Merge pull request 'feat(file): 新增文件哈希去重与文本提取功能- 在多个模块中引入 FileHashUtil 并用于文件上传前的哈希计算' (#19) from liwenlong/fys-Multi-tenant:jingquan into jingquan

Reviewed-on: #19

fys-admin/src/main/java/com/fuyuanshen/app/controller/AppVideoController.java

@@ -1,5 +1,6 @@
 package com.fuyuanshen.app.controller;
 
+import cn.dev33.satoken.annotation.SaIgnore;
 import com.fuyuanshen.app.service.AudioProcessService;
 import com.fuyuanshen.app.service.VideoProcessService;
 import com.fuyuanshen.common.core.domain.R;
@@ -51,4 +52,13 @@ public class AppVideoController extends BaseController {
     public R<List<String>> uploadAudioTTS(@RequestParam String text) throws IOException {
         return R.ok(audioProcessService.generateStandardPcmData(text));
     }
+
+    /**
+     * 提取文本内容（只支持txt/docx）
+     */
+    @PostMapping(value = "/extract", consumes = MediaType.MULTIPART_FORM_DATA_VALUE)
+    @RepeatSubmit(interval = 2, timeUnit = TimeUnit.SECONDS,message = "请勿重复提交！")
+    public R<String> extract(@RequestParam("file") MultipartFile file) throws Exception {
+        return R.ok("Success",audioProcessService.extract(file));
+    }
 }

fys-admin/src/main/java/com/fuyuanshen/app/service/AudioProcessService.java

@@ -7,11 +7,17 @@ import lombok.extern.slf4j.Slf4j;
 import org.springframework.stereotype.Service;
 import org.springframework.web.multipart.MultipartFile;
 
-import java.io.File;
-import java.io.IOException;
+import javax.xml.stream.XMLInputFactory;
+import javax.xml.stream.XMLStreamConstants;
+import javax.xml.stream.XMLStreamException;
+import javax.xml.stream.XMLStreamReader;
+import java.io.*;
+import java.nio.charset.StandardCharsets;
 import java.nio.file.Files;
 import java.util.Arrays;
 import java.util.List;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipInputStream;
 
 /**
  * 音频处理服务
@@ -170,5 +176,75 @@ public class AudioProcessService {
         }
     }
 
+    /**
+     * 提取文本
+     */
+    public String extract(MultipartFile file) throws Exception {
+        String name = file.getOriginalFilename();
+        if (name == null ||
+                (!name.endsWith(".txt") && !name.endsWith(".docx"))) {
+            throw new IllegalArgumentException("仅支持 .txt 或 .docx");
+        }
+        if (file.getSize() > MAX_AUDIO_SIZE) {
+            throw new IllegalArgumentException("文件超过5MB");
+        }
+
+        String text;
+        /* 全程流式，不落地磁盘，不一次性读字节数组 */
+        try (InputStream in = file.getInputStream()) {
+            if (name.endsWith(".txt")) {
+                text = readTxt(in);
+            } else {
+                text = readDocx(in);
+            }
+        }
+        return text;
+    }
+
+    /* ---------- txt：按行读，StringBuilder 复用 ---------- */
+    private String readTxt(InputStream in) throws IOException {
+        BufferedReader br = new BufferedReader(new InputStreamReader(in, StandardCharsets.UTF_8));
+        StringBuilder sb = new StringBuilder(4096);
+        String line;
+        while ((line = br.readLine()) != null) {
+            sb.append(line).append('\n');
+        }
+        return sb.toString();
+    }
+
+    /* ---------- docx：ZipInputStream 只扫 document.xml ---------- */
+    private String readDocx(InputStream in) throws IOException {
+        ZipInputStream zin = new ZipInputStream(in);
+        ZipEntry e;
+        while ((e = zin.getNextEntry()) != null) {
+            if ("word/document.xml".equals(e.getName())) {
+                return staxExtract(zin);   // 流式读 XML
+            }
+        }
+        return "";
+    }
+
+    /* ---------- StAX 流式提取 <w:t> ---------- */
+    private String staxExtract(InputStream xml) throws IOException {
+        XMLStreamReader r = null;
+        StringBuilder sb = new StringBuilder(4096);
+        try {
+            //System.out.println(new String(xml.readAllBytes()));
+            r = XMLInputFactory.newInstance().createXMLStreamReader(xml);
+            while (r.hasNext()) {
+                if (r.next() == XMLStreamConstants.START_ELEMENT &&
+                        "t".equals(r.getLocalName())) {
+                    String elementText = r.getElementText();
+                    sb.append(elementText);
+                }
+            }
+        } catch (XMLStreamException ex) {
+            throw new IOException(ex);
+        } finally {
+            if (r != null) try { r.close(); } catch (XMLStreamException ignore) {}
+        }
+        return sb.toString();
+    }
+
 
 }

fys-admin/src/main/java/com/fuyuanshen/web/service/device/DeviceDebugService.java

@@ -39,6 +39,7 @@ public class DeviceDebugService {
     private final IAppBusinessFileService appBusinessFileService;
     private final IAppOperationVideoService appOperationVideoService;
     private final DeviceService deviceService;
+    private final FileHashUtil fileHashUtil;
 
     /**
      * 文件上传并添加文件信息哈希去重
@@ -62,26 +63,12 @@ public class DeviceDebugService {
         Map<String, Long> hash2OssId = new LinkedHashMap<>(files.length);
         for (MultipartFile file : files) {
             // 1. 计算文件哈希
-            String hash = FileHashUtil.hash(file);
+            String hash = fileHashUtil.hash(file);
 
-            // 2. 先根据 hash 查库（秒传）
-            SysOssVo exist = sysOssService.selectByHash(hash);
-            Long ossId;
-            if (exist != null) {
-                // 2.1 已存在，直接复用
-                ossId = exist.getOssId();
-                hash2OssId.putIfAbsent(hash, ossId);
-            } else {
-                // 2.2 不存在，真正上传
-                SysOssVo upload = sysOssService.upload(file);
-                if (upload == null) {
-                    return false;
-                }
-                ossId = upload.getOssId();
-                hash2OssId.putIfAbsent(hash, ossId);
-                // 2.3 把 hash 写回记录（供下次去重）
-                sysOssService.updateHashById(ossId, hash);
-            }
+            SysOssVo exist = sysOssService.updateHash(file, hash);
+            // 2.1 已存在，直接复用
+            long ossId = exist.getOssId();
+            hash2OssId.putIfAbsent(hash, ossId);
         }
         // 4. 组装业务中间表
         List<AppBusinessFile> bizList = new ArrayList<>(bo.getDeviceIds().length * hash2OssId.size());

fys-modules/fys-equipment/src/main/java/com/fuyuanshen/equipment/service/impl/DeviceRepairRecordsServiceImpl.java

@@ -50,6 +50,7 @@ public class DeviceRepairRecordsServiceImpl extends ServiceImpl<DeviceRepairReco
     private final DeviceRepairRecordsMapper baseMapper;
     private final DeviceRepairImagesMapper imagesMapper;
     private final ISysOssService ossService;
+    private final FileHashUtil fileHashUtil;
 
     /**
      * 查询设备维修记录
@@ -210,19 +211,13 @@ public class DeviceRepairRecordsServiceImpl extends ServiceImpl<DeviceRepairReco
         // 1. 计算文件哈希
         String hash = null;
         try {
-            hash = FileHashUtil.hash(file);
+            hash = fileHashUtil.hash(file);
         } catch (IOException e) {
             throw new RuntimeException(e);
         }
 
         // 2. 先根据 hash 查库（秒传）
-        SysOssVo exist = ossService.selectByHash(hash);
-        if (exist == null) {
-            // 2.2 不存在，真正上传
-            exist = ossService.upload(file);
-            // 2.3 把 hash 写回记录（供下次去重）
-            ossService.updateHashById(exist.getOssId(), hash);
-        }
+        SysOssVo exist = ossService.updateHash(file,hash);
 
         DeviceRepairImages image = new DeviceRepairImages();
         image.setRecordId(recordId);

fys-modules/fys-equipment/src/main/java/com/fuyuanshen/equipment/service/impl/DeviceServiceImpl.java

@@ -34,6 +34,7 @@ import com.fuyuanshen.equipment.enums.CommunicationModeEnum;
 import com.fuyuanshen.equipment.enums.DeviceActiveStatusEnum;
 import com.fuyuanshen.equipment.mapper.*;
 import com.fuyuanshen.equipment.service.*;
+import com.fuyuanshen.equipment.utils.FileHashUtil;
 import com.fuyuanshen.system.domain.vo.SysOssVo;
 import com.fuyuanshen.system.domain.vo.SysRoleVo;
 import com.fuyuanshen.system.service.ISysOssService;
@@ -84,6 +85,7 @@ public class DeviceServiceImpl extends ServiceImpl<DeviceMapper, Device> impleme
     private final DeviceTypeGrantsMapper deviceTypeGrantsMapper;
 
     private final DeviceFenceAccessRecordMapper deviceFenceAccessRecordMapper;
+    private final FileHashUtil fileHashUtil;
 
 
     /**
@@ -209,7 +211,8 @@ public class DeviceServiceImpl extends ServiceImpl<DeviceMapper, Device> impleme
 
         // 保存图片并获取URL
         if (deviceForm.getFile() != null) {
-            SysOssVo upload = ossService.upload(deviceForm.getFile());
+            String fileHash = fileHashUtil.hash(deviceForm.getFile());
+            SysOssVo upload = ossService.updateHash(deviceForm.getFile(),fileHash);
             // 设置图片路径
             deviceForm.setDevicePic(upload.getUrl());
         }
@@ -283,8 +286,8 @@ public class DeviceServiceImpl extends ServiceImpl<DeviceMapper, Device> impleme
 
         // 处理上传的图片
         if (deviceForm.getFile() != null) {
-            // 设置图片路径
-            SysOssVo oss = ossService.upload(deviceForm.getFile());
+            String fileHash = fileHashUtil.hash(deviceForm.getFile());
+            SysOssVo oss = ossService.updateHash(deviceForm.getFile(),fileHash);
             // 强制将HTTP替换为HTTPS
             if (oss.getUrl() != null && oss.getUrl().startsWith("http://")) {
                 oss.setUrl(oss.getUrl().replaceFirst("^http://", "https://"));

fys-modules/fys-equipment/src/main/java/com/fuyuanshen/equipment/utils/FileHashUtil.java

@@ -1,28 +1,70 @@
 package com.fuyuanshen.equipment.utils;
 
 import org.apache.commons.codec.digest.DigestUtils;
+import org.springframework.stereotype.Component;
 import org.springframework.web.multipart.MultipartFile;
 
 import java.io.IOException;
 import java.io.InputStream;
 import java.security.MessageDigest;
+import java.security.NoSuchAlgorithmException;
 import java.util.HexFormat;
 
 /**
  * 文件哈希工具类
  */
+@Component  // 如果使用 Spring 可以注入
 public class FileHashUtil {
+    /* 算法常量 */
     private static final String ALGORITHM = "SHA-256";
+    /* 缓冲区大小 8 KB */
+    private static final int BUFFER_SIZE = 8192;
 
-    public static String hash(MultipartFile file) throws IOException {
-        MessageDigest digest = DigestUtils.getDigest(ALGORITHM);
+    /**
+     * 计算上传文件的 SHA-256 十六进制哈希
+     *
+     * @param file 上传文件；不能为 null，且必须非空
+     * @return 64 位小写十六进制字符串
+     * @throws IllegalArgumentException 参数不合法
+     * @throws IOException               流读取失败
+     * @throws IllegalStateException     算法运行时异常（不会触发）
+     */
+    public String hash(MultipartFile file) throws IOException {
+        validate(file);
+
+        /* 每个请求新建实例，保证线程安全 */
+        MessageDigest digest = newDigest();
+
+        /* try-with-resources 自动关闭流 */
         try (InputStream in = file.getInputStream()) {
-            byte[] buf = new byte[8192];
+            byte[] buf = new byte[BUFFER_SIZE];
             int len;
             while ((len = in.read(buf)) != -1) {
                 digest.update(buf, 0, len);
             }
         }
+
+        /* JDK 17+ 的 HexFormat，比 Apache Commons 更快且无需额外依赖 */
         return HexFormat.of().formatHex(digest.digest());
     }
+
+    /* -------------------- 私有辅助方法 -------------------- */
+
+    private static void validate(MultipartFile file) {
+        if (file == null) {
+            throw new IllegalArgumentException("MultipartFile 不能为 null");
+        }
+        if (file.isEmpty()) {
+            throw new IllegalArgumentException("上传文件不能为空");
+        }
+    }
+
+    private static MessageDigest newDigest() {
+        try {
+            return MessageDigest.getInstance(ALGORITHM);
+        } catch (NoSuchAlgorithmException e) {
+            /* SHA-256 是 JDK 必现算法，走到这里说明 JDK 实现损坏 */
+            throw new IllegalStateException("算法 " + ALGORITHM + " 不可用", e);
+        }
+    }
 }

fys-modules/fys-system/src/main/java/com/fuyuanshen/system/service/ISysOssService.java

@@ -60,6 +60,14 @@ public interface ISysOssService {
      */
     int updateHashById(long ossId,String fileHash);
 
+    /**
+     * 更新文件 hash 值
+     *
+     * @param file 文件对象
+     * @return 匹配的 SysOssVo 列表
+     */
+    SysOssVo updateHash(MultipartFile file, String hash);
+
     /**
      * 上传 MultipartFile 到对象存储服务，并保存文件信息到数据库
      *

fys-modules/fys-system/src/main/java/com/fuyuanshen/system/service/impl/SysOssServiceImpl.java

@@ -191,6 +191,32 @@ public class SysOssServiceImpl implements ISysOssService, OssService {
         storage.download(sysOss.getFileName(), response.getOutputStream(), response::setContentLengthLong);
     }
 
+    /**
+     * 上传 MultipartFile 到对象存储服务，并保存文件信息到数据库
+     *
+     * @param file 要上传的 MultipartFile 对象
+     * @return 保存到数据库的 SysOssVo 对象
+     */
+    @Override
+    public SysOssVo updateHash(MultipartFile file, String hash) {
+        // 2. 先根据 hash 查库（秒传）
+        SysOssVo exist = baseMapper.selectByHash(hash);
+        if (exist != null) {
+            return exist;
+        }
+        String originalfileName = file.getOriginalFilename();
+        String suffix = StringUtils.substring(originalfileName, originalfileName.lastIndexOf("."), originalfileName.length());
+        OssClient storage = OssFactory.instance();
+        UploadResult uploadResult;
+        try {
+            uploadResult = storage.uploadSuffix(file.getBytes(), suffix, file.getContentType());
+        } catch (IOException e) {
+            throw new ServiceException(e.getMessage());
+        }
+        // 保存文件信息
+        return buildResultEntity(originalfileName, suffix, storage.getConfigKey(), uploadResult,hash);
+    }
+
     /**
      * 上传 MultipartFile 到对象存储服务，并保存文件信息到数据库
      *
@@ -210,7 +236,7 @@ public class SysOssServiceImpl implements ISysOssService, OssService {
             throw new ServiceException(e.getMessage());
         }
         // 保存文件信息
-        return buildResultEntity(originalfileName, suffix, storage.getConfigKey(), uploadResult);
+        return buildResultEntity(originalfileName, suffix, storage.getConfigKey(), uploadResult,null);
     }
 
     /**
@@ -226,7 +252,7 @@ public class SysOssServiceImpl implements ISysOssService, OssService {
         OssClient storage = OssFactory.instance();
         UploadResult uploadResult = storage.uploadSuffix(file, suffix);
         // 保存文件信息
-        return buildResultEntity(originalfileName, suffix, storage.getConfigKey(), uploadResult);
+        return buildResultEntity(originalfileName, suffix, storage.getConfigKey(), uploadResult,null);
     }
 
 
@@ -255,18 +281,19 @@ public class SysOssServiceImpl implements ISysOssService, OssService {
         uploadResult = storage.uploadSuffix(data, suffix, "image/jpeg"); // 假设是图片类型，可以根据实际需要修改
 
         // 保存文件信息
-        return buildResultEntity(fileName, suffix, storage.getConfigKey(), uploadResult);
+        return buildResultEntity(fileName, suffix, storage.getConfigKey(), uploadResult,null);
     }
 
 
     @NotNull
-    private SysOssVo buildResultEntity(String originalfileName, String suffix, String configKey, UploadResult uploadResult) {
+    private SysOssVo buildResultEntity(String originalfileName, String suffix, String configKey, UploadResult uploadResult, String hash) {
         SysOss oss = new SysOss();
         oss.setUrl(uploadResult.getUrl());
         oss.setFileSuffix(suffix);
         oss.setFileName(uploadResult.getFilename());
         oss.setOriginalName(originalfileName);
         oss.setService(configKey);
+        oss.setFileHash(hash); // 设置哈希值
         baseMapper.insert(oss);
         SysOssVo sysOssVo = MapstructUtils.convert(oss, SysOssVo.class);
         return this.matchingUrl(sysOssVo);