fix(aihr): sanitize multipart filenames before storage

2026-07-14 21:34:45 +08:00
parent 39febf87b2
commit dd5c0dadff
8 changed files with 126 additions and 8 deletions
@@ -6,6 +6,7 @@ import lombok.extern.slf4j.Slf4j;
 import org.dromara.aihr.domain.AihrSpeechDto.AsrResponse;
 import org.dromara.aihr.domain.AihrSpeechDto.TtsRequest;
 import org.dromara.aihr.domain.AihrSpeechDto.TtsResponse;
+import org.dromara.aihr.service.AihrMultipartFiles;
 import org.dromara.aihr.service.AihrSpeechService;
 import org.dromara.common.core.domain.R;
 import org.dromara.system.domain.vo.SysOssVo;
@@ -52,10 +53,12 @@ public class AihrSpeechController {
        if (file.getSize() > MAX_ASR_BYTES) {
            return R.fail("音频过大（上限 5MB），请缩短录音时长");
        }
-        return speechService.transcribe(file.getBytes(), file.getOriginalFilename(), file.getContentType())
+        String fileName = AihrMultipartFiles.sanitizeFileName(file.getOriginalFilename(), "audio.webm");
+        MultipartFile sanitizedFile = AihrMultipartFiles.withOriginalFilename(file, fileName);
+        return speechService.transcribe(sanitizedFile.getBytes(), fileName, sanitizedFile.getContentType())
            .map(text -> {
                try {
-                    SysOssVo oss = ossService.upload(file);
+                    SysOssVo oss = ossService.upload(sanitizedFile);
                    return R.ok(new AsrResponse(text, "openai-compatible", oss.getOssId(), oss.getUrl()));
                } catch (Exception e) {
                    log.warn("asr audio upload failed（处理错误已隐藏）");
@@ -50,7 +50,7 @@ public class AihrCandidateMaterialService {
            throw new IllegalArgumentException("候选人资料仅支持 PDF、Word 或 JPG/PNG/WebP 图片");
        }
        String type = clean(materialType, materialType(fileName));
-        SysOssVo oss = ossService.upload(file);
+        SysOssVo oss = ossService.upload(AihrMultipartFiles.withOriginalFilename(file, fileName));
        KeyHolder keyHolder = new GeneratedKeyHolder();
        jdbcTemplate.update(connection -> {
            PreparedStatement ps = connection.prepareStatement("""
@@ -67,7 +67,7 @@ public class AihrCaseService {
            String transcript = speechService.transcribe(file.getBytes(), fileName, contentType)
                .orElseThrow(() -> new IllegalStateException("ASR 未配置或转写失败"));
            transcript = maskSensitiveText(transcript);
-            MediaRef media = storeSourceAudio(file);
+            MediaRef media = storeSourceAudio(file, fileName);
            String caseId = "case-" + UUID.randomUUID();
            CaseState state = new CaseState(caseId, fileName, scopedProject, transcript, tagsFromText(transcript), null,
                media.ossId(), media.url(), currentCaseOwner());
@@ -440,12 +440,12 @@ public class AihrCaseService {
        return String.join(",", Collections.nCopies(size, "?"));
    }

-    private MediaRef storeSourceAudio(MultipartFile file) {
+    private MediaRef storeSourceAudio(MultipartFile file, String fileName) {
        // Unit tests can construct this service without the system OSS bean; production always wires it.
        if (ossService == null) {
            return new MediaRef(null, "");
        }
-        SysOssVo oss = ossService.upload(file);
+        SysOssVo oss = ossService.upload(AihrMultipartFiles.withOriginalFilename(file, fileName));
        return new MediaRef(oss.getOssId(), firstNonBlank(oss.getUrl(), ""));
    }

@@ -0,0 +1,75 @@
+package org.dromara.aihr.service;
+
+import org.springframework.web.multipart.MultipartFile;
+
+import java.io.File;
+import java.io.IOException;
+import java.io.InputStream;
+
+/**
+ * Multipart 文件边界工具：清理后的文件名必须继续贯穿到 OSS 元数据和外部请求。
+ */
+public final class AihrMultipartFiles {
+
+    private AihrMultipartFiles() {
+    }
+
+    public static MultipartFile withOriginalFilename(MultipartFile delegate, String originalFilename) {
+        if (delegate == null) {
+            throw new IllegalArgumentException("文件不能为空");
+        }
+        String safeName = sanitizeFileName(originalFilename, "upload.bin");
+        return new MultipartFile() {
+            @Override
+            public String getName() {
+                return delegate.getName();
+            }
+
+            @Override
+            public String getOriginalFilename() {
+                return safeName;
+            }
+
+            @Override
+            public String getContentType() {
+                return delegate.getContentType();
+            }
+
+            @Override
+            public boolean isEmpty() {
+                return delegate.isEmpty();
+            }
+
+            @Override
+            public long getSize() {
+                return delegate.getSize();
+            }
+
+            @Override
+            public byte[] getBytes() throws IOException {
+                return delegate.getBytes();
+            }
+
+            @Override
+            public InputStream getInputStream() throws IOException {
+                return delegate.getInputStream();
+            }
+
+            @Override
+            public void transferTo(File dest) throws IOException, IllegalStateException {
+                delegate.transferTo(dest);
+            }
+        };
+    }
+
+    public static String sanitizeFileName(String originalFilename, String fallback) {
+        String name = originalFilename == null || originalFilename.isBlank()
+            ? fallback
+            : originalFilename.trim();
+        name = name.replaceAll("[\\\\/\\r\\n\\t]", "_");
+        if (name.length() > 200) {
+            name = name.substring(name.length() - 200);
+        }
+        return name.isBlank() ? fallback : name;
+    }
+}
@@ -389,7 +389,7 @@ public class AihrSopSeedService {
        if (file == null || file.isEmpty()) {
            throw new ServiceException("上传文件不能为空");
        }
-        String fileName = Optional.ofNullable(file.getOriginalFilename()).orElse("knowledge.txt").trim();
+        String fileName = AihrMultipartFiles.sanitizeFileName(file.getOriginalFilename(), "knowledge.txt");
        if (!supportedFile(fileName)) {
            throw new ServiceException("仅支持 txt/md/markdown/pdf/doc/docx/xls/xlsx/ppt/pptx/图片文件");
        }
@@ -397,7 +397,8 @@ public class AihrSopSeedService {
            throw new ServiceException("文件不能超过 100MB");
        }

-        return saveDocument(fileName, category, () -> ossService.upload(file), () -> readContent(file, fileName), () -> fileFingerprint(file));
+        MultipartFile sanitizedFile = AihrMultipartFiles.withOriginalFilename(file, fileName);
+        return saveDocument(fileName, category, () -> ossService.upload(sanitizedFile), () -> readContent(sanitizedFile, fileName), () -> fileFingerprint(sanitizedFile));
    }

    /**
@@ -0,0 +1,32 @@
+package org.dromara.aihr.service;
+
+import org.junit.jupiter.api.Tag;
+import org.junit.jupiter.api.Test;
+import org.springframework.mock.web.MockMultipartFile;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+
+class AihrMultipartFilesTest {
+
+    @Test
+    @Tag("dev")
+    void sanitizedNameMustReachTheMultipartPassedToStorage() {
+        MockMultipartFile source = new MockMultipartFile(
+            "file", "../../case\r\n.webm", "audio/webm", new byte[]{1, 2, 3}
+        );
+
+        var sanitized = AihrMultipartFiles.withOriginalFilename(source, AihrMultipartFiles.sanitizeFileName(
+            source.getOriginalFilename(), "audio.webm"
+        ));
+
+        assertEquals(".._.._case__.webm", sanitized.getOriginalFilename());
+        assertEquals(source.getSize(), sanitized.getSize());
+        assertEquals("audio/webm", sanitized.getContentType());
+    }
+
+    @Test
+    @Tag("dev")
+    void blankNameUsesSafeFallback() {
+        assertEquals("upload.bin", AihrMultipartFiles.sanitizeFileName(" \t", "upload.bin"));
+    }
+}
@@ -1,5 +1,7 @@
 # BRD 功能审查与未完成项记录

+- 2026-07-14 BRD G3 文件名边界复核：候选资料、案例音频、同步 SOP 文档和 ASR 音频此前虽然分别校验或清理文件名，但仍有路径把原始 `MultipartFile` 直接交给 `sys_oss`，清理后的名字没有贯通到 OSS 元数据；现新增共用 `AihrMultipartFiles` 包装器，四条链路均以清理后的文件名上传/外发，并补单测与 `demo-check` marker。该修复不改变文件类型、大小、权限和业务状态，仍需正式环境抽查 OSS 原始文件名与下载响应头。
+
 更新时间：2026-07-14

 - 2026-07-14 21:28 BRD 发布只读复核：生产根站、`/h5/` 和 `/prod-api/auth/tenant/list` 均返回 `200`；线上管理端仍加载 `assets/index-CJZ3Ax3Z.js`（SHA-256 `8b9278a26ccb760abce489b12a20b221748c64d4c4153613d2511777bb9677bc`），H5 仍加载 `assets/index-D4-NrEpb.js`（SHA-256 `ab15bfd17cabe58e2f34b0ac61ee198a837128ed85dc6bb0ddf77bc3e22dcd9c`）。本地最新提交 `d2b0035e` 的租户隔离修复尚未发布；本轮仅做 GET/hash 核验，未执行生产静态同步、后端重启或业务数据写入。
@@ -625,6 +625,11 @@ contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/control
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "仅支持 txt/md/markdown/pdf/doc/docx/xls/xlsx/ppt/pptx/图片文件"
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "/embeddings"
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "aihr_knowledge_gap"
+contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrMultipartFiles.java "withOriginalFilename"
+contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrCandidateMaterialService.java "AihrMultipartFiles.withOriginalFilename"
+contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrCaseService.java "AihrMultipartFiles.withOriginalFilename"
+contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "AihrMultipartFiles.withOriginalFilename"
+contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/controller/AihrSpeechController.java "AihrMultipartFiles.withOriginalFilename"
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "o.tenant_id = a.tenant_id"
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "where tenant_id = ? and oss_id = ? limit 1"
 contains backend/ruoyi-modules/ruoyi-aihr/src/main/java/org/dromara/aihr/service/AihrSopSeedService.java "deleteQdrantTenantPoints()"