refactor(module-llm):调整文件处理逻辑

- 恢复了对 .doc 文件的处理逻辑,将其转换为 .docx
- 注释掉了对 .md 文件的处理逻辑- 优化了代码格式和缩进
This commit is contained in:
Liuyang 2025-03-01 20:58:40 +08:00
parent 6e19d81a72
commit 3fecb7e378

View File

@ -415,24 +415,24 @@ public class RagHttpService {
Path tempFilePath = downloadFileToTemp(fileUrl, fileName);
log.info("文件已下载到临时目录: {}", tempFilePath);
// String fileSuffix = getFileSuffix(fileName);
// if ("doc".equals(fileSuffix)) {
// log.info("正在处理 doc 文件");
// try {
// tempFilePath= converterDocToDocx(tempFilePath.toString(), tempFilePath.toString().replace(".doc", ".docx"));
// } catch (Exception e) {
// throw new RuntimeException(e);
// }
// }
String fileSuffix = getFileSuffix(fileName);
if ("doc".equals(fileSuffix)) {
log.info("正在处理 doc 文件");
try {
tempFilePath = converterDocToDocx(tempFilePath.toString(), tempFilePath.toString().replace(".doc", ".docx"));
} catch (Exception e) {
throw new RuntimeException(e);
}
}
// if ("md".equals(fileSuffix)) {
// log.info("正在处理 md 文件");
// try {
// tempFilePath= converterMdToTxt(tempFilePath.toString(), tempFilePath.toString().replace(".md", ".txt"));
// } catch (Exception e) {
// throw new RuntimeException(e);
// }
// }
// if ("md".equals(fileSuffix)) {
// log.info("正在处理 md 文件");
// try {
// tempFilePath= converterMdToTxt(tempFilePath.toString(), tempFilePath.toString().replace(".md", ".txt"));
// } catch (Exception e) {
// throw new RuntimeException(e);
// }
// }
// 创建 OkHttpClient 实例
log.info("创建 OkHttpClient 实例,设置超时时间为 3 分钟");
@ -626,7 +626,7 @@ public class RagHttpService {
return path;
}
public static Path converterDocToDocx(String inputPath, String outputPath) throws Exception {
public static Path converterDocToDocx (String inputPath, String outputPath) throws Exception {
// 读取DOC文档
try (HWPFDocument doc = new HWPFDocument(Files.newInputStream(Paths.get(inputPath)))) {
XWPFDocument docx = new XWPFDocument();
@ -650,6 +650,7 @@ public class RagHttpService {
return Paths.get(outputPath);
}
}
/**
* 处理响应结果
*/