feat: add local hexagram content for offline reading

Load a versioned Wikisource jing plus project-authored plain drafts so results can show labeled original and vernacular texts without unauthorized modern translations. Co-authored-by: Cursor <cursoragent@cursor.com>
2026-08-19 11:48:16 +08:00
parent 4235cf9011
commit 534c88993e
33 changed files with 5626 additions and 29 deletions
@@ -0,0 +1,85 @@
+import { readFile, writeFile, mkdir } from "node:fs/promises";
+import path from "node:path";
+import { kingWenEntries, validateContentPackage } from "./content-contract.mjs";
+import { repositoryRoot } from "./repository-files.mjs";
+import { lingjiPlain } from "../content/plain/lingji-plain.mjs";
+
+const names = [
+  "乾", "坤", "屯", "蒙", "需", "讼", "师", "比",
+  "小畜", "履", "泰", "否", "同人", "大有", "谦", "豫",
+  "随", "蛊", "临", "观", "噬嗑", "贲", "剥", "复",
+  "无妄", "大畜", "颐", "大过", "坎", "离", "咸", "恒",
+  "遁", "大壮", "晋", "明夷", "家人", "睽", "蹇", "解",
+  "损", "益", "夬", "姤", "萃", "升", "困", "井",
+  "革", "鼎", "震", "艮", "渐", "归妹", "丰", "旅",
+  "巽", "兑", "涣", "节", "中孚", "小过", "既济", "未济",
+];
+
+if (lingjiPlain.length !== 64) {
+  throw new Error(`lingji plain table must contain 64 hexagrams, got ${lingjiPlain.length}`);
+}
+
+const jingPath = path.join(repositoryRoot, "content", "raw", "zhouyi-wikisource-jing.json");
+const jing = JSON.parse(await readFile(jingPath, "utf8"));
+if (!Array.isArray(jing) || jing.length !== 64) {
+  throw new Error("Wikisource jing import must contain 64 hexagrams; run scripts/import-zhouyi-wikisource.mjs");
+}
+
+const jingById = new Map(jing.map((item) => [item.kingWenNumber, item]));
+const entries = kingWenEntries();
+const originalSource = "wikisource-zhouyi-jing-zh-Hans";
+const plainSource = "lingji-plain-zh-Hans-2026.1";
+
+const contentPackage = {
+  schemaVersion: 1,
+  contentVersion: "zh-Hans-2026.1",
+  specialUsageTexts: { qian: true, kun: true },
+  sources: [
+    {
+      id: originalSource,
+      title: "维基文库《周易》易经",
+      edition: "仅卦辞、爻辞与乾用九、坤用六，不含十翼；由 zh.wikisource.org 导入并经 Wikimedia zh-hans 转换",
+      license: "公版（Wikimedia PD-old）",
+      url: "https://zh.wikisource.org/wiki/周易",
+    },
+    {
+      id: plainSource,
+      title: "灵机本地白话",
+      edition: "2026.1 项目自撰草稿，待内容负责人审校后视为可再分发",
+      license: "项目自撰；审校完成前仅用于本机构建，不以现代译注冒充公版",
+      url: "https://creativecommons.org/licenses/by/4.0/deed.zh-Hans",
+    },
+  ],
+  hexagrams: entries.map((entry) => {
+    const original = jingById.get(entry.kingWenNumber);
+    if (!original) {
+      throw new Error(`Missing jing text for King Wen ${entry.kingWenNumber}`);
+    }
+    const plain = lingjiPlain[entry.kingWenNumber - 1];
+    if (!Array.isArray(plain) || plain.length !== 7 || plain.some((text) => !String(text).trim())) {
+      throw new Error(`Plain text for King Wen ${entry.kingWenNumber} must contain 7 non-blank strings`);
+    }
+    return {
+      ...entry,
+      name: names[entry.kingWenNumber - 1],
+      symbol: String.fromCodePoint(0x4DC0 + entry.kingWenNumber - 1),
+      judgmentOriginal: original.judgmentOriginal,
+      judgmentPlain: plain[0],
+      lineTextsBottomUp: original.lineTextsBottomUp,
+      linePlainBottomUp: plain.slice(1),
+      specialUsageText: original.specialUsageText ?? null,
+      sourceRefs: [originalSource, plainSource],
+    };
+  }),
+};
+
+const errors = validateContentPackage(contentPackage);
+if (errors.length > 0) {
+  throw new Error(`Built content package failed validation:\n${errors.join("\n")}`);
+}
+
+const outputDirectory = path.join(repositoryRoot, "content", "packages");
+await mkdir(outputDirectory, { recursive: true });
+const outputPath = path.join(outputDirectory, "hexagram-content.json");
+await writeFile(outputPath, `${JSON.stringify(contentPackage, null, 2)}\n`);
+process.stdout.write(`wrote ${outputPath}\n`);
@@ -0,0 +1,221 @@
+import { execFile } from "node:child_process";
+import { mkdir, readFile, writeFile } from "node:fs/promises";
+import path from "node:path";
+import { promisify } from "node:util";
+import { kingWenEntries } from "./content-contract.mjs";
+import { repositoryRoot } from "./repository-files.mjs";
+
+const execFileAsync = promisify(execFile);
+const proxyUrl = process.env.HTTPS_PROXY || process.env.HTTP_PROXY || "http://127.0.0.1:1080";
+const limit = Number.parseInt(process.env.IMPORT_LIMIT ?? "64", 10);
+const userAgent = "LingjiContentImport/1.0 (local Zhouyi pipeline; public-domain jingwen only)";
+const outputDirectory = path.join(repositoryRoot, "content", "raw");
+const outputPath = path.join(outputDirectory, "zhouyi-wikisource-jing.json");
+
+const titles = [
+  "乾", "坤", "屯", "蒙", "需", "訟", "師", "比",
+  "小畜", "履", "泰", "否", "同人", "大有", "謙", "豫",
+  "隨", "蠱", "臨", "觀", "噬嗑", "賁", "剝", "復",
+  "无妄", "大畜", "頤", "大過", "坎", "離", "咸", "恆",
+  "遯", "大壯", "晉", "明夷", "家人", "睽", "蹇", "解",
+  "損", "益", "夬", "姤", "萃", "升", "困", "井",
+  "革", "鼎", "震", "艮", "漸", "歸妹", "豐", "旅",
+  "巽", "兌", "渙", "節", "中孚", "小過", "既濟", "未濟",
+];
+
+const linePattern = /^(初九|九二|九三|九四|九五|上九|初六|六二|六三|六四|六五|上六|用九|用六)[：:，,](.*)$/u;
+
+function stripMarkup(wikitext) {
+  return wikitext
+    .replace(/-\{([^}|]+)-\}/gu, "$1")
+    .replace(/\{\{[^}]*\}\}/gu, "")
+    .replace(/<[^>]+>/gu, "")
+    .replace(/'{2,}/gu, "")
+    .replace(/\[\[File:[^\]]*\]\]/gu, "")
+    .replace(/\[\[(?:[^\|\]]*\|)?([^\]]+)\]\]/gu, "$1")
+    .replace(/&nbsp;/gu, " ")
+    .replace(/\r/gu, "");
+}
+
+function parseClassicFields(section) {
+  const rawLines = section
+    .split("\n")
+    .map((line) => line.replace(/^[ *#:]+/u, "").trim())
+    .filter(Boolean);
+
+  let judgment = "";
+  const lines = [];
+  let special = null;
+  for (const rawLine of rawLines) {
+    const compact = rawLine.replace(/\s+/gu, "");
+    const hit = linePattern.exec(compact) ?? linePattern.exec(rawLine);
+    if (hit) {
+      const label = hit[1];
+      const text = `${label}：${hit[2].trim()}`;
+      if (label === "用九" || label === "用六") {
+        special = text;
+      } else {
+        lines.push(text);
+      }
+    } else if (lines.length === 0 && special == null) {
+      const fragment = rawLine.replace(/\s+/gu, "");
+      judgment = judgment ? `${judgment}${fragment}` : fragment;
+    }
+  }
+
+  judgment = judgment
+    .replace(/^(?:周易)?[\u4e00-\u9fff]{1,3}[：:]/u, "")
+    .trim();
+  if (!judgment) throw new Error("missing judgment");
+  if (lines.length !== 6) {
+    throw new Error(`expected 6 lines, got ${lines.length}: ${lines.join(" | ")}`);
+  }
+  return { judgmentOriginal: judgment, lineTextsBottomUp: lines, specialUsageText: special };
+}
+
+function extractClassicSection(wikitext) {
+  const normalized = stripMarkup(wikitext).replace(/\u3000/gu, " ");
+  const classicIndex = normalized.search(/易[經经][：:]/u);
+  if (classicIndex < 0) throw new Error("missing 易經 section");
+  const afterClassic = normalized.slice(classicIndex);
+  const stop = afterClassic.search(/\n[ *#]*彖曰[：:]/u);
+  return (stop >= 0 ? afterClassic.slice(0, stop) : afterClassic)
+    .replace(/^易[經经][：:]\s*/u, "")
+    .trim();
+}
+
+function stripHtml(value) {
+  return value
+    .replace(/<[^>]+>/gu, "")
+    .replace(/&quot;/gu, "\"")
+    .replace(/&amp;/gu, "&")
+    .replace(/&lt;/gu, "<")
+    .replace(/&gt;/gu, ">")
+    .replace(/\n+/gu, "\n")
+    .trim();
+}
+
+async function curlJson(args) {
+  const { stdout } = await execFileAsync("curl.exe", [
+    "-sS",
+    "--fail",
+    "--retry",
+    "12",
+    "--retry-delay",
+    "20",
+    "--retry-all-errors",
+    "--max-time",
+    "45",
+    "-x",
+    proxyUrl,
+    "-H",
+    `User-Agent: ${userAgent}`,
+    "-H",
+    "Accept: application/json",
+    ...args,
+  ], { encoding: "utf8", maxBuffer: 8 * 1024 * 1024 });
+  return JSON.parse(stdout);
+}
+
+const titleFallbacks = {
+  恆: ["恒"],
+  遯: ["遁"],
+  晉: ["晋"],
+  大壯: ["大壮"],
+  歸妹: ["归妹"],
+  豐: ["丰"],
+  兌: ["兑"],
+  渙: ["涣"],
+  節: ["节"],
+  既濟: ["既济"],
+  未濟: ["未济"],
+};
+
+async function fetchWikitext(title) {
+  const candidates = [title, ...(titleFallbacks[title] ?? [])];
+  for (const candidate of candidates) {
+    const url = new URL("https://zh.wikisource.org/w/api.php");
+    url.searchParams.set("action", "parse");
+    url.searchParams.set("page", `周易/${candidate}`);
+    url.searchParams.set("prop", "wikitext");
+    url.searchParams.set("format", "json");
+    const payload = await curlJson([url.toString()]);
+    const wikitext = payload?.parse?.wikitext?.["*"];
+    if (typeof wikitext === "string") return wikitext;
+    process.stdout.write(`missing ${candidate}, trying fallback\n`);
+  }
+  throw new Error(`no wikitext for ${title}`);
+}
+
+async function toHansFields(classic) {
+  const pieces = [
+    classic.judgmentOriginal,
+    ...classic.lineTextsBottomUp,
+    ...(classic.specialUsageText == null ? [] : [classic.specialUsageText]),
+  ];
+  const converted = await convertToHans(pieces.join("\n¶\n"));
+  const parts = converted
+    .split("¶")
+    .map((part) => part.replace(/\s+/gu, "").trim())
+    .filter(Boolean);
+  if (parts.length !== pieces.length) {
+    throw new Error(`zh-hans field count ${parts.length} != ${pieces.length}: ${converted}`);
+  }
+  return {
+    judgmentOriginal: parts[0],
+    lineTextsBottomUp: parts.slice(1, 7),
+    specialUsageText: classic.specialUsageText == null ? null : parts[7],
+  };
+}
+async function convertToHans(text) {
+  const payload = await curlJson([
+    "--data-urlencode", "action=parse",
+    "--data-urlencode", `text=${text}`,
+    "--data-urlencode", "prop=text",
+    "--data-urlencode", "variant=zh-hans",
+    "--data-urlencode", "disablelimitreport=1",
+    "--data-urlencode", "wrapoutputclass=",
+    "--data-urlencode", "contentmodel=wikitext",
+    "--data-urlencode", "format=json",
+    "https://zh.wikisource.org/w/api.php",
+  ]);
+  const html = payload?.parse?.text?.["*"];
+  if (typeof html !== "string") throw new Error("zh-hans conversion returned no text");
+  return stripHtml(html);
+}
+
+const entries = kingWenEntries();
+await mkdir(outputDirectory, { recursive: true });
+let imported = [];
+try {
+  imported = JSON.parse(await readFile(outputPath, "utf8"));
+  if (!Array.isArray(imported)) imported = [];
+} catch {
+  imported = [];
+}
+const done = new Set(imported.map((item) => item.kingWenNumber));
+
+for (const [index, title] of titles.entries()) {
+  if (index >= limit) break;
+  const kingWenNumber = index + 1;
+  if (done.has(kingWenNumber)) {
+    process.stdout.write(`skip ${kingWenNumber}/64 ${title}\n`);
+    continue;
+  }
+  const wikitext = await fetchWikitext(title);
+  const traditional = extractClassicSection(wikitext);
+  const classic = await toHansFields(parseClassicFields(traditional));
+  imported.push({
+    kingWenNumber,
+    wikisourceTitle: title,
+    sourcePage: `https://zh.wikisource.org/wiki/周易/${title}`,
+    ...entries[index],
+    ...classic,
+  });
+  imported.sort((left, right) => left.kingWenNumber - right.kingWenNumber);
+  await writeFile(outputPath, `${JSON.stringify(imported, null, 2)}\n`);
+  process.stdout.write(`imported ${kingWenNumber}/64 ${title}\n`);
+  await new Promise((resolve) => setTimeout(resolve, 1200));
+}
+
+process.stdout.write(`wrote ${outputPath} (${imported.length} hexagrams)\n`);
@@ -104,3 +104,52 @@ for (const [name, mutate, expected] of negativeCases) {
 console.log(
  `Content contract verification passed (64 entries cross-checked with Kotlin, ${negativeCases.length} negative fixtures, digest ${digest.slice(0, 12)}…).`,
 );
+
+const packagePath = path.join(repositoryRoot, "content", "packages", "hexagram-content.json");
+const namesPath = path.join(
+  repositoryRoot,
+  "app",
+  "src",
+  "main",
+  "java",
+  "net",
+  "opcapp",
+  "flash",
+  "core",
+  "model",
+  "HexagramNames.kt",
+);
+const namesSource = await readFile(namesPath, "utf8");
+const namesBlock = namesSource.match(/private val names = listOf\(([\s\S]*?)\)/u);
+if (!namesBlock) {
+  throw new Error("HexagramNames.kt name table could not be parsed");
+}
+const catalogNames = [...namesBlock[1].matchAll(/"([^"]+)"/gu)].map((match) => match[1]);
+if (catalogNames.length !== 64) {
+  throw new Error("HexagramNames.kt must contain 64 names");
+}
+
+const published = JSON.parse(await readFile(packagePath, "utf8"));
+const publishedErrors = validateContentPackage(published);
+if (publishedErrors.length > 0) {
+  throw new Error(`Published content package failed validation:\n${publishedErrors.join("\n")}`);
+}
+published.hexagrams.forEach((hexagram) => {
+  const expectedName = catalogNames[hexagram.kingWenNumber - 1];
+  const expectedSymbol = String.fromCodePoint(0x4DC0 + hexagram.kingWenNumber - 1);
+  if (hexagram.name !== expectedName) {
+    throw new Error(`Published package name for ${hexagram.kingWenNumber} must be ${expectedName}`);
+  }
+  if (hexagram.symbol !== expectedSymbol) {
+    throw new Error(`Published package symbol for ${hexagram.kingWenNumber} must be ${expectedSymbol}`);
+  }
+});
+if (published.contentVersion !== "zh-Hans-2026.1") {
+  throw new Error("Published contentVersion must be zh-Hans-2026.1");
+}
+if (!published.specialUsageTexts.qian || !published.specialUsageTexts.kun) {
+  throw new Error("Published package must include Qian 用九 and Kun 用六");
+}
+console.log(
+  `Published content package ${published.contentVersion} passed (digest ${contentDigest(published).slice(0, 12)}…).`,
+);