Preserve exact original download filenames

backend/record_metadata.py

@@ -172,6 +172,9 @@ def reader_resolved_record(record: dict, asset: dict) -> dict:
     return {
         "url": source, "download": source, "extension": extension,
         "original_extension": extension, "title": record.get("File") or record.get("Name") or "在线阅读",
+        # Keep the exact indexed filename separate from the display title.
+        # The stem may itself end in an extension, or include trailing spaces.
+        "filename": build_record_relative_path(record).rsplit("/", 1)[-1],
         "path": "/".join([repo, *folders]), "folder": "/".join(folders), "repo": repo,
         **reader_asset_fields(asset, extension),
     }

static/app.js

@@ -2546,7 +2546,7 @@ function getReaderLink(rec, returnUrl = location.href, cacheMetadata = true) {
     readerRecord.OcrUrl = `/txt/${encodeRecordPath(VoiceOfMLReader.txtRelativePath(relPath))}`;
   }
   const readerUrl = VoiceOfMLReader.readerUrl(readerRecord, "/static/reader.html");
-  if (cacheMetadata) try { const readerId = new URL(readerUrl, location.origin).searchParams.get("id"); if (readerId) { const sourceData = { url: readerRecord.ReaderLink || readerRecord.Link, download: readerRecord.Link, title: readerRecord.File, extension: readerRecord.ReaderExtension || readerRecord.Extension, original_extension: readerRecord.Extension, repo: String(readerRecord.Repo || "").split("/").pop(), folder: readerRecord.Folder, chapter_manifest: readerRecord.ReaderChapterManifest || "", pdf_document: readerRecord.ReaderPdfDocument || "" }; cacheReaderMetadata(readerId, sourceData, true); } } catch (_) {}
+  if (cacheMetadata) try { const readerId = new URL(readerUrl, location.origin).searchParams.get("id"); if (readerId) { const sourceData = { url: readerRecord.ReaderLink || readerRecord.Link, download: readerRecord.Link, title: readerRecord.File, filename: String(readerRecord.File || "") + (readerRecord.Extension ? "." + readerRecord.Extension : ""), extension: readerRecord.ReaderExtension || readerRecord.Extension, original_extension: readerRecord.Extension, repo: String(readerRecord.Repo || "").split("/").pop(), folder: readerRecord.Folder, chapter_manifest: readerRecord.ReaderChapterManifest || "", pdf_document: readerRecord.ReaderPdfDocument || "" }; cacheReaderMetadata(readerId, sourceData, true); } } catch (_) {}
   return readerUrl;
 }
 

static/reader/core/reader-source.mjs

@@ -18,8 +18,16 @@ export function createReaderSource(readerContext) {
     return path;
   }
 
+  function readerOriginalFilename(fallback) {
+    // The human-readable title can omit a repeated suffix or normalize spaces;
+    // the authoritative metadata keeps the exact source filename separately.
+    const name = readerContext.resolvedReaderData?.filename;
+    return typeof name === "string" && name.trim() && !/[\/\\\x00-\x1f\x7f]/.test(name)
+      ? name : fallback;
+  }
+
   function readerDownloadUrl(filename, link) {
-    return readerContext.readerDownloadApiUrl(`/api/download?file=${encodeURIComponent(filename || "file")}&link=${encodeURIComponent(link || "")}`);
+    return readerContext.readerDownloadApiUrl(`/api/download?file=${encodeURIComponent(readerOriginalFilename(filename || "file"))}&link=${encodeURIComponent(link || "")}`);
   }
 
   async function startReaderDownload(event) {

tests/test_fastapi_endpoints.py

@@ -409,9 +409,17 @@ class FastApiEndpointTests(unittest.TestCase):
             response=self.client.get('/api/reader-resolve',params={'id':'2w2m8oowhjtl9'})
         self.assertEqual(response.status_code,200)
         self.assertTrue(response.json()['download'].endswith('/www.maopost.com_postl_0566_001M.jpg.jpg'))
+        self.assertEqual(response.json()['filename'],'www.maopost.com_postl_0566_001M.jpg.jpg')
         self.assertEqual(response.json()['extension'],'webp')
         self.assertIn(asset['p'],response.json()['url'])
 
+    def test_resolved_filename_preserves_case_spaces_and_repeated_extension(self):
+        for stem, extension in [('poster.jpg','jpg'),('Lecture','PDF'),('Book ','pdf'),('README','')]:
+            with self.subTest(stem=stem,extension=extension):
+                record={'Repo':'VoiceOfML/Test','File':stem,'Extension':extension,'Folder':['docs']}
+                data=app.reader_resolved_record(record,{})
+                self.assertEqual(data['filename'],stem+('.'+extension if extension else ''))
+
     def test_legacy_double_extension_alias_collision_fails_closed(self):
         app.records=[{"Repo":"VoiceOfML/Test", "File":name, "Extension":"jpg", "Folder":[]}
                      for name in ('one.jpg','two.jpg')]

tests/test_reader_navigation_host.py

@@ -110,6 +110,14 @@ process.stdout.write(stripBundledScripts(readFileSync('static/index.html','utf8'
             self.assertTrue(row['link'].endswith('/目录/'+row['filename']))
         self.assertTrue(result[0]['reader'])
 
+    def test_session_handoff_keeps_exact_original_filename(self):
+        result=self.page.evaluate('''()=>{
+          const record={Repo:'VoiceOfML/Test',Folder:['docs'],File:'poster.jpg',Extension:'JPG'};
+          const url=getReaderLink(record),id=new URL(url,location.href).searchParams.get('id');
+          return JSON.parse(sessionStorage.getItem('reader-source:'+id));
+        }''')
+        self.assertEqual(result['filename'],'poster.jpg.JPG')
+
     def test_explicit_close_ignores_late_panel_state_before_history_pop(self):
         self.page.evaluate('navigateToReader("/static/reader.html?ext=txt")')
         self.page.frame_locator('.reader-overlay').locator('body').wait_for()

tests/test_reader_pdf_loading.py

@@ -60,6 +60,48 @@ class PdfLoadingTests(unittest.TestCase):
         self.assertEqual(len(self.modules), 1)
         self.assertNotIn('reader-module-retry', self.modules[0])
 
+    def test_resolved_original_download_name_keeps_exact_repeated_extension(self):
+        original='https://huggingface.co/datasets/VoiceOfML/Test/resolve/main/poster.jpg.jpg'
+        image=b'GIF89a\x01\x00\x01\x00\x80\x00\x00\x00\x00\x00\xff\xff\xff!\xf9\x04\x01\x00\x00\x00\x00,\x00\x00\x00\x00\x01\x00\x01\x00\x00\x02\x02D\x01\x00;'
+        self.page.route('**/api/reader-resolve?id=filename-test',lambda route:route.fulfill(json={
+            'url':original,'download':original,'extension':'jpg','original_extension':'jpg',
+            'title':'poster.jpg','filename':'poster.jpg.jpg'}))
+        self.page.route('**/api/reader-content**',lambda route:route.fulfill(content_type='image/gif',body=image))
+        self.page.goto(self.origin+'/static/reader.html?id=filename-test&ext=jpg',wait_until='domcontentloaded')
+        self.page.locator('.reader-image').wait_for()
+        query=urllib.parse.parse_qs(urllib.parse.urlsplit(self.page.locator('#download').get_attribute('href')).query)
+        self.assertEqual(query['file'],['poster.jpg.jpg'])
+        self.assertEqual(query['link'],[original])
+
+    def test_malformed_original_filename_metadata_is_not_used_for_download(self):
+        source='https://huggingface.co/datasets/VoiceOfML/Test/resolve/main/safe.txt'
+        self.page.route('**/api/reader-resolve?id=filename-invalid',lambda route:route.fulfill(json={
+            'url':source,'download':source,'extension':'txt','original_extension':'txt',
+            'title':'safe','filename':'../unsafe.txt'}))
+        self.page.route('**/api/reader-content**',lambda route:route.fulfill(content_type='text/plain',body='safe'))
+        self.page.goto(self.origin+'/static/reader.html?id=filename-invalid&ext=txt',wait_until='domcontentloaded')
+        self.page.locator('.reader-text').wait_for()
+        query=urllib.parse.parse_qs(urllib.parse.urlsplit(self.page.locator('#download').get_attribute('href')).query)
+        self.assertEqual(query['file'],['safe.txt'])
+
+    def test_original_download_name_survives_title_change_and_button_handoff(self):
+        source='https://huggingface.co/datasets/VoiceOfML/Test/resolve/main/Sample%20.PDF'
+        checks=[]
+        self.page.route('**/api/reader-resolve?id=filename-click',lambda route:route.fulfill(json={
+            'url':source,'download':source,'extension':'txt','original_extension':'pdf',
+            'title':'Sample ','filename':'Sample .PDF'}))
+        self.page.route('**/api/reader-content**',lambda route:route.fulfill(content_type='text/plain',body='safe'))
+        self.page.route('**/api/download/check?**',lambda route:route.fulfill(json={'ok':True}))
+        self.page.route('**/api/download?**',lambda route:(checks.append(route.request.url),route.fulfill(body='file')))
+        self.page.goto(self.origin+'/static/reader.html?id=filename-click&ext=txt',wait_until='domcontentloaded')
+        self.page.locator('.reader-text').wait_for()
+        self.page.locator('#download').click()
+        self.page.wait_for_function("()=>document.querySelector('#download-feedback')?.textContent.includes('已发起下载')")
+        self.assertEqual(len(checks),1)
+        query=urllib.parse.parse_qs(urllib.parse.urlsplit(checks[0]).query)
+        self.assertEqual(query['file'],['Sample .PDF'])
+        self.assertEqual(query['link'],[source])
+
     def test_first_range_overlaps_held_engine_and_is_consumed_once(self):
         held = []
         self.page.route(MODULE, lambda route: held.append(route))