();
|
||
it.enBlock?.remove();
|
||
for (const el of [...it.zhCont, ...it.enCont]) el.textContent = "";
|
||
removed++;
|
||
}
|
||
filled.push({ day: rd.day, meal: rm.meal, replaced, added, removed, oldCount: oldItems.length, newCount: newItems.length });
|
||
}
|
||
}
|
||
|
||
// 校验: 重新收集逐项比对
|
||
const after = snapshot(body);
|
||
const recipe2 = daysToRecipe(collect(body));
|
||
const mismatch: string[] = [];
|
||
for (const rd of recipe.days ?? []) {
|
||
if (!templateDayNames.has(rd.day)) continue; // 模板本来就没有该天(如周末), 已在 warnings 提示, 不算 mismatch
|
||
const d2 = recipe2.days.find((x) => x.day === rd.day);
|
||
if (!d2) { mismatch.push(`${rd.day} 未收集到`); continue; }
|
||
for (const rm of rd.meals ?? []) {
|
||
const m2 = d2.meals.find((x) => x.meal === rm.meal);
|
||
const exp = rm.items ?? [];
|
||
const got = m2?.items ?? [];
|
||
if (got.length !== exp.length) { mismatch.push(`${rd.day} ${rm.meal} 条数 ${exp.length}→${got.length}`); continue; }
|
||
for (let i = 0; i < exp.length; i++) {
|
||
if ((got[i].zh ?? "") !== (exp[i].zh ?? "") || (got[i].en ?? "") !== (exp[i].en ?? "")) {
|
||
mismatch.push(`${rd.day} ${rm.meal} #${i + 1} 文本不符`);
|
||
break;
|
||
}
|
||
}
|
||
}
|
||
}
|
||
|
||
const html = (body as any).innerHTML as string;
|
||
return {
|
||
html,
|
||
report: { filled, warnings, repaired, mismatch, removedDays, verification: diffSnapshot(baseline, after) },
|
||
};
|
||
}
|
||
|
||
/** 提取模板当前食谱(可用于基线对照/测试) */
|
||
export function extractTemplateRecipe(templateHtml: string): Recipe {
|
||
const dom = new JSDOM(`<html><body>${templateHtml}</body></html>`, { url: "https://www.135editor.com/" });
|
||
const body = dom.window.document.body as unknown as HTMLElement;
|
||
return daysToRecipe(collect(body));
|
||
}
|