| | |
| | | import com.documents4j.api.IConverter; |
| | | import com.documents4j.job.LocalConverter; |
| | | import lombok.extern.slf4j.Slf4j; |
| | | import org.apache.poi.xwpf.usermodel.*; |
| | | import org.springframework.beans.factory.annotation.Autowired; |
| | | import org.springframework.mock.web.MockMultipartFile; |
| | | import org.springframework.stereotype.Component; |
| | |
| | | |
| | | import java.io.*; |
| | | import java.net.URL; |
| | | import java.nio.charset.StandardCharsets; |
| | | import java.nio.file.Files; |
| | | import java.util.ArrayList; |
| | | import java.util.List; |
| | | import java.util.Map; |
| | | import java.util.concurrent.TimeUnit; |
| | | |
| | | @Slf4j |
| | |
| | | /** |
| | | * word 转 pdf |
| | | * |
| | | * @param url |
| | | */ |
| | | // public String wordToPdf(String url,String filePath, String fileName) { |
| | | // try { |
| | |
| | | // return null; |
| | | // } |
| | | |
| | | public String wordToPdf(String url, String filePath, String fileName) { |
| | | public String wordToPdf(String filePath, String fileName) { |
| | | try { |
| | | DocumentType documentType = DocumentType.DOC; |
| | | if (url.contains(".docx")) { |
| | | documentType = DocumentType.DOCX; |
| | | } else if (url.contains(".doc")) { |
| | | documentType = DocumentType.DOC; |
| | | } else if (url.contains(".xlsx")) { |
| | | documentType = DocumentType.XLSX; |
| | | } else if (url.contains(".xls")) { |
| | | documentType = DocumentType.XLS; |
| | | } |
| | | // 1. 首先确保输入文件是UTF-8编码 |
| | | String inputFile = filePath + fileName; |
| | | String outputDir = filePath + "/pdf"; |
| | | String outputFileName = fileName.substring(0, fileName.lastIndexOf(".")) + ".pdf"; |
| | | |
| | | // Ensure the URL has a protocol part |
| | | if (!url.startsWith("file://") && !url.startsWith("http://") && !url.startsWith("https://")) { |
| | | url = "file://" + url; |
| | | } |
| | | // 2. 创建临时文件用于转换 |
| | | String tempDocx = createTempFileWithEncoding(inputFile); |
| | | |
| | | // 使用LibreOffice进行转换 |
| | | ProcessBuilder pb = new ProcessBuilder( |
| | | "soffice", |
| | | "--headless", |
| | | "--convert-to", "pdf", |
| | | "--outdir", new File(filePath+ "/pdf"+fileName.substring(0, fileName.lastIndexOf(".")) + ".pdf").getParent(), |
| | | filePath+fileName |
| | | ); |
| | | // 3. 使用更详细的LibreOffice转换参数 |
| | | List<String> command = new ArrayList<>(); |
| | | command.add("/usr/bin/soffice"); |
| | | command.add("--headless"); |
| | | command.add("--convert-to"); |
| | | command.add("pdf:writer_pdf_Export:PDFExport{'EmbedStandardFonts':true}"); |
| | | command.add("--outdir"); |
| | | command.add(outputDir); |
| | | command.add(tempDocx); |
| | | |
| | | ProcessBuilder pb = new ProcessBuilder(command); |
| | | |
| | | // 4. 设置更完整的环境变量 |
| | | Map<String, String> env = pb.environment(); |
| | | env.put("LC_ALL", "zh_CN.UTF-8"); |
| | | env.put("LANG", "zh_CN.UTF-8"); |
| | | env.put("LANGUAGE", "zh_CN.UTF-8"); |
| | | env.put("PYTHONIOENCODING", "utf8"); |
| | | env.put("JAVA_TOOL_OPTIONS", "-Dfile.encoding=UTF-8"); |
| | | |
| | | // 5. 执行转换 |
| | | Process process = pb.start(); |
| | | |
| | | // 等待转换完成 |
| | | int exitCode = process.waitFor(); |
| | | if (exitCode == 0) { |
| | | System.out.println("PDF转换成功!"); |
| | | } else { |
| | | System.out.println("PDF转换失败!"); |
| | | // 6. 读取输出 |
| | | StringBuilder output = new StringBuilder(); |
| | | try (BufferedReader reader = new BufferedReader( |
| | | new InputStreamReader(process.getInputStream(), StandardCharsets.UTF_8))) { |
| | | String line; |
| | | while ((line = reader.readLine()) != null) { |
| | | output.append(line).append("\n"); |
| | | } |
| | | } |
| | | return ""; |
| | | |
| | | // 7. 等待进程完成 |
| | | int exitCode = process.waitFor(); |
| | | |
| | | // 8. 清理临时文件 |
| | | new File(tempDocx).delete(); |
| | | |
| | | if (exitCode == 0) { |
| | | return outputDir + "/" + outputFileName; |
| | | } else { |
| | | throw new RuntimeException("转换失败: " + output.toString()); |
| | | } |
| | | |
| | | } catch (Exception e) { |
| | | e.printStackTrace(); |
| | | throw new RuntimeException("Failed to generate PDF: " + e.getMessage(), e); |
| | | throw new RuntimeException("PDF转换失败: " + e.getMessage(), e); |
| | | } |
| | | } |
| | | private static String createTempFileWithEncoding(String inputFile) { |
| | | try { |
| | | // 1. 读取原始文档 |
| | | XWPFDocument doc = new XWPFDocument(new FileInputStream(inputFile)); |
| | | |
| | | // 2. 修改文档字体和编码 |
| | | for (XWPFParagraph paragraph : doc.getParagraphs()) { |
| | | for (XWPFRun run : paragraph.getRuns()) { |
| | | // 设置中文字体 |
| | | run.setFontFamily("WenQuanYi Zen Hei"); |
| | | run.setFontFamily("WenQuanYi Zen Hei", XWPFRun.FontCharRange.eastAsia); |
| | | |
| | | // 确保文本是UTF-8编码 |
| | | String text = run.getText(0); |
| | | if (text != null) { |
| | | text = new String(text.getBytes(StandardCharsets.UTF_8), StandardCharsets.UTF_8); |
| | | run.setText(text, 0); |
| | | } |
| | | } |
| | | } |
| | | |
| | | // 3. 处理表格中的文本 |
| | | for (XWPFTable table : doc.getTables()) { |
| | | for (XWPFTableRow row : table.getRows()) { |
| | | for (XWPFTableCell cell : row.getTableCells()) { |
| | | for (XWPFParagraph paragraph : cell.getParagraphs()) { |
| | | for (XWPFRun run : paragraph.getRuns()) { |
| | | run.setFontFamily("WenQuanYi Zen Hei"); |
| | | run.setFontFamily("WenQuanYi Zen Hei", XWPFRun.FontCharRange.eastAsia); |
| | | |
| | | String text = run.getText(0); |
| | | if (text != null) { |
| | | text = new String(text.getBytes(StandardCharsets.UTF_8), StandardCharsets.UTF_8); |
| | | run.setText(text, 0); |
| | | } |
| | | } |
| | | } |
| | | } |
| | | } |
| | | } |
| | | |
| | | // 4. 保存为临时文件 |
| | | String tempFile = inputFile + ".temp.docx"; |
| | | try (FileOutputStream out = new FileOutputStream(tempFile)) { |
| | | doc.write(out); |
| | | } |
| | | |
| | | return tempFile; |
| | | |
| | | } catch (Exception e) { |
| | | throw new RuntimeException("处理文档编码失败: " + e.getMessage(), e); |
| | | } |
| | | } |
| | | |
| | | // 在转换之前执行的环境检查 |
| | | private static void preConversionCheck() { |
| | | try { |
| | | // 1. 检查并安装必要的字体 |
| | | installRequiredFonts(); |
| | | |
| | | // 2. 验证LibreOffice安装 |
| | | verifyLibreOfficeInstallation(); |
| | | |
| | | // 3. 设置字体配置 |
| | | setupFontConfig(); |
| | | |
| | | } catch (Exception e) { |
| | | throw new RuntimeException("环境检查失败: " + e.getMessage(), e); |
| | | } |
| | | } |
| | | |
| | | private static void installRequiredFonts() { |
| | | try { |
| | | ProcessBuilder pb = new ProcessBuilder( |
| | | "sudo", "apt-get", "install", "-y", |
| | | "fonts-wqy-zenhei", |
| | | "fonts-wqy-microhei", |
| | | "fonts-arphic-ukai", |
| | | "fonts-arphic-uming" |
| | | ); |
| | | pb.inheritIO(); |
| | | Process p = pb.start(); |
| | | p.waitFor(); |
| | | } catch (Exception e) { |
| | | System.err.println("警告: 安装字体失败 - " + e.getMessage()); |
| | | } |
| | | } |
| | | |
| | | private static void verifyLibreOfficeInstallation() { |
| | | try { |
| | | Process p = Runtime.getRuntime().exec("soffice --version"); |
| | | try (BufferedReader reader = new BufferedReader( |
| | | new InputStreamReader(p.getInputStream()))) { |
| | | String version = reader.readLine(); |
| | | if (version == null || !version.contains("LibreOffice")) { |
| | | throw new RuntimeException("LibreOffice未正确安装"); |
| | | } |
| | | } |
| | | } catch (Exception e) { |
| | | throw new RuntimeException("LibreOffice检查失败: " + e.getMessage()); |
| | | } |
| | | } |
| | | |
| | | private static void setupFontConfig() { |
| | | try { |
| | | String fontConfig = |
| | | "<?xml version='1.0'?>\n" + |
| | | "<!DOCTYPE fontconfig SYSTEM 'fonts.dtd'>\n" + |
| | | "<fontconfig>\n" + |
| | | " <match target=\"pattern\">\n" + |
| | | " <test name=\"family\"><string>SimSun</string></test>\n" + |
| | | " <edit name=\"family\" mode=\"assign\" binding=\"same\">\n" + |
| | | " <string>WenQuanYi Zen Hei</string>\n" + |
| | | " </edit>\n" + |
| | | " </match>\n" + |
| | | "</fontconfig>"; |
| | | |
| | | String userHome = System.getProperty("user.home"); |
| | | File fontConfigFile = new File(userHome + "/.fonts.conf"); |
| | | |
| | | try (FileWriter writer = new FileWriter(fontConfigFile)) { |
| | | writer.write(fontConfig); |
| | | } |
| | | } catch (Exception e) { |
| | | System.err.println("警告: 设置字体配置失败 - " + e.getMessage()); |
| | | } |
| | | } |
| | | |
| | | public static MultipartFile convertToMultipartFile(ByteArrayOutputStream baos, String fileName) throws IOException { |
| | | // 创建一个临时文件 |
| | |
| | | } |
| | | |
| | | public String test(String fileName){ |
| | | String url = "file:///usr/local/project/file/"+fileName; |
| | | // String url = "file:///usr/local/project/file/"+fileName; |
| | | // String filePath = "E:\\qiyeweixin\\WXWork\\1688855207501340\\Cache\\File\\2024-09"; |
| | | // String fileName = "专业技术工作总结.docx";4 |
| | | String filePath = "/usr/local/project/file/"; |
| | | |
| | | String s = wordToPdf(url, filePath, fileName); |
| | | String s = wordToPdf(filePath, fileName); |
| | | System.err.println(s); |
| | | |
| | | return s; |