fix(libreoffice): PDF 轉 DOCX 修正 (v0.1.8)

- 新增 PDF 匯入管線:PDF → DOCX/ODT/RTF/TXT/HTML 使用 --infilter=writer_pdf_import
- 新增輸出檔案存在性驗證,避免 ENOENT 錯誤
- 改善錯誤訊息:密碼保護、檔案損壞、無匯出過濾器等情境
- 重寫測試套件:19 個測試全數通過
This commit is contained in:
Your Name 2026-01-20 14:38:17 +08:00
parent 03be03bbc2
commit a0eccf0437
5 changed files with 369 additions and 107 deletions

View file

@ -1,6 +1,35 @@
import { execFile as execFileOriginal } from "node:child_process";
import { existsSync as existsSyncOriginal } from "node:fs";
import { basename, dirname, join } from "node:path";
import { ExecFileFn } from "./types";
// ==============================================================================
// LibreOffice 轉換器
// ==============================================================================
//
// 🔑 關鍵技術說明:
//
// LibreOffice 有兩種轉換模式:
//
// 1. Export Pipeline輸出轉換
// - 用於DOCX → PDF, ODT → PDF 等
// - 原生格式 → 導出格式
// - 使用 --convert-to 參數
//
// 2. Import Pipeline輸入轉換
// - 用於PDF → DOCX, PDF → ODT 等
// - 非原生格式 → 原生格式
// - 必須使用 --infilter 參數指定 import filter
//
// ⚠️ 常見錯誤:
// PDF → DOCX 若不指定 --infilter=writer_pdf_import
// 會得到 "no export filter" 錯誤
//
// ==============================================================================
// 用於測試的依賴注入類型
export type ExistsSyncFn = (path: string) => boolean;
export const properties = {
from: {
text: [
@ -102,7 +131,7 @@ const filters: Record<FileCategories, Record<string, string>> = {
odt: "writer8",
ott: "writer8_template",
pages: "Apple Pages",
// pdf: "writer_pdf_import",
pdf: "writer_pdf_Export", // PDF 作為輸出格式
psw: "PocketWord File",
rtf: "Rich Text Format",
sdw: "StarOffice_Writer",
@ -123,6 +152,16 @@ const filters: Record<FileCategories, Record<string, string>> = {
calc: {},
};
// ==============================================================================
// PDF Import FilterPDF 作為輸入格式時使用)
// ==============================================================================
const PDF_IMPORT_FILTER = "writer_pdf_import";
// 需要使用 PDF import pipeline 的情況
function needsPdfImportPipeline(inputExt: string, outputExt: string): boolean {
return inputExt === "pdf" && ["docx", "doc", "odt", "rtf", "txt", "html"].includes(outputExt);
}
const getFilters = (fileType: string, converto: string) => {
if (fileType in filters.text && converto in filters.text) {
return [filters.text[fileType], filters.text[converto]];
@ -139,39 +178,115 @@ export function convert(
targetPath: string,
options?: unknown,
execFile: ExecFileFn = execFileOriginal,
existsSync: ExistsSyncFn = existsSyncOriginal,
): Promise<string> {
const outputPath = targetPath.split("/").slice(0, -1).join("/").replace("./", "") ?? targetPath;
const outputDir = dirname(targetPath).replace("./", "") || ".";
const inputFileName = basename(filePath);
const inputBaseName = inputFileName.replace(/\.[^.]+$/, "");
const expectedOutputFile = join(outputDir, `${inputBaseName}.${convertTo}`);
// Build arguments array
const args: string[] = [];
args.push("--headless");
const [inFilter, outFilter] = getFilters(fileType, convertTo);
if (inFilter) {
args.push(`--infilter="${inFilter}"`);
}
// ==============================================================================
// 關鍵分流PDF → 文字格式 vs 其他轉換
// ==============================================================================
if (needsPdfImportPipeline(fileType, convertTo)) {
// PDF → DOCX/ODT 等:必須使用 import pipeline
console.log(`[LibreOffice] Using PDF import pipeline: ${fileType}${convertTo}`);
args.push(`--infilter=${PDF_IMPORT_FILTER}`);
if (outFilter) {
args.push("--convert-to", `${convertTo}:${outFilter}`, "--outdir", outputPath, filePath);
// 輸出格式仍需指定 filter
const outFilter = filters.text[convertTo];
if (outFilter && convertTo !== "pdf") {
args.push("--convert-to", `${convertTo}:${outFilter}`);
} else {
args.push("--convert-to", convertTo);
}
} else {
args.push("--convert-to", convertTo, "--outdir", outputPath, filePath);
// 一般轉換流程export pipeline
const [inFilter, outFilter] = getFilters(fileType, convertTo);
if (inFilter && fileType !== "pdf") {
args.push(`--infilter=${inFilter}`);
}
if (outFilter) {
args.push("--convert-to", `${convertTo}:${outFilter}`);
} else {
args.push("--convert-to", convertTo);
}
}
args.push("--outdir", outputDir, filePath);
console.log(`[LibreOffice] Command: soffice ${args.join(" ")}`);
return new Promise((resolve, reject) => {
execFile("soffice", args, (error, stdout, stderr) => {
if (error) {
reject(`error: ${error}`);
}
// ==============================================================================
// 錯誤處理與輸出檔案驗證
// ==============================================================================
if (stdout) {
console.log(`stdout: ${stdout}`);
console.log(`[LibreOffice] stdout: ${stdout}`);
}
if (stderr) {
console.error(`stderr: ${stderr}`);
console.error(`[LibreOffice] stderr: ${stderr}`);
}
// 檢查 LibreOffice 執行錯誤
if (error) {
const errorMsg = getLibreOfficeErrorMessage(error, stderr);
console.error(`[LibreOffice] Error: ${errorMsg}`);
reject(errorMsg);
return;
}
// ==============================================================================
// 關鍵防呆:檢查輸出檔案是否實際存在
// ==============================================================================
if (!existsSync(expectedOutputFile)) {
const errorMsg = `LibreOffice 轉換失敗:輸出檔案不存在 (${expectedOutputFile})。可能原因1) 輸入檔案損壞或加密 2) 缺少必要字型 3) 格式不支援`;
console.error(`[LibreOffice] ${errorMsg}`);
reject(errorMsg);
return;
}
console.log(`[LibreOffice] Successfully created: ${expectedOutputFile}`);
resolve("Done");
});
});
}
// ==============================================================================
// 錯誤訊息解析
// ==============================================================================
function getLibreOfficeErrorMessage(
error: Error & { code?: number | string },
stderr: string,
): string {
const stderrLower = stderr.toLowerCase();
// 常見錯誤類型判斷
if (stderrLower.includes("no export filter")) {
return "LibreOffice 錯誤:找不到 export filter。可能是格式不支援或需要使用不同的轉換路徑。";
}
if (stderrLower.includes("password") || stderrLower.includes("encrypted")) {
return "LibreOffice 錯誤:檔案已加密或需要密碼。請先解除密碼保護。";
}
if (stderrLower.includes("corrupt") || stderrLower.includes("damaged")) {
return "LibreOffice 錯誤:檔案已損壞或格式無效。";
}
if (error.code === "ENOENT") {
return "LibreOffice 錯誤:找不到 soffice 執行檔。請確認 LibreOffice 已正確安裝。";
}
// 通用錯誤
return `LibreOffice 轉換失敗 (exit code: ${error.code || "unknown"}): ${stderr || error.message}`;
}