资料上传支持 Excel 全格式：xlsx、xls、xlsm、xlsb、ods

- 改用 SheetJS（官方 CDN 0.20.3，devDependency）读取表格：按工作表逐行转文字，
  公式取计算结果，日期、百分比按单元格格式显示，去掉空行和行尾空格
- 兼容业务系统导出的网页格式 .xls；损坏或加密的文件给出明确提示，不存乱码
- 资料页与使用说明列出支持的格式

Co-Authored-By: Claude Opus 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01MMDFmVTvRksi1ZT8aPGXoL

docs/user-guide.md

@@ -64,7 +64,7 @@
 
 ### 7. 资料与企业画像
 
-- **资料库**：上传经营总结、行业报告等（Word、PDF、Excel、txt / md / csv，单个不超过 20MB），平台在浏览器里提取文字后保存；也可以直接粘贴文字。数字咨询师会检索资料并注明出处。
+- **资料库**：上传经营总结、行业报告、经营数据表等（Word docx、PDF、Excel xlsx / xls / xlsm / xlsb / ods、txt / md / csv，单个不超过 20MB），平台在浏览器里提取文字后保存；也可以直接粘贴文字。Excel 按工作表逐行读取，公式取计算结果、日期和百分比按表格里的显示；业务系统导出的网页格式 .xls 也能读。数字咨询师会检索资料并注明出处。
 - **企业画像**：按“事实 / 假设 / 未知 / 待补”分类记录企业情况，每条注明来源。可以让数字咨询师根据资料起草画像条目（需确认写入，起草的条目一律为草稿）。
 - 画像条目由咨询师**复核**、企业项目负责人**确认**；他人修改已复核或已确认的条目，会回到草稿。两人同时编辑时，后保存的一方会被提示刷新。
 - 画像摘要会随页面上下文提供给各阶段的数字咨询师。

package.json

@@ -90,6 +90,7 @@
     "typescript-eslint": "^8.30.1",
     "vite": "^6.3.5",
     "vitest": "^4.1.10",
+    "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz",
     "zod": "^3.25.76"
   },
   "publishConfig": {

pnpm-lock.yaml

@@ -257,6 +257,9 @@ importers:
       vitest:
         specifier: ^4.1.10
         version: 4.1.10(@opentelemetry/[email protected])(@types/[email protected])(@vitest/[email protected])([email protected])([email protected](@types/[email protected])([email protected])([email protected])([email protected]))
+      xlsx:
+        specifier: https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz
+        version: https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz
       zod:
         specifier: ^3.25.76
         version: 3.25.76
@@ -5887,6 +5890,12 @@ packages:
     resolution: {integrity: sha512-g/eziiSUNBSsdDJtCLB8bdYEUMj4jR7AGeUo96p/3dTafgjHhpF4RiCFPiRILwjQoDXx5MqkBr4fwWtR3Ky4Wg==}
     engines: {node: '>=20'}
 
+  xlsx@https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz:
+    resolution: {integrity: sha512-oLDq3jw7AcLqKWH2AhCpVTZl8mf6X2YReP+Neh0SJUzV/BdZYjth94tG5toiMB1PPrYtxOCfaoUCkvtuH+3AJA==, tarball: https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz}
+    version: 0.20.3
+    engines: {node: '>=0.8'}
+    hasBin: true
+
   [email protected]:
     resolution: {integrity: sha512-7rVi2KMfwfWFl+GpPg6m80IVMWXLRjO+PxTq7V2CDhoGak0wzYzFgUY2m4XJ47OGdXd8eLE8EmwfAmdjw7lC1g==}
     hasBin: true
@@ -12127,6 +12136,8 @@ snapshots:
       is-wsl: 3.1.1
       powershell-utils: 0.1.0
 
+  xlsx@https://cdn.sheetjs.com/xlsx-0.20.3/xlsx-0.20.3.tgz: {}
+
   [email protected]:
     dependencies:
       sax: 1.6.1

src/lib/dingze/extract-text.ts

@@ -1,7 +1,9 @@
 // Turn an uploaded file into plain text in the browser; only the text is stored. The parsers
 // load on demand so they stay out of the main bundle.
 
-export const MATERIAL_ACCEPT = ".txt,.md,.csv,.docx,.pdf,.xlsx";
+/** Spreadsheet formats read with SheetJS: Excel 2007+ and 97–2003, macro / binary workbooks, OpenDocument. */
+export const SPREADSHEET_EXTENSIONS = ["xlsx", "xls", "xlsm", "xlsb", "ods"];
+export const MATERIAL_ACCEPT = [".txt", ".md", ".csv", ".docx", ".pdf", ...SPREADSHEET_EXTENSIONS.map((e) => `.${e}`)].join(",");
 export const MATERIAL_MAX_BYTES = 20 * 1024 * 1024;
 
 export class UnsupportedFileError extends Error {}
@@ -29,19 +31,34 @@ export async function extractText(file: File): Promise<string> {
     }
     return pages.join("\n\n");
   }
-  if (ext === "xlsx") {
-    const { Workbook } = await import("exceljs");
-    const book = new Workbook();
-    await book.xlsx.load(await file.arrayBuffer());
-    const lines: string[] = [];
-    book.eachSheet((sheet) => {
-      lines.push(`【${sheet.name}】`);
-      sheet.eachRow((row) => {
-        const values = (row.values as unknown[]).slice(1).map((v) => (v && typeof v === "object" && "text" in (v as object) ? String((v as { text: unknown }).text) : v ?? ""));
-        lines.push(values.join("\t"));
-      });
-    });
-    return lines.join("\n");
+  if (SPREADSHEET_EXTENSIONS.includes(ext)) return spreadsheetText(await file.arrayBuffer());
+  throw new UnsupportedFileError("暂不支持这种格式：请上传 txt、md、csv、docx、pdf 或 Excel（xlsx、xls 等），或把文字粘贴进来");
+}
+
+/**
+ * Every sheet as a 【name】 heading followed by its rows, cells joined by tabs. Cells show
+ * their formatted text as Excel would (formula results, dates, percentages); empty rows and
+ * trailing empty cells are dropped.
+ */
+export async function spreadsheetText(data: ArrayBuffer | Uint8Array): Promise<string> {
+  const XLSX = await import("xlsx");
+  let book: import("xlsx").WorkBook;
+  try {
+    book = XLSX.read(data, { type: "array", cellDates: true, dense: true });
+  } catch {
+    throw new UnsupportedFileError("读不了这个表格文件，可能已损坏或设了打开密码；请另存为 xlsx 后再上传");
   }
-  throw new UnsupportedFileError("暂不支持这种格式：请上传 txt、md、csv、docx、pdf 或 xlsx，或把文字粘贴进来");
+  const parts: string[] = [];
+  for (const name of book.SheetNames) {
+    const rows = XLSX.utils.sheet_to_json<unknown[]>(book.Sheets[name], { header: 1, raw: false, defval: "", blankrows: false });
+    const lines = rows
+      .map((row) => {
+        const cells = row.map((v) => String(v ?? "").replace(/\s*\n\s*/g, " ").trim());
+        while (cells.length && !cells[cells.length - 1]) cells.pop();
+        return cells.join("\t");
+      })
+      .filter((line) => line.trim());
+    if (lines.length) parts.push(`【${name}】`, ...lines);
+  }
+  return parts.join("\n");
 }

src/pages/projects/materials.tsx

@@ -26,7 +26,7 @@ import { Textarea } from "@/components/ui/textarea";
 import { type ProposedChange, applyChanges } from "@/lib/dingze/changes";
 import { PROPOSE_TOOL } from "@/lib/dingze/coach";
 import { errorMessage } from "@/lib/dingze/errors";
-import { MATERIAL_ACCEPT, extractText } from "@/lib/dingze/extract-text";
+import { MATERIAL_ACCEPT, SPREADSHEET_EXTENSIONS, extractText } from "@/lib/dingze/extract-text";
 import { SEARCH_MATERIALS_TOOL } from "@/lib/dingze/material-tools";
 import { type MaterialSummary, useAddMaterial, useDeleteMaterial, useMaterial, useMaterials, useProfile, useSaveProfile } from "@/lib/dingze/materials-api";
 import { formatTime } from "@/lib/dingze/records";
@@ -108,7 +108,11 @@ function MaterialLibrary({ materials, canEdit, canDeleteAny, currentUserId }: {
       try {
         const text = await extractText(file);
         if (!text.trim()) {
-          toast.error(`《${file.name}》没有读到文字，可能是扫描件；请粘贴文字或上传可复制文字的版本`);
+          toast.error(
+            SPREADSHEET_EXTENSIONS.some((e) => file.name.toLowerCase().endsWith(`.${e}`))
+              ? `《${file.name}》的表格里没有内容`
+              : `《${file.name}》没有读到文字，可能是扫描件；请粘贴文字或上传可复制文字的版本`
+          );
           continue;
         }
         const result = await add.mutateAsync({ title: file.name.replace(/\.[^.]+$/, ""), kind: "file", fileName: file.name, size: file.size, text });
@@ -142,7 +146,7 @@ function MaterialLibrary({ materials, canEdit, canDeleteAny, currentUserId }: {
       </div>
       {materials.length === 0 ? (
         <p className="rounded-xl border border-dashed px-4 py-6 text-center text-sm text-muted-foreground">
-          还没有资料。可上传 txt、md、csv、docx、pdf、xlsx（不超过 20 MB），或直接粘贴文字。
+          还没有资料。可上传 Word（docx）、PDF、Excel（xlsx、xls、xlsm、xlsb、ods）、txt、md、csv（不超过 20 MB），或直接粘贴文字。
         </p>
       ) : (
         <ul className="overflow-hidden rounded-xl border bg-card">

tests/logic/dingze/extract-text.test.ts

@@ -0,0 +1,55 @@
+import * as XLSX from "xlsx";
+import { describe, expect, it } from "vitest";
+
+import { MATERIAL_ACCEPT, UnsupportedFileError, spreadsheetText } from "@/lib/dingze/extract-text";
+
+function workbook(bookType: XLSX.BookType): Uint8Array {
+  const book = XLSX.utils.book_new();
+  const sales = XLSX.utils.aoa_to_sheet([
+    ["业务", "2025 销量（万方）", "占比"],
+    ["车用气", 360, 0.4],
+    [],
+    ["工商业", 540, 0.6],
+  ]);
+  sales["C2"].z = "0%";
+  sales["C4"].z = "0%";
+  // A formula with its cached result, as Excel saves it.
+  sales["B5"] = { t: "n", f: "SUM(B2:B4)", v: 900 };
+  sales["A5"] = { t: "s", v: "合计" };
+  sales["!ref"] = "A1:C5";
+  XLSX.utils.book_append_sheet(book, sales, "销量");
+  XLSX.utils.book_append_sheet(book, XLSX.utils.aoa_to_sheet([["事项", "日期"], ["站点开工", new Date(Date.UTC(2026, 2, 15))]], { cellDates: true, dateNF: "yyyy-mm-dd" }), "节点");
+  XLSX.utils.book_append_sheet(book, XLSX.utils.aoa_to_sheet([]), "空表");
+  return new Uint8Array(XLSX.write(book, { type: "array", bookType }));
+}
+
+describe("spreadsheet materials", () => {
+  for (const bookType of ["xlsx", "xls", "ods"] as const) {
+    it(`reads every sheet of a .${bookType} as text`, async () => {
+      const text = await spreadsheetText(workbook(bookType));
+      const lines = text.split("\n");
+      expect(lines[0]).toBe("【销量】");
+      // ODS as written here keeps the value but not the percent format.
+      expect(lines).toContain(bookType === "ods" ? "车用气\t360\t0.4" : "车用气\t360\t40%");
+      expect(lines).toContain("合计\t900");
+      expect(lines).not.toContain("");
+      expect(text).toContain("【节点】");
+      expect(text).toMatch(/站点开工\t2026-03-15/);
+      expect(text).not.toContain("【空表】");
+    });
+  }
+
+  it("explains a damaged file instead of storing garbage", async () => {
+    await expect(spreadsheetText(workbook("xlsx").slice(0, 200))).rejects.toBeInstanceOf(UnsupportedFileError);
+    await expect(spreadsheetText(workbook("xls").slice(0, 600))).rejects.toBeInstanceOf(UnsupportedFileError);
+  });
+
+  it("reads the HTML tables that business systems export with an .xls name", async () => {
+    const html = "<html><body><table><tr><td>项目</td><td>金额</td></tr><tr><td>光伏</td><td>120</td></tr></table></body></html>";
+    expect(await spreadsheetText(new TextEncoder().encode(html))).toBe("【Sheet1】\n项目\t金额\n光伏\t120");
+  });
+
+  it("offers the Excel formats in the file picker", () => {
+    for (const ext of [".xlsx", ".xls", ".xlsm", ".xlsb", ".ods", ".docx", ".pdf"]) expect(MATERIAL_ACCEPT.split(",")).toContain(ext);
+  });
+});