@qcplay/cli 1.0.13 → 1.0.15

package/lib/lark-article.js

@@ -0,0 +1,294 @@
+import { spawn } from "child_process";
+
+import * as cheerio from "cheerio";
+
+import { normalizeArticleColor, sanitizeArticleRichHtml } from "./wechat-article.js";
+
+const LARK_DOCUMENT_PATH = /^\/(?:wiki|docx)\/([A-Za-z0-9_-]+)\/?$/;
+
+function normalizeText(value) {
+  return String(value || "")
+    .replace(/\u00a0/g, " ")
+    .replace(/[\t\r\n ]+/g, " ")
+    .trim();
+}
+
+function appendStyle(element, declarations) {
+  const existing = String(element.attr("style") || "").trim().replace(/;+$/, "");
+  const added = declarations.filter(Boolean).join(";");
+  element.attr("style", [existing, added].filter(Boolean).join(";"));
+}
+
+function larkHostname(hostname) {
+  const value = String(hostname || "").toLowerCase();
+  return (
+    value === "feishu.cn" ||
+    value.endsWith(".feishu.cn") ||
+    value === "larksuite.com" ||
+    value.endsWith(".larksuite.com") ||
+    value === "larkoffice.com" ||
+    value.endsWith(".larkoffice.com")
+  );
+}
+
+export function parseLarkDocumentUrl(value) {
+  let url;
+  try {
+    url = new URL(String(value || ""));
+  } catch {
+    throw new Error("飞书文档链接无效");
+  }
+  if (url.protocol !== "https:" || !larkHostname(url.hostname) || !LARK_DOCUMENT_PATH.test(url.pathname)) {
+    throw new Error("仅支持飞书 Docx 或 Wiki 文档链接");
+  }
+  url.search = "";
+  url.hash = "";
+  return url;
+}
+
+function larkError(output, fallback) {
+  try {
+    const envelope = JSON.parse(output);
+    const message = normalizeText(envelope.error?.message);
+    const hint = normalizeText(envelope.error?.hint);
+    return [message, hint].filter(Boolean).join("；") || fallback;
+  } catch {
+    return normalizeText(output) || fallback;
+  }
+}
+
+function runLarkJson(args) {
+  return new Promise((resolve, reject) => {
+    const executable = process.platform === "win32" ? "lark-cli.cmd" : "lark-cli";
+    const child = spawn(executable, args, {
+      windowsHide: true,
+      shell: process.platform === "win32",
+      stdio: ["ignore", "pipe", "pipe"],
+      env: {
+        ...process.env,
+        LARKSUITE_CLI_NO_UPDATE_NOTIFIER: "1",
+        LARKSUITE_CLI_NO_SKILLS_NOTIFIER: "1"
+      }
+    });
+    const stdout = [];
+    const stderr = [];
+    child.stdout.on("data", chunk => stdout.push(chunk));
+    child.stderr.on("data", chunk => stderr.push(chunk));
+    child.once("error", error => reject(new Error(`无法启动 lark-cli: ${error.message}`)));
+    child.once("close", code => {
+      const output = Buffer.concat(stdout).toString("utf8").trim();
+      const diagnostics = Buffer.concat(stderr).toString("utf8").trim();
+      if (code !== 0) {
+        reject(new Error(`飞书文档读取失败: ${larkError(diagnostics || output, `lark-cli 退出码 ${code}`)}`));
+        return;
+      }
+      try {
+        resolve(JSON.parse(output));
+      } catch {
+        reject(new Error("飞书文档读取失败: lark-cli 没有返回有效 JSON"));
+      }
+    });
+  });
+}
+
+export async function fetchLarkDocument(value, options = {}) {
+  const sourceUrl = parseLarkDocumentUrl(value);
+  const run = options.runLarkJson || runLarkJson;
+  const envelope = await run([
+    "docs",
+    "+fetch",
+    "--doc",
+    sourceUrl.toString(),
+    "--detail",
+    "full",
+    "--as",
+    "user",
+    "--format",
+    "json"
+  ]);
+  if (envelope?.ok !== true) {
+    throw new Error(`飞书文档读取失败: ${larkError(JSON.stringify(envelope || {}), "返回状态异常")}`);
+  }
+  const document = envelope.data?.document;
+  if (!normalizeText(document?.content)) {
+    throw new Error("飞书文档正文为空");
+  }
+  return {
+    sourceUrl: sourceUrl.toString(),
+    documentId: normalizeText(document.document_id),
+    revisionId: document.revision_id,
+    xml: String(document.content)
+  };
+}
+
+function styleLarkElement($, node) {
+  const element = $(node);
+  const tag = String(node.tagName || node.name || "").toLowerCase();
+  const align = String(element.attr("align") || "").toLowerCase();
+  if (["left", "center", "right", "justify"].includes(align)) {
+    appendStyle(element, [`text-align:${align}`]);
+  }
+
+  const textColor = normalizeArticleColor(element.attr("text-color"));
+  const backgroundColor = normalizeArticleColor(element.attr("background-color"));
+  appendStyle(element, [textColor ? `color:${textColor}` : "", backgroundColor ? `background-color:${backgroundColor}` : ""]);
+
+  if (/^h[1-6]$/.test(tag)) {
+    const level = Number(tag.slice(1));
+    const sizes = [28, 24, 21, 19, 17, 16];
+    appendStyle(element, [
+      "margin:30px 0 14px",
+      `font-size:${sizes[level - 1]}px`,
+      "line-height:1.5",
+      "font-weight:700",
+      "color:#24292f"
+    ]);
+  } else if (tag === "p") {
+    appendStyle(element, [element.parents("td,th").length ? "margin:0 0 8px" : "margin:0 0 18px", "line-height:1.8"]);
+  } else if (tag === "blockquote") {
+    appendStyle(element, [
+      "margin:22px 0",
+      "padding:14px 18px",
+      "border-left:4px solid #3370ff",
+      "background-color:#f5f7fa",
+      "color:#3f4654"
+    ]);
+  } else if (tag === "table") {
+    const totalWidth = element
+      .find("col")
+      .toArray()
+      .reduce((sum, col) => sum + Number.parseInt($(col).attr("width") || "0", 10), 0);
+    appendStyle(element, [
+      "width:100%",
+      `min-width:${Math.max(560, Math.min(totalWidth || 560, 1200))}px`,
+      "margin:22px 0",
+      "border-collapse:collapse",
+      "table-layout:fixed",
+      "color:#24292f"
+    ]);
+  } else if (tag === "td" || tag === "th") {
+    appendStyle(element, [
+      "padding:10px 12px",
+      "border:1px solid #d9dce3",
+      "vertical-align:middle",
+      "font-size:15px",
+      "line-height:1.7",
+      "overflow-wrap:anywhere",
+      "word-break:break-word"
+    ]);
+  } else if (tag === "col") {
+    const width = Number.parseInt(element.attr("width") || "0", 10);
+    if (width > 0) {
+      appendStyle(element, [`width:${Math.min(width, 1600)}px`]);
+    }
+  } else if (tag === "ol" || tag === "ul") {
+    appendStyle(element, ["margin:8px 0 16px", "padding-left:1.8em", "line-height:1.8"]);
+  } else if (tag === "li") {
+    appendStyle(element, ["margin:5px 0"]);
+  } else if (tag === "a") {
+    appendStyle(element, ["color:#245bdb", "text-decoration:underline", "overflow-wrap:anywhere"]);
+  }
+}
+
+export function convertLarkDocumentXml(xml, sourceUrl = "") {
+  const $ = cheerio.load(`<qcplay-root>${String(xml || "")}</qcplay-root>`, { xmlMode: true }, false);
+  const root = $("qcplay-root").first();
+  if (!root.length) {
+    throw new Error("飞书文档 XML 结构无效");
+  }
+
+  const titleElement = root.children("title").first();
+  const title = normalizeText(titleElement.text());
+  if (!title) {
+    throw new Error("飞书文档缺少标题");
+  }
+  if (titleElement.length) {
+    titleElement.get(0).name = "h1";
+  }
+
+  root.find("ol").each((_, list) => {
+    const firstSequence = Number.parseInt($(list).children("li").first().attr("seq") || "1", 10);
+    if (firstSequence > 1) {
+      $(list).attr("start", String(firstSequence));
+    }
+  });
+
+  root.find("img").each((_, image) => {
+    const element = $(image);
+    const url = normalizeText(element.attr("url") || element.attr("href") || element.attr("src"));
+    if (!url) {
+      throw new Error(`飞书文档图片缺少可访问地址: ${normalizeText(element.attr("token")) || "未知图片"}`);
+    }
+    let imageUrl;
+    try {
+      imageUrl = new URL(url);
+    } catch {
+      throw new Error("飞书文档图片地址无效");
+    }
+    if (!new Set(["http:", "https:"]).has(imageUrl.protocol)) {
+      throw new Error("飞书文档图片地址仅支持 HTTP/HTTPS");
+    }
+    element.attr("src", imageUrl.toString());
+    const width = Number.parseInt(element.attr("width") || "0", 10);
+    const height = Number.parseInt(element.attr("height") || "0", 10);
+    appendStyle(element, [
+      "display:block",
+      "max-width:100%",
+      "height:auto",
+      "margin:20px auto",
+      width > 0 ? `width:${Math.min(width, 1600)}px` : ""
+    ]);
+    if (width > 0) element.attr("width", String(width));
+    if (height > 0) element.attr("height", String(height));
+  });
+
+  root.find("source").each((_, source) => {
+    const element = $(source);
+    const url = normalizeText(element.attr("url"));
+    const name = normalizeText(element.attr("name")) || "飞书文档附件";
+    const paragraph = $("<p></p>");
+    if (url) {
+      paragraph.append($("<a></a>").attr("href", url).text(name));
+    } else {
+      paragraph.text(name);
+    }
+    element.replaceWith(paragraph);
+  });
+
+  root.find("*").each((_, node) => styleLarkElement($, node));
+  root.find("*").each((_, node) => {
+    const element = $(node);
+    for (const attribute of ["id", "token", "seq", "text-color", "background-color", "align", "url", "name"]) {
+      element.removeAttr(attribute);
+    }
+  });
+
+  const richHtml = sanitizeArticleRichHtml(root.html() || "", { maxWidth: 960 });
+  if (!richHtml) {
+    throw new Error("飞书文档没有可转换的正文内容");
+  }
+  const excerpt = normalizeText(
+    root
+      .children()
+      .filter((_, node) => String(node.tagName || node.name || "").toLowerCase() !== "h1")
+      .first()
+      .text()
+  ).slice(0, 160);
+  return {
+    title,
+    author: "飞书文档",
+    releaseDate: "",
+    excerpt,
+    thumbnail: normalizeText(root.find("img").first().attr("src")),
+    markdown: "",
+    richHtml,
+    imageCount: root.find("img").length,
+    tableCount: root.find("table").length,
+    sourceUrl: parseLarkDocumentUrl(sourceUrl).toString()
+  };
+}
+
+export async function importLarkArticle(value, options = {}) {
+  const document = await (options.fetchDocument || fetchLarkDocument)(value, options);
+  return convertLarkDocumentXml(document.xml, document.sourceUrl || value);
+}