Java实现Word转PDF的两种常见方案和性能实测

<dependencies>
      <!-- docx4j -->
        <dependency>
            <groupId>org.docx4j</groupId>
            <artifactId>docx4j-core</artifactId>
            <version>11.4.8</version>
        </dependency>

        <dependency>
            <groupId>org.docx4j</groupId>
            <artifactId>docx4j-JAXB-ReferenceImpl</artifactId>
            <version>11.4.8</version>
        </dependency>

        <dependency>
            <groupId>org.docx4j</groupId>
            <artifactId>docx4j-export-fo</artifactId>
            <version>11.4.8</version>
        </dependency>

        <!-- Apache FOP -->
        <dependency>
            <groupId>org.apache.xmlgraphics</groupId>
            <artifactId>fop</artifactId>
            <version>2.9</version>
        </dependency>
</dependencies>

public class LibreOfficeConverter {

  private final String soffice;

  public LibreOfficeConverter(String soffice) {
    this.soffice = soffice;
  }

  public void convert(Path docx, Path outDir) throws Exception {
    ProcessBuilder pb = new ProcessBuilder(
        soffice,
        "--headless",
        "--nologo",
        "--nolockcheck",
        "--nodefault",
        "--nofirststartwizard",
        "--convert-to", "pdf",
        "--outdir", outDir.toString(),
        docx.toString()
    );
    pb.redirectErrorStream(true);

    Process p = pb.start();
    int code = p.waitFor();
    if (code != 0) {
      throw new RuntimeException("LibreOffice failed, exit=" + code);
    }
  }
}

public class Docx4jConverter {

    public void convert(Path docx, Path pdf) throws Exception {

        // 1️⃣ 加载 Word
        WordprocessingMLPackage wordMLPackage =
                WordprocessingMLPackage.load(docx.toFile());

        // 2️⃣ 字体映射（核心）
        IdentityPlusMapper fontMapper = new IdentityPlusMapper();

        // 宋体（核心兜底）
        PhysicalFont simsun = PhysicalFonts.get("SimSun");
        if (simsun == null) {
            throw new RuntimeException("未找到 SimSun 字体，请确认系统已安装宋体");
        }

        fontMapper.put("SimSun", simsun);

        // ===== 常用中文字体映射 =====
        fontMapper.put("隶书", PhysicalFonts.get("LiSu"));
        fontMapper.put("宋体", PhysicalFonts.get("SimSun"));
        fontMapper.put("微软雅黑", PhysicalFonts.get("Microsoft YaHei"));
        fontMapper.put("黑体", PhysicalFonts.get("SimHei"));
        fontMapper.put("楷体", PhysicalFonts.get("KaiTi"));
        fontMapper.put("新宋体", PhysicalFonts.get("NSimSun"));
        fontMapper.put("华文行楷", PhysicalFonts.get("STXingkai"));
        fontMapper.put("华文仿宋", PhysicalFonts.get("STFangsong"));
        fontMapper.put("仿宋", PhysicalFonts.get("FangSong"));
        fontMapper.put("幼圆", PhysicalFonts.get("YouYuan"));
        fontMapper.put("华文宋体", PhysicalFonts.get("STSong"));
        fontMapper.put("华文中宋", PhysicalFonts.get("STZhongsong"));
        fontMapper.put("等线", PhysicalFonts.get("SimSun"));
        fontMapper.put("等线 Light", PhysicalFonts.get("SimSun"));
        fontMapper.put("华文琥珀", PhysicalFonts.get("STHupo"));
        fontMapper.put("华文隶书", PhysicalFonts.get("STLiti"));
        fontMapper.put("华文新魏", PhysicalFonts.get("STXinwei"));
        fontMapper.put("华文彩云", PhysicalFonts.get("STCaiyun"));
        fontMapper.put("方正姚体", PhysicalFonts.get("FZYaoti"));
        fontMapper.put("方正舒体", PhysicalFonts.get("FZShuTi"));
        fontMapper.put("华文细黑", PhysicalFonts.get("STXihei"));
        fontMapper.put("宋体扩展", PhysicalFonts.get("simsun-extB"));
        fontMapper.put("仿宋_GB2312", PhysicalFonts.get("FangSong_GB2312"));
        fontMapper.put("新細明體", PhysicalFonts.get("SimSun"));

        // ===== 修复 “宋体（正文）/ 宋体（标题）/ 台湾字体” 乱码 =====
        PhysicalFonts.put("PMingLiU", simsun);
        PhysicalFonts.put("新細明體", simsun);

        // 3️⃣ 设置到 Word 包
        wordMLPackage.setFontMapper(fontMapper);

        // 4️⃣ 转 PDF
        try (OutputStream os = Files.newOutputStream(pdf)) {
            Docx4J.toPDF(wordMLPackage, os);
        }
    }
}

public class Stats {

  private final List<Long> times = new ArrayList<>();

  public void add(long nanos) {
    times.add(nanos);
  }

  public void print(String name) {
    Collections.sort(times);
    double avg = times.stream().mapToLong(x -> x).average().orElse(0) / 1e6;
    double p50 = times.get(times.size() / 2) / 1e6;
    double p95 = times.get((int)(times.size() * 0.95)) / 1e6;

    System.out.printf(
      "%s -> avg=%.2fms, p50=%.2fms, p95=%.2fms%n",
      name, avg, p50, p95
    );
  }
}

public class BenchMain {

  public static void main(String[] args) throws Exception {

    Path docx = Path.of("samples/sample.docx");
    Path out = Path.of("out");
    Files.createDirectories(out);

    LibreOfficeConverter lo =
        new LibreOfficeConverter("soffice");
    Docx4jConverter d4j =
        new Docx4jConverter();

    int rounds = 5;

    // LibreOffice
    Stats loStats = new Stats();
    for (int i = 0; i < rounds; i++) {
      long t0 = System.nanoTime();
      lo.convert(docx, out);
      loStats.add(System.nanoTime() - t0);
    }

    // docx4j
    Stats d4jStats = new Stats();
    for (int i = 0; i < rounds; i++) {
      long t0 = System.nanoTime();
      d4j.convert(docx, out.resolve("d4j_" + i + ".pdf"));
      d4jStats.add(System.nanoTime() - t0);
    }

    loStats.print("LibreOffice");
    d4jStats.print("docx4j");
  }
}