fix: streamline document parsing settings

src/main/document-ocr-broker.test.ts

@@ -50,4 +50,93 @@ describe('DocumentOcrBroker', () => {
     ).toThrow('OCR 解析已取消')
     broker.dispose()
   })
+
+  it('rejects requests whose selected OCR pages exceed the limit', () => {
+    const broker = new DocumentOcrBroker({
+      isDestroyed: vi.fn(() => false),
+      webContents: { send: vi.fn() }
+    } as never)
+
+    expect(() =>
+      broker.recognize({
+        ...request(),
+        maximumPages: 1,
+        pageNumbers: [1, 2]
+      })
+    ).toThrow('OCR 页数超过当前文档限制')
+    broker.dispose()
+  })
+
+  it('queues OCR requests and starts timeout accounting on dispatch', async () => {
+    const send = vi.fn()
+    const broker = new DocumentOcrBroker({
+      isDestroyed: vi.fn(() => false),
+      webContents: { send }
+    } as never)
+    const first = broker.recognize(request())
+    const second = broker.recognize({
+      ...request(),
+      fileName: 'second.pdf'
+    })
+    const requests = send.mock.calls.filter(
+      ([channel]) => channel === ipcChannels.documentParsingOcrRequest
+    )
+
+    expect(requests).toHaveLength(1)
+    const firstRequest = requests[0]?.[1] as {
+      requestId: string
+    }
+    broker.respond({
+      requestId: firstRequest.requestId,
+      sections: [],
+      pageCount: 1,
+      warnings: []
+    })
+    await expect(first).resolves.toEqual(
+      expect.objectContaining({ requestId: firstRequest.requestId })
+    )
+
+    const dispatched = send.mock.calls.filter(
+      ([channel]) => channel === ipcChannels.documentParsingOcrRequest
+    )
+    expect(dispatched).toHaveLength(2)
+    const secondRequest = dispatched[1]?.[1] as {
+      requestId: string
+    }
+    broker.respond({
+      requestId: secondRequest.requestId,
+      sections: [],
+      pageCount: 1,
+      warnings: []
+    })
+    await expect(second).resolves.toEqual(
+      expect.objectContaining({ requestId: secondRequest.requestId })
+    )
+    broker.dispose()
+  })
+
+  it('cancels a queued request without interrupting the active request', async () => {
+    const send = vi.fn()
+    const broker = new DocumentOcrBroker({
+      isDestroyed: vi.fn(() => false),
+      webContents: { send }
+    } as never)
+    const active = broker.recognize(request())
+    const controller = new AbortController()
+    const queued = broker.recognize(
+      { ...request(), fileName: 'queued.pdf' },
+      controller.signal
+    )
+
+    controller.abort()
+
+    await expect(queued).rejects.toThrow('OCR 解析已取消')
+    expect(
+      send.mock.calls.filter(
+        ([channel]) => channel === ipcChannels.documentParsingOcrCancel
+      )
+    ).toHaveLength(0)
+    broker.dispose()
+    await expect(active).rejects.toThrow('OCR 解析已取消')
+  })
 })

src/main/document-ocr-broker.ts

@@ -9,17 +9,23 @@ import {
 } from '../shared/document-parsing-contracts'
 
 type PendingRequest = {
+  request: DocumentOcrRequest
   resolve: (result: DocumentOcrResult) => void
   reject: (error: Error) => void
-  timer: ReturnType<typeof setTimeout>
+  timer?: ReturnType<typeof setTimeout>
+  timeoutMs: number
   detachAbort: () => void
+  dispatched: boolean
 }
 
 const maximumPendingRequests = 4
 const maximumTotalTimeoutMs = 10 * 60 * 1_000
+const workerStartupTimeoutMs = 60 * 1_000
 
 export class DocumentOcrBroker {
   private readonly pending = new Map<string, PendingRequest>()
+  private readonly queue: string[] = []
+  private activeRequestId?: string
   private disposed = false
 
   constructor(private readonly window: BrowserWindow) {}
@@ -41,50 +47,37 @@ export class DocumentOcrBroker {
     if (signal?.aborted) {
       throw new Error('OCR 解析已取消')
     }
+    const pageCount =
+      request.pageNumbers?.length ?? request.maximumPages
     const timeoutMs = Math.min(
       maximumTotalTimeoutMs,
       Math.max(
         request.pageTimeoutSeconds * 1_000,
-        request.pageTimeoutSeconds *
+        workerStartupTimeoutMs +
-          request.maximumPages *
+          request.pageTimeoutSeconds *
-          1_000
+            pageCount *
+            1_000
       )
     )
     return new Promise<DocumentOcrResult>((resolve, reject) => {
-      const cancel = (message: string): void => {
+      const onAbort = (): void =>
-        const pending = this.pending.get(request.requestId)
+        this.cancelRequest(request.requestId, 'OCR 解析已取消')
-        if (!pending) {
-          return
-        }
-        clearTimeout(pending.timer)
-        pending.detachAbort()
-        this.pending.delete(request.requestId)
-        this.window.webContents.send(
-          ipcChannels.documentParsingOcrCancel,
-          request.requestId
-        )
-        reject(new Error(message))
-      }
-      const timer = setTimeout(() => {
-        cancel('OCR 解析超时')
-      }, timeoutMs)
-      const onAbort = (): void => cancel('OCR 解析已取消')
       signal?.addEventListener('abort', onAbort, { once: true })
       this.pending.set(request.requestId, {
+        request,
         resolve,
         reject,
-        timer,
+        timeoutMs,
         detachAbort: () =>
-          signal?.removeEventListener('abort', onAbort)
+          signal?.removeEventListener('abort', onAbort),
+        dispatched: false
       })
+      this.queue.push(request.requestId)
       if (signal?.aborted) {
-        cancel('OCR 解析已取消')
+        this.cancelRequest(request.requestId, 'OCR 解析已取消')
         return
       }
-      this.window.webContents.send(
+      this.dispatchNext()
-        ipcChannels.documentParsingOcrRequest,
-        request
-      )
     })
   }
 
@@ -102,30 +95,111 @@ export class DocumentOcrBroker {
       throw new Error('OCR 响应无效')
     }
     const pending = this.pending.get(requestId)
-    if (!pending) {
+    if (!pending || !pending.dispatched) {
       return
     }
-    clearTimeout(pending.timer)
-    pending.detachAbort()
-    this.pending.delete(requestId)
     if (result.success) {
-      pending.resolve(result.data)
+      this.finishRequest(requestId, () =>
+        pending.resolve(result.data)
+      )
     } else {
       if (!failure?.success) {
-        pending.reject(new Error('OCR 响应无效'))
+        this.finishRequest(requestId, () =>
+          pending.reject(new Error('OCR 响应无效'))
+        )
         return
       }
-      pending.reject(new Error(failure.data.error))
+      this.finishRequest(requestId, () =>
+        pending.reject(new Error(failure.data.error))
+      )
     }
   }
 
   dispose(): void {
     this.disposed = true
     for (const pending of this.pending.values()) {
-      clearTimeout(pending.timer)
+      if (pending.timer) {
+        clearTimeout(pending.timer)
+      }
       pending.detachAbort()
       pending.reject(new Error('OCR 解析已取消'))
     }
     this.pending.clear()
+    this.queue.length = 0
+    this.activeRequestId = undefined
+  }
+
+  private dispatchNext(): void {
+    if (
+      this.disposed ||
+      this.activeRequestId ||
+      this.window.isDestroyed()
+    ) {
+      return
+    }
+    let requestId = this.queue.shift()
+    while (requestId && !this.pending.has(requestId)) {
+      requestId = this.queue.shift()
+    }
+    if (!requestId) {
+      return
+    }
+    const pending = this.pending.get(requestId)
+    if (!pending) {
+      return
+    }
+    this.activeRequestId = requestId
+    pending.dispatched = true
+    pending.timer = setTimeout(() => {
+      this.cancelRequest(requestId, 'OCR 解析超时')
+    }, pending.timeoutMs)
+    try {
+      this.window.webContents.send(
+        ipcChannels.documentParsingOcrRequest,
+        pending.request
+      )
+    } catch (error) {
+      const detail =
+        error instanceof Error ? error.message : 'OCR 渲染服务不可用'
+      this.finishRequest(requestId, () =>
+        pending.reject(new Error(detail))
+      )
+    }
+  }
+
+  private cancelRequest(requestId: string, message: string): void {
+    const pending = this.pending.get(requestId)
+    if (!pending) {
+      return
+    }
+    if (pending.dispatched && !this.window.isDestroyed()) {
+      this.window.webContents.send(
+        ipcChannels.documentParsingOcrCancel,
+        requestId
+      )
+    }
+    this.finishRequest(requestId, () =>
+      pending.reject(new Error(message))
+    )
+  }
+
+  private finishRequest(
+    requestId: string,
+    settle: () => void
+  ): void {
+    const pending = this.pending.get(requestId)
+    if (!pending) {
+      return
+    }
+    if (pending.timer) {
+      clearTimeout(pending.timer)
+    }
+    pending.detachAbort()
+    this.pending.delete(requestId)
+    if (this.activeRequestId === requestId) {
+      this.activeRequestId = undefined
+    }
+    settle()
+    this.dispatchNext()
   }
 }

src/main/document-ocr-model-manager.test.ts

@@ -163,6 +163,17 @@ afterEach(async () => {
 })
 
 describe('DocumentOcrModelManager', () => {
+  it('reports a removed catalog model as unavailable', async () => {
+    const { manager } = await createManager()
+
+    await expect(manager.getStatus('retired-model')).resolves.toMatchObject({
+      id: 'retired-model',
+      available: false,
+      verified: false,
+      detail: expect.stringContaining('不再提供')
+    })
+  })
+
   it('uses immutable SHA-256 verified ModelScope catalog files', () => {
     expect(DOCUMENT_OCR_MODEL_CATALOG).toHaveLength(3)
     expect(

src/main/document-ocr-model-manager.ts

@@ -205,7 +205,18 @@ export class DocumentOcrModelManager {
   async getStatus(
     modelId: string
   ): Promise<ReturnType<typeof documentParsingModelStatusSchema.parse>> {
-    const entry = this.requireCatalogEntry(modelId)
+    const id = localOcrModelIdSchema.parse(modelId)
+    const entry = this.catalog.find((candidate) => candidate.id === id)
+    if (!entry) {
+      return documentParsingModelStatusSchema.parse({
+        id,
+        displayName: id,
+        available: false,
+        verified: false,
+        runtime: 'onnxruntime-web-wasm',
+        detail: '当前版本不再提供此 OCR 模型，请选择其他模型'
+      })
+    }
     try {
       await this.getVerifiedStatus(entry)
       return documentParsingModelStatusSchema.parse({

src/main/document-parsing-service.test.ts

@@ -4,14 +4,29 @@ import {
 } from './document-parsing-settings-store'
 import { DocumentParsingService } from './document-parsing-service'
 
-function createPdfFixture(text: string): Buffer {
+function createPdfFixture(...pageTexts: string[]): Buffer {
-  const stream = `BT /F1 18 Tf 50 100 Td (${text}) Tj ET`
+  const texts = pageTexts.length > 0 ? pageTexts : ['']
+  const fontObjectId = texts.length + 3
+  const firstContentObjectId = fontObjectId + 1
   const objects = [
     '<< /Type /Catalog /Pages 2 0 R >>',
-    '<< /Type /Pages /Kids [3 0 R] /Count 1 >>',
+    `<< /Type /Pages /Kids [${texts
-    '<< /Type /Page /Parent 2 0 R /MediaBox [0 0 300 200] /Resources << /Font << /F1 4 0 R >> >> /Contents 5 0 R >>',
+      .map((_, index) => `${index + 3} 0 R`)
+      .join(' ')}] /Count ${texts.length} >>`,
+    ...texts.map(
+      (_, index) =>
+        '<< /Type /Page /Parent 2 0 R /MediaBox [0 0 300 200] ' +
+        `/Resources << /Font << /F1 ${fontObjectId} 0 R >> >> ` +
+        `/Contents ${firstContentObjectId + index} 0 R >>`
+    ),
     '<< /Type /Font /Subtype /Type1 /BaseFont /Helvetica >>',
-    `<< /Length ${Buffer.byteLength(stream)} >>\nstream\n${stream}\nendstream`
+    ...texts.map((text) => {
+      const stream = `BT /F1 18 Tf 50 100 Td (${text}) Tj ET`
+      return (
+        `<< /Length ${Buffer.byteLength(stream)} >>\n` +
+        `stream\n${stream}\nendstream`
+      )
+    })
   ]
   let content = '%PDF-1.4\n'
   const offsets = [0]
@@ -33,41 +48,62 @@ function createPdfFixture(text: string): Buffer {
 
 function createService(overrides?: {
   settings?: Partial<typeof defaultDocumentParsingSettings>
+  modelStatus?: {
+    available: boolean
+    verified: boolean
+    detail: string
+  }
+  recognize?: () => Promise<{
+    requestId: string
+    sections: Array<{
+      locator: string
+      content: string
+      confidence: number
+    }>
+    pageCount: number
+    warnings: string[]
+  }>
 }) {
   const settings = {
     ...defaultDocumentParsingSettings,
     ...overrides?.settings
   }
-  const recognize = vi.fn(async () => ({
+  const recognize = vi.fn(
-    requestId: crypto.randomUUID(),
+    overrides?.recognize ??
-    sections: [
+      (async () => ({
-      {
+        requestId: crypto.randomUUID(),
-        locator: '第 1 页',
+        sections: [
-        content: '扫描件识别正文',
+          {
-        confidence: 0.93
+            locator: '第 1 页',
-      }
+            content: '扫描件识别正文',
-    ],
+            confidence: 0.93
-    pageCount: 1,
+          }
-    warnings: []
+        ],
-  }))
+        pageCount: 1,
-  const service = new DocumentParsingService(
+        warnings: []
-    {
+      }))
-      get: vi.fn(async () => settings),
+  )
-      update: vi.fn(async () => settings)
+  const settingsStore = {
-    } as never,
+    get: vi.fn(async () => settings),
-    {
+    update: vi.fn(async () => settings)
-      getStatus: vi.fn(async () => ({
+  }
+  const modelManager = {
+    getStatus: vi.fn(async () => ({
         id: 'pp-ocrv6-tiny',
         displayName: 'PP-OCRv6 Tiny',
-        available: true,
+        available: overrides?.modelStatus?.available ?? true,
-        verified: true,
+        verified: overrides?.modelStatus?.verified ?? true,
         runtime: 'onnxruntime-web-wasm',
-        detail: '可用'
+        detail: overrides?.modelStatus?.detail ?? '可用'
-      }))
+      })),
-    } as never,
+    getSnapshot: vi.fn()
+  }
+  const service = new DocumentParsingService(
+    settingsStore as never,
+    modelManager as never,
     { recognize } as never
   )
-  return { recognize, service }
+  return { modelManager, recognize, service, settingsStore }
 }
 
 describe('DocumentParsingService', () => {
@@ -127,4 +163,100 @@ describe('DocumentParsingService', () => {
     ).rejects.toThrow('未启用 OCR')
     expect(recognize).not.toHaveBeenCalled()
   })
+
+  it('falls back to useful native text when automatic OCR is unavailable', async () => {
+    const { recognize, service } = createService({
+      modelStatus: {
+        available: false,
+        verified: false,
+        detail: '模型尚未安装'
+      }
+    })
+    const pdf = createPdfFixture('Native PDF body text', '')
+    const originalExtract = await service.parse(
+      'native.pdf',
+      pdf,
+      'chat-attachment'
+    )
+    expect(originalExtract.content).toContain('Native PDF body text')
+    expect(originalExtract.warnings).toEqual([
+      expect.stringContaining('模型尚未安装')
+    ])
+    expect(recognize).not.toHaveBeenCalled()
+  })
+
+  it('falls back to native text when automatic OCR fails', async () => {
+    const { service } = createService({
+      recognize: async () => {
+        throw new Error('OCR runtime failed')
+      }
+    })
+    const parsed = await service.parse(
+      'native.pdf',
+      createPdfFixture('Native PDF body text', ''),
+      'chat-attachment'
+    )
+
+    expect(parsed.content).toContain('Native PDF body text')
+    expect(parsed.warnings).toEqual([
+      expect.stringContaining('OCR runtime failed')
+    ])
+  })
+
+  it('limits the number of pages sent to OCR rather than total PDF pages', async () => {
+    const { recognize, service } = createService({
+      settings: { maximumPages: 1 }
+    })
+
+    await service.parse(
+      'mixed.pdf',
+      createPdfFixture('Native PDF body text', ''),
+      'chat-attachment'
+    )
+
+    expect(recognize).toHaveBeenCalledWith(
+      expect.objectContaining({
+        maximumPages: 1,
+        pageNumbers: [2]
+      })
+    )
+  })
+
+  it('rejects high-fidelity parsing when OCR pages exceed the limit', async () => {
+    const { recognize, service } = createService({
+      settings: {
+        chatWorkflow: 'high-fidelity',
+        maximumPages: 1
+      }
+    })
+
+    await expect(
+      service.parse(
+        'two-pages.pdf',
+        createPdfFixture('First page text', 'Second page text'),
+        'chat-attachment'
+      )
+    ).rejects.toThrow('有 2 页需要 OCR')
+    expect(recognize).not.toHaveBeenCalled()
+  })
+
+  it('rejects selecting an OCR model that is not installed', async () => {
+    const { modelManager, service, settingsStore } = createService()
+    modelManager.getStatus.mockResolvedValueOnce({
+      id: 'pp-ocrv6-medium',
+      displayName: 'PP-OCRv6 Medium',
+      available: false,
+      verified: false,
+      runtime: 'onnxruntime-web-wasm',
+      detail: '模型尚未安装'
+    })
+
+    await expect(
+      service.update({
+        ...defaultDocumentParsingSettings,
+        localOcrModelId: 'pp-ocrv6-medium'
+      })
+    ).rejects.toThrow('请先安装并校验')
+    expect(settingsStore.update).not.toHaveBeenCalled()
+  })
 })

src/main/document-parsing-service.ts

@@ -1,6 +1,7 @@
 import { extname } from 'node:path'
 import {
   documentParsingDiagnosticSchema,
+  documentParsingSettingsUpdateSchema,
   documentParsingSnapshotSchema,
   type DocumentParsingDiagnostic,
   type DocumentParsingPurpose,
@@ -52,9 +53,10 @@ function hasUsefulText(content: string): boolean {
 function effectiveOcrMode(
   settings: DocumentParsingSettings,
   purpose: DocumentParsingPurpose
-): DocumentParsingSettings['pdfOcrMode'] {
+): 'auto' | 'always' | 'disabled' {
   if (
-    (purpose === 'chat-attachment' &&
+    ((purpose === 'chat-attachment' ||
+      purpose === 'artifact-import') &&
       settings.chatWorkflow === 'fast-text') ||
     (purpose === 'knowledge-index' &&
       settings.knowledgeWorkflow === 'fast-index')
@@ -62,14 +64,15 @@ function effectiveOcrMode(
     return 'disabled'
   }
   if (
-    (purpose === 'chat-attachment' &&
+    ((purpose === 'chat-attachment' ||
+      purpose === 'artifact-import') &&
       settings.chatWorkflow === 'high-fidelity') ||
     (purpose === 'knowledge-index' &&
       settings.knowledgeWorkflow === 'high-fidelity')
   ) {
     return 'always'
   }
-  return settings.pdfOcrMode
+  return 'auto'
 }
 
 function buildPdfDocument(
@@ -130,7 +133,21 @@ export class DocumentParsingService {
   }
 
   async update(input: unknown): Promise<DocumentParsingSnapshot> {
-    await this.settingsStore.update(input)
+    const nextSettings =
+      documentParsingSettingsUpdateSchema.parse(input)
+    const currentSettings = await this.settingsStore.get()
+    if (
+      nextSettings.localOcrModelId !==
+      currentSettings.localOcrModelId
+    ) {
+      const status = await this.modelManager.getStatus(
+        nextSettings.localOcrModelId
+      )
+      if (!status.available || !status.verified) {
+        throw new Error('请先安装并校验所选 OCR 模型')
+      }
+    }
+    await this.settingsStore.update(nextSettings)
     return this.snapshot()
   }
 
@@ -159,7 +176,7 @@ export class DocumentParsingService {
           ? pagesWithoutUsefulText
           : []
 
-    if (mode === 'disabled' || !settings.localOcrEnabled) {
+    if (mode === 'disabled') {
       const native = nativePdfSections(pages)
       if (native.length > 0) {
         return buildPdfDocument(
@@ -182,15 +199,24 @@ export class DocumentParsingService {
         pages.length
       )
     }
-    if (pages.length > settings.maximumPages) {
+    if (ocrPageNumbers.length > settings.maximumPages) {
       throw new Error(
-        `PDF 共 ${pages.length} 页，超过本地 OCR 的 ${settings.maximumPages} 页限制`
+        `PDF 有 ${ocrPageNumbers.length} 页需要 OCR，超过 ${settings.maximumPages} 页限制`
       )
     }
+    const native = nativePdfSections(pages)
     const modelStatus = await this.modelManager.getStatus(
       settings.localOcrModelId
     )
     if (!modelStatus.available || !modelStatus.verified) {
+      if (
+        mode === 'auto' &&
+        native.some((section) => hasUsefulText(section.content))
+      ) {
+        return buildPdfDocument(name, native, pages.length, [
+          `本地 OCR 不可用，已保留 PDF 文本层内容：${modelStatus.detail}`
+        ])
+      }
       throw new Error(modelStatus.detail)
     }
 
@@ -203,9 +229,25 @@ export class DocumentParsingService {
       pageNumbers: ocrPageNumbers,
       pageTimeoutSeconds: settings.pageTimeoutSeconds
     }
-    const ocr = await (signal
+    let ocr
-      ? this.ocrBroker.recognize(ocrRequest, signal)
+    try {
-      : this.ocrBroker.recognize(ocrRequest))
+      ocr = await (signal
+        ? this.ocrBroker.recognize(ocrRequest, signal)
+        : this.ocrBroker.recognize(ocrRequest))
+    } catch (error) {
+      ensureNotAborted(signal)
+      if (
+        mode === 'auto' &&
+        native.some((section) => hasUsefulText(section.content))
+      ) {
+        const detail =
+          error instanceof Error ? error.message : '本地 OCR 识别失败'
+        return buildPdfDocument(name, native, pages.length, [
+          `本地 OCR 失败，已保留 PDF 文本层内容：${detail}`
+        ])
+      }
+      throw error
+    }
     ensureNotAborted(signal)
     const ocrByLocator = new Map(
       ocr.sections.map((section) => [section.locator, section])

src/main/document-parsing-settings-store.test.ts

@@ -60,7 +60,7 @@ describe('DocumentParsingSettingsStore', () => {
 
     await expect(store.update(settings)).resolves.toEqual(settings)
     expect(JSON.parse(await readFile(filePath, 'utf8'))).toEqual({
-      version: 2,
+      version: 3,
       ...settings
     })
     await expect(
@@ -68,27 +68,59 @@ describe('DocumentParsingSettingsStore', () => {
     ).resolves.toEqual(settings)
   })
 
-  it('migrates legacy cloud permissions to the local OCR provider', async () => {
+  it('migrates version 2 OCR switches into scenario modes', async () => {
+    const { filePath, store } = await createStore()
+    await writeFile(
+      filePath,
+      JSON.stringify({
+        version: 2,
+        chatWorkflow: 'auto',
+        knowledgeWorkflow: 'complete-index',
+        pdfOcrMode: 'always',
+        ocrProvider: 'local',
+        localOcrEnabled: true,
+        localOcrModelId: 'pp-ocrv6-small',
+        maximumPages: 42,
+        ocrConcurrency: 4,
+        pageTimeoutSeconds: 90
+      }),
+      'utf8'
+    )
+
+    await expect(store.get()).resolves.toEqual({
+      chatWorkflow: 'high-fidelity',
+      knowledgeWorkflow: 'high-fidelity',
+      localOcrModelId: 'pp-ocrv6-small',
+      maximumPages: 42,
+      pageTimeoutSeconds: 90
+    })
+  })
+
+  it('migrates version 1 cloud settings into local scenario modes', async () => {
     const { filePath, store } = await createStore()
-    const {
-      ocrProvider: _ocrProvider,
-      ...legacySettings
-    } = defaultDocumentParsingSettings
-    void _ocrProvider
     await writeFile(
       filePath,
       JSON.stringify({
         version: 1,
-        ...legacySettings,
+        chatWorkflow: 'auto',
+        knowledgeWorkflow: 'complete-index',
+        pdfOcrMode: 'auto',
+        localOcrEnabled: false,
+        localOcrModelId: 'pp-ocrv6-tiny',
+        maximumPages: 100,
+        ocrConcurrency: 1,
+        pageTimeoutSeconds: 60,
         chatCloudPermission: 'always',
         knowledgeCloudPermission: 'never'
       }),
       'utf8'
     )
 
-    await expect(store.get()).resolves.toEqual(
+    await expect(store.get()).resolves.toEqual({
-      defaultDocumentParsingSettings
+      ...defaultDocumentParsingSettings,
-    )
+      chatWorkflow: 'fast-text',
+      knowledgeWorkflow: 'fast-index'
+    })
   })
 
   it('rejects incomplete or out-of-range settings', async () => {

src/main/document-parsing-settings-store.ts

@@ -14,7 +14,7 @@ import {
   type DocumentParsingSettings
 } from '../shared/document-parsing-contracts'
 
-const CURRENT_SETTINGS_VERSION = 2
+const CURRENT_SETTINGS_VERSION = 3
 
 const storedDocumentParsingSettingsSchema =
   documentParsingSettingsSchema
@@ -27,9 +27,32 @@ type StoredDocumentParsingSettings = z.infer<
   typeof storedDocumentParsingSettingsSchema
 >
 
-const legacyDocumentParsingSettingsSchema =
+const legacyVersionTwoSettingsSchema = z
-  documentParsingSettingsSchema
+  .object({
-    .omit({ ocrProvider: true })
+    version: z.literal(2),
+    chatWorkflow: z.enum(['auto', 'fast-text', 'high-fidelity']),
+    knowledgeWorkflow: z.enum([
+      'complete-index',
+      'fast-index',
+      'high-fidelity'
+    ]),
+    pdfOcrMode: z.enum(['auto', 'always', 'disabled']),
+    ocrProvider: z.literal('local'),
+    localOcrEnabled: z.boolean(),
+    localOcrModelId: z
+      .string()
+      .min(1)
+      .max(96)
+      .regex(/^[a-z0-9]+(?:-[a-z0-9]+)*$/u),
+    maximumPages: z.number().int().min(1).max(500),
+    ocrConcurrency: z.number().int().min(1).max(4),
+    pageTimeoutSeconds: z.number().int().min(10).max(300)
+  })
+  .strict()
+
+const legacyVersionOneSettingsSchema =
+  legacyVersionTwoSettingsSchema
+    .omit({ version: true, ocrProvider: true })
     .extend({
       version: z.literal(1),
       chatCloudPermission: z.enum(['ask', 'always', 'never']),
@@ -40,15 +63,39 @@ const legacyDocumentParsingSettingsSchema =
 export const defaultDocumentParsingSettings: DocumentParsingSettings = {
   chatWorkflow: 'auto',
   knowledgeWorkflow: 'complete-index',
-  pdfOcrMode: 'auto',
-  ocrProvider: 'local',
-  localOcrEnabled: true,
   localOcrModelId: 'pp-ocrv6-tiny',
   maximumPages: 100,
-  ocrConcurrency: 1,
   pageTimeoutSeconds: 60
 }
 
+type LegacySettings = z.infer<
+  typeof legacyVersionTwoSettingsSchema
+>
+
+function migrateLegacySettings(
+  legacy: LegacySettings
+): StoredDocumentParsingSettings {
+  const ocrDisabled =
+    !legacy.localOcrEnabled || legacy.pdfOcrMode === 'disabled'
+  const ocrAlways = legacy.pdfOcrMode === 'always'
+  return {
+    version: CURRENT_SETTINGS_VERSION,
+    chatWorkflow: ocrDisabled
+      ? 'fast-text'
+      : legacy.chatWorkflow === 'auto' && ocrAlways
+        ? 'high-fidelity'
+        : legacy.chatWorkflow,
+    knowledgeWorkflow: ocrDisabled
+      ? 'fast-index'
+      : legacy.knowledgeWorkflow === 'complete-index' && ocrAlways
+        ? 'high-fidelity'
+        : legacy.knowledgeWorkflow,
+    localOcrModelId: legacy.localOcrModelId,
+    maximumPages: legacy.maximumPages,
+    pageTimeoutSeconds: legacy.pageTimeoutSeconds
+  }
+}
+
 function isMissingFile(error: unknown): boolean {
   return (
     error !== null &&
@@ -99,23 +146,27 @@ export class DocumentParsingSettingsStore {
       const result =
         storedDocumentParsingSettingsSchema.safeParse(parsed)
       if (!result.success) {
-        const legacy =
+        const versionTwo =
-          legacyDocumentParsingSettingsSchema.safeParse(parsed)
+          legacyVersionTwoSettingsSchema.safeParse(parsed)
-        if (legacy.success) {
+        if (versionTwo.success) {
+          this.settings = migrateLegacySettings(versionTwo.data)
+          return this.settings
+        }
+        const versionOne =
+          legacyVersionOneSettingsSchema.safeParse(parsed)
+        if (versionOne.success) {
           const {
-            version: _version,
             chatCloudPermission: _chatCloudPermission,
             knowledgeCloudPermission: _knowledgeCloudPermission,
-            ...settings
+            ...legacy
-          } = legacy.data
+          } = versionOne.data
-          void _version
           void _chatCloudPermission
           void _knowledgeCloudPermission
-          this.settings = {
+          this.settings = migrateLegacySettings({
-            version: CURRENT_SETTINGS_VERSION,
+            ...legacy,
+            version: 2,
             ocrProvider: 'local',
-            ...settings
+          })
-          }
           return this.settings
         }
         await this.isolateCorruptFile()

src/main/ipc.test.ts

@@ -784,6 +784,149 @@ describe('registerIpcHandlers model ZIP dialogs', () => {
   })
 })
 
+describe('registerIpcHandlers document parsing', () => {
+  const temporaryDirectories: string[] = []
+
+  afterEach(async () => {
+    electronMocks.handlers.clear()
+    vi.clearAllMocks()
+    await Promise.all(
+      temporaryDirectories.splice(0).map((directory) =>
+        rm(directory, { recursive: true, force: true })
+      )
+    )
+  })
+
+  it('tests an explicit scenario and parses artifact PDFs through the shared service', async () => {
+    const directory = await mkdtemp(
+      join(tmpdir(), 'goodbuddy-ipc-document-parsing-')
+    )
+    temporaryDirectories.push(directory)
+    const diagnosticPath = join(directory, 'diagnostic.pdf')
+    const artifactPath = join(directory, 'artifact.pdf')
+    await writeFile(diagnosticPath, 'diagnostic')
+    await writeFile(artifactPath, 'artifact')
+    const diagnostic = {
+      fileName: 'diagnostic.pdf',
+      sourceFormat: 'PDF',
+      pageCount: 1,
+      ocrPageCount: 0,
+      characterCount: 4,
+      method: 'native',
+      durationMs: 1,
+      preview: 'text',
+      warnings: []
+    }
+    const documentParsingService = {
+      diagnose: vi.fn(async () => diagnostic),
+      parse: vi.fn(async () => ({
+        title: 'artifact',
+        sourceFormat: '.pdf',
+        content: 'parsed artifact',
+        sections: [
+          {
+            locator: '第 1 页',
+            content: 'parsed artifact',
+            method: 'native'
+          }
+        ],
+        pageCount: 1,
+        warnings: []
+      }))
+    }
+    const createInlineArtifact = vi.fn((input) => input)
+    const assistantDatabase = {
+      claimDueSchedules: vi.fn(() => []),
+      createInlineArtifact
+    }
+    const webContents = {
+      mainFrame: { url: 'file:///goodbuddy/index.html' },
+      getURL: vi.fn(() => 'file:///goodbuddy/index.html'),
+      send: vi.fn()
+    }
+    const window = {
+      webContents,
+      isDestroyed: vi.fn(() => false),
+      isMaximized: vi.fn(() => false),
+      on: vi.fn(),
+      removeListener: vi.fn()
+    }
+    const event = {
+      sender: webContents,
+      senderFrame: webContents.mainFrame
+    }
+    const dispose = registerIpcHandlers(
+      window as never,
+      { capability: 'text' } as never,
+      'CommandOrControl+Shift+Space',
+      {} as never,
+      {} as never,
+      { clear: vi.fn() } as never,
+      {} as never,
+      assistantDatabase as never,
+      { clear: vi.fn() } as never,
+      {} as never,
+      vi.fn(async () => undefined),
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      undefined,
+      documentParsingService as never
+    )
+
+    electronMocks.showOpenDialog.mockResolvedValueOnce({
+      canceled: false,
+      filePaths: [diagnosticPath]
+    })
+    await expect(
+      electronMocks.handlers.get(
+        ipcChannels.documentParsingTest
+      )?.(event, { purpose: 'knowledge-index' })
+    ).resolves.toBe(diagnostic)
+    expect(documentParsingService.diagnose).toHaveBeenCalledWith(
+      'diagnostic.pdf',
+      expect.any(Buffer),
+      'knowledge-index'
+    )
+
+    await expect(
+      electronMocks.handlers.get(
+        ipcChannels.documentParsingTest
+      )?.(event, { purpose: 'diagnostic' })
+    ).rejects.toThrow()
+
+    electronMocks.showOpenDialog.mockResolvedValueOnce({
+      canceled: false,
+      filePaths: [artifactPath]
+    })
+    await expect(
+      electronMocks.handlers.get(
+        ipcChannels.artifactsImportFiles
+      )?.(event)
+    ).resolves.toEqual([
+      expect.objectContaining({
+        title: 'artifact.pdf',
+        content: '## 第 1 页\n\nparsed artifact'
+      })
+    ])
+    expect(documentParsingService.parse).toHaveBeenCalledWith(
+      'artifact.pdf',
+      expect.any(Buffer),
+      'artifact-import'
+    )
+
+    await dispose()
+  })
+})
+
 describe('registerIpcHandlers connection tests', () => {
   afterEach(() => {
     electronMocks.handlers.clear()

src/main/ipc.ts

@@ -102,7 +102,8 @@ import {
   documentOcrModelActionInputSchema,
   documentOcrFailureSchema,
   documentOcrResultSchema,
-  documentParsingSettingsUpdateSchema
+  documentParsingSettingsUpdateSchema,
+  documentParsingTestInputSchema
 } from '../shared/document-parsing-contracts'
 import {
   agentRuntimeSelectionSchema,
@@ -2785,11 +2786,13 @@ export function registerIpcHandlers(
 
   ipcMain.handle(
     ipcChannels.documentParsingTest,
-    async (event) => {
+    async (event, input: unknown) => {
       assertTrustedSender(event, window)
       if (!documentParsingService) {
         throw new Error('文档解析设置服务不可用')
       }
+      const { purpose } =
+        documentParsingTestInputSchema.parse(input)
       const result = await dialog.showOpenDialog(window, {
         title: '选择测试文档',
         properties: ['openFile'],
@@ -2814,7 +2817,8 @@ export function registerIpcHandlers(
         }
         return documentParsingService.diagnose(
           basename(canonicalPath),
-          await readFile(canonicalPath)
+          await readFile(canonicalPath),
+          purpose
         )
       } catch (error) {
         if (error instanceof Error && !('code' in error)) {
@@ -3436,7 +3440,13 @@ export function registerIpcHandlers(
           )
           continue
         }
-        const parsed = await parseDocument(name, file)
+        const parsed = documentParsingService
+          ? await documentParsingService.parse(
+              name,
+              file,
+              'artifact-import'
+            )
+          : await parseDocument(name, file)
         artifacts.push(
           assistantDatabase.createInlineArtifact({
             projectId,

src/preload/index.ts

@@ -85,7 +85,8 @@ import type {
   DocumentOcrResult,
   DocumentParsingDiagnostic,
   DocumentParsingSettings,
-  DocumentParsingSnapshot
+  DocumentParsingSnapshot,
+  DocumentParsingTestPurpose
 } from '../shared/document-parsing-contracts'
 import type { AgentRuntimeSelection } from '../shared/runtime-selection-contracts'
 import type { WeixinBindingSnapshot } from '../shared/weixin-channel-contracts'
@@ -435,9 +436,10 @@ const desktopApi: DesktopApi = {
         ipcChannels.documentParsingUpdate,
         input
       ) as Promise<DocumentParsingSnapshot>,
-    test: () =>
+    test: (purpose: DocumentParsingTestPurpose) =>
       ipcRenderer.invoke(
-        ipcChannels.documentParsingTest
+        ipcChannels.documentParsingTest,
+        { purpose }
       ) as Promise<DocumentParsingDiagnostic | undefined>,
     installOcrModel: (modelId: string) =>
       ipcRenderer.invoke(

src/preload/preload-sandbox.test.ts

@@ -50,6 +50,19 @@ describe('sandboxed preload', () => {
     expect(source).not.toContain('importOcrModel:')
   })
 
+  it('passes an explicit parsing scenario to document diagnostics', () => {
+    const source = readFileSync(
+      join(process.cwd(), 'src', 'preload', 'index.ts'),
+      'utf8'
+    )
+    expect(source).toContain(
+      'test: (purpose: DocumentParsingTestPurpose)'
+    )
+    expect(source).toContain(
+      'ipcChannels.documentParsingTest,\n        { purpose }'
+    )
+  })
+
   it('exposes a removable attachment parsing progress listener', () => {
     const source = readFileSync(
       join(process.cwd(), 'src', 'preload', 'index.ts'),

src/renderer/src/DocumentParsingSettingsSection.test.tsx

@@ -7,6 +7,7 @@ import {
 } from '@testing-library/react'
 import { afterEach, beforeEach, describe, expect, it, vi } from 'vitest'
 import type {
+  InstalledDocumentOcrModel,
   DocumentParsingSettings,
   DocumentParsingSnapshot
 } from '../../shared/document-parsing-contracts'
@@ -16,12 +17,8 @@ import { DocumentParsingSettingsSection } from './DocumentParsingSettingsSection
 const settings: DocumentParsingSettings = {
   chatWorkflow: 'auto',
   knowledgeWorkflow: 'complete-index',
-  pdfOcrMode: 'auto',
-  ocrProvider: 'local',
-  localOcrEnabled: true,
   localOcrModelId: 'pp-ocrv6-tiny',
   maximumPages: 100,
-  ocrConcurrency: 1,
   pageTimeoutSeconds: 60
 }
 
@@ -139,35 +136,36 @@ const test = vi.fn(async () => ({
   preview: '扫描件识别正文',
   warnings: []
 }))
-const installOcrModel = vi.fn(async () => ({
+const installOcrModel =
-  ...snapshot,
+  vi.fn<() => Promise<DocumentParsingSnapshot>>(async () => ({
-  status: {
+    ...snapshot,
-    ...snapshot.status,
+    status: {
-    localOcr: {
+      ...snapshot.status,
-      ...snapshot.status.localOcr,
+      localOcr: {
-      available: true,
+        ...snapshot.status.localOcr,
-      verified: true,
+        available: true,
-      detail: '模型已安装并校验'
+        verified: true,
-    }
+        detail: '模型已安装并校验'
-  },
-  ocrModels: {
-    ...snapshot.ocrModels,
-    installed: [
-      {
-        id: 'pp-ocrv6-tiny' as const,
-        displayName: 'PP-OCRv6 Tiny',
-        source: 'download' as const,
-        installedAt: '2026-08-11T00:00:00.000Z',
-        files: modelEntry.files.map((file) => ({
-          name: file.name,
-          role: file.role,
-          size: file.download.size,
-          sha256: file.download.sha256
-        }))
       }
-    ]
+    },
-  }
+    ocrModels: {
-}))
+      ...snapshot.ocrModels,
+      installed: [
+        {
+          id: 'pp-ocrv6-tiny',
+          displayName: 'PP-OCRv6 Tiny',
+          source: 'download',
+          installedAt: '2026-08-11T00:00:00.000Z',
+          files: modelEntry.files.map((file) => ({
+            name: file.name,
+            role: file.role,
+            size: file.download.size,
+            sha256: file.download.sha256
+          }))
+        }
+      ]
+    }
+  }))
 const importOcrModelArchive = vi.fn(async () => snapshot)
 const exportOcrModelArchive = vi.fn(async () => snapshot)
 const openOcrModelRepository = vi.fn(async () => undefined)
@@ -212,17 +210,7 @@ describe('DocumentParsingSettingsSection', () => {
     expect(screen.getByText('质量：基础')).toBeInTheDocument()
     expect(screen.getByText('速度：快')).toBeInTheDocument()
     expect(screen.getByText('旧版 Office 转换')).toBeInTheDocument()
-    expect(
+    expect(screen.queryByRole('switch')).not.toBeInTheDocument()
-      screen.getByRole('switch', { name: /启用本地 OCR/u })
-    ).toBeChecked()
-    expect(
-      screen.getByRole('button', { name: '本地模型' })
-    ).toHaveAttribute('aria-pressed', 'true')
-    expect(
-      screen.getByRole('button', {
-        name: '远程服务（即将支持）'
-      })
-    ).toBeDisabled()
     expect(screen.queryByText('隐私与云端处理')).not.toBeInTheDocument()
     expect(
       screen.queryByText('模型详情与手动导入')
@@ -237,12 +225,15 @@ describe('DocumentParsingSettingsSection', () => {
     )
     expect(openOcrModelRepository).toHaveBeenCalledWith('pp-ocrv6-tiny')
 
-    fireEvent.change(screen.getByLabelText('聊天附件'), {
+    fireEvent.change(screen.getByLabelText('聊天与成果文件'), {
       target: { value: 'fast-text' }
     })
-    fireEvent.click(
+    expect(
-      screen.getByRole('button', { name: '保存设置' })
+      screen.getByRole('button', {
-    )
+        name: '测试聊天与成果模式'
+      })
+    ).toBeDisabled()
+    fireEvent.click(screen.getByRole('button', { name: '保存设置' }))
 
     await waitFor(() =>
       expect(update).toHaveBeenCalledWith(
@@ -278,6 +269,59 @@ describe('DocumentParsingSettingsSection', () => {
     )
   })
 
+  it('downloads and selects an uninstalled model in one action', async () => {
+    const installedMedium = {
+      ...snapshot,
+      settings: {
+        ...snapshot.settings,
+        localOcrModelId: 'pp-ocrv6-medium'
+      }
+    }
+    installOcrModel.mockResolvedValueOnce({
+      ...snapshot,
+      ocrModels: {
+        ...snapshot.ocrModels,
+        installed: [
+          ...snapshot.ocrModels.installed,
+          {
+            id: 'pp-ocrv6-medium',
+            displayName: 'PP-OCRv6 Medium',
+            source: 'download',
+            installedAt: '2026-08-11T00:00:00.000Z',
+            files: thirdModelEntry.files.map((file) => ({
+              name: file.name,
+              role: file.role,
+              size: file.download.size,
+              sha256: file.download.sha256
+            }))
+          } satisfies InstalledDocumentOcrModel
+        ]
+      }
+    })
+    update.mockResolvedValueOnce(installedMedium)
+    render(<DocumentParsingSettingsSection />)
+
+    fireEvent.change(await screen.findByLabelText('当前 OCR 模型'), {
+      target: { value: 'pp-ocrv6-medium' }
+    })
+    expect(
+      screen.getByRole('button', { name: '保存设置' })
+    ).toBeDisabled()
+    fireEvent.click(
+      screen.getByRole('button', {
+        name: '下载 PP-OCRv6 Medium'
+      })
+    )
+
+    await waitFor(() =>
+      expect(update).toHaveBeenCalledWith(
+        expect.objectContaining({
+          localOcrModelId: 'pp-ocrv6-medium'
+        })
+      )
+    )
+  })
+
   it('imports and exports verified OCR model ZIP archives', async () => {
     const onNotify = vi.fn()
     render(
@@ -370,7 +414,9 @@ describe('DocumentParsingSettingsSection', () => {
     await screen.findByText('PP-OCRv6 Tiny')
 
     fireEvent.click(
-      screen.getByRole('button', { name: '测试解析' })
+      screen.getByRole('button', {
+        name: '测试聊天与成果模式'
+      })
     )
 
     expect(
@@ -378,6 +424,7 @@ describe('DocumentParsingSettingsSection', () => {
         name: '解析测试结果'
       })
     ).toHaveTextContent('扫描件识别正文')
-    expect(test).toHaveBeenCalledOnce()
+    expect(test).toHaveBeenCalledWith('chat-attachment')
+    expect(update).not.toHaveBeenCalled()
   })
 })

src/renderer/src/DocumentParsingSettingsSection.tsx

@@ -25,11 +25,11 @@ import type {
   DocumentOcrModelCatalogEntry,
   DocumentOcrModelOperation,
   DocumentParsingSettings,
-  DocumentParsingSnapshot
+  DocumentParsingSnapshot,
+  DocumentParsingTestPurpose
 } from '../../shared/document-parsing-contracts'
 import type { AppNotificationInput } from './notifications'
 import { SettingsCategoryHeader } from './SettingsPrimitives'
-import { SegmentedControl } from './WorkspacePrimitives'
 
 type DocumentParsingSettingsSectionProps = {
   onNotify?: (notification: AppNotificationInput) => void
@@ -211,7 +211,8 @@ export function DocumentParsingSettingsSection({
   const [draft, setDraft] = useState<DocumentParsingSettings>()
   const [error, setError] = useState<string>()
   const [saving, setSaving] = useState(false)
-  const [testing, setTesting] = useState(false)
+  const [testingPurpose, setTestingPurpose] =
+    useState<DocumentParsingTestPurpose>()
   const [busyModelId, setBusyModelId] = useState<string>()
   const [confirmingRemove, setConfirmingRemove] = useState<string>()
   const [diagnostic, setDiagnostic] =
@@ -283,9 +284,7 @@ export function DocumentParsingSettingsSection({
     )
   }
 
-  const save = async (
+  const save = async (): Promise<DocumentParsingSnapshot | undefined> => {
-    notify = true
-  ): Promise<DocumentParsingSnapshot | undefined> => {
     const api = window.goodbuddy.documentParsing
     if (!api || !draft) {
       return undefined
@@ -296,13 +295,11 @@ export function DocumentParsingSettingsSection({
       const next = await api.update(draft)
       setSnapshot(next)
       setDraft(next.settings)
-      if (notify) {
+      onNotify?.({
-        onNotify?.({
+        tone: 'success',
-          tone: 'success',
+        message: t('notifications.documentParsingSaved'),
-          message: t('notifications.documentParsingSaved'),
+        dedupeKey: 'document-parsing-saved'
-          dedupeKey: 'document-parsing-saved'
+      })
-        })
-      }
       return next
     } catch (reason) {
       setError(
@@ -365,19 +362,18 @@ export function DocumentParsingSettingsSection({
     )
   }
 
-  const testParsing = async (): Promise<void> => {
+  const testParsing = async (
+    purpose: DocumentParsingTestPurpose
+  ): Promise<void> => {
     const api = window.goodbuddy.documentParsing
     if (!api) {
       setError(t('errors.documentParsingUnavailable'))
       return
     }
-    setTesting(true)
+    setTestingPurpose(purpose)
     setError(undefined)
     try {
-      if (!(await save(false))) {
+      const result = await api.test(purpose)
-        return
-      }
-      const result = await api.test()
       if (result) {
         setDiagnostic(result)
         onNotify?.({
@@ -391,7 +387,7 @@ export function DocumentParsingSettingsSection({
         errorMessage(reason, t('errors.testDocumentParsing'))
       )
     } finally {
-      setTesting(false)
+      setTestingPurpose(undefined)
     }
   }
 
@@ -415,6 +411,10 @@ export function DocumentParsingSettingsSection({
   const model = snapshot.ocrModels.catalog.find(
     (entry) => entry.id === draft.localOcrModelId
   )
+  const selectedModelInCatalog = model !== undefined
+  const currentModelInCatalog = snapshot.ocrModels.catalog.some(
+    (entry) => entry.id === snapshot.settings.localOcrModelId
+  )
   const installedModel = snapshot.ocrModels.installed.find(
     (entry) => entry.id === draft.localOcrModelId
   )
@@ -426,38 +426,41 @@ export function DocumentParsingSettingsSection({
     : undefined
   const pendingModelSelection =
     draft.localOcrModelId !== snapshot.settings.localOcrModelId
+  const settingsDirty =
+    JSON.stringify(draft) !== JSON.stringify(snapshot.settings)
+  const selectedModelReady = installedModel !== undefined
+  const invalidPendingModel =
+    pendingModelSelection && !selectedModelReady
+  const testing = testingPurpose !== undefined
 
   return (
     <>
       <SettingsCategoryHeader
         actions={
-          <>
+          <button
-            <button
+            className="primary-button"
-              className="secondary-button"
+            disabled={
-              disabled={saving || testing}
+              saving || testing || !settingsDirty || invalidPendingModel
-              onClick={() => void testParsing()}
+            }
-              type="button"
+            onClick={() => void save()}
-            >
+            type="button"
-              <FileSearch aria-hidden="true" size={14} />
+          >
-              {testing
+            {saving
-                ? t('actions.testingParsing')
+              ? t('actions.saving')
-                : t('actions.testParsing')}
+              : t('actions.saveSettings')}
-            </button>
+          </button>
-            <button
-              className="primary-button"
-              disabled={saving || testing}
-              onClick={() => void save()}
-              type="button"
-            >
-              {saving
-                ? t('actions.saving')
-                : t('actions.saveSettings')}
-            </button>
-          </>
         }
         category="document-parsing"
         error={error}
       />
+      {settingsDirty && (
+        <p
+          className="settings-notice"
+          id="document-parsing-unsaved-notice"
+        >
+          {t('documentParsing.workflows.unsavedNotice')}
+        </p>
+      )}
 
       <section
         aria-labelledby="document-parsing-status-title"
@@ -483,9 +486,13 @@ export function DocumentParsingSettingsSection({
             detail={t(
               snapshot.status.localOcr.available
                 ? 'documentParsing.status.ocrReady'
-                : 'documentParsing.status.ocrUnavailable'
+                : currentModelInCatalog
+                  ? 'documentParsing.status.ocrUnavailable'
+                  : 'documentParsing.ocr.selectedModelUnavailable'
             )}
-            label={t('documentParsing.status.localOcr')}
+            label={t('documentParsing.status.localOcrModel', {
+              name: snapshot.status.localOcr.displayName
+            })}
           />
           <StatusRow
             available={snapshot.status.conversionAvailable}
@@ -514,10 +521,13 @@ export function DocumentParsingSettingsSection({
           </div>
         </div>
         <div className="document-parsing-grid">
-          <label className="field">
+          <div className="field">
-            <span>{t('documentParsing.workflows.chat')}</span>
+            <label htmlFor="document-parsing-chat-workflow">
+              {t('documentParsing.workflows.chat')}
+            </label>
             <select
               aria-label={t('documentParsing.workflows.chat')}
+              id="document-parsing-chat-workflow"
               onChange={(event) =>
                 updateDraft(
                   'chatWorkflow',
@@ -540,13 +550,36 @@ export function DocumentParsingSettingsSection({
               </option>
             </select>
             <small>
-              {t('documentParsing.workflows.chatDescription')}
+              {t(
+                `documentParsing.workflows.chatDescriptions.${draft.chatWorkflow}`
+              )}
             </small>
-          </label>
+            <button
-          <label className="field">
+              aria-describedby={
-            <span>{t('documentParsing.workflows.knowledge')}</span>
+                settingsDirty
+                  ? 'document-parsing-unsaved-notice'
+                  : undefined
+              }
+              className="secondary-button document-parsing-workflow-test"
+              disabled={saving || testing || settingsDirty}
+              onClick={() =>
+                void testParsing('chat-attachment')
+              }
+              type="button"
+            >
+              <FileSearch aria-hidden="true" size={14} />
+              {testingPurpose === 'chat-attachment'
+                ? t('actions.testingParsing')
+                : t('documentParsing.workflows.testChat')}
+            </button>
+          </div>
+          <div className="field">
+            <label htmlFor="document-parsing-knowledge-workflow">
+              {t('documentParsing.workflows.knowledge')}
+            </label>
             <select
               aria-label={t('documentParsing.workflows.knowledge')}
+              id="document-parsing-knowledge-workflow"
               onChange={(event) =>
                 updateDraft(
                   'knowledgeWorkflow',
@@ -573,9 +606,29 @@ export function DocumentParsingSettingsSection({
               </option>
             </select>
             <small>
-              {t('documentParsing.workflows.knowledgeDescription')}
+              {t(
+                `documentParsing.workflows.knowledgeDescriptions.${draft.knowledgeWorkflow}`
+              )}
             </small>
-          </label>
+            <button
+              aria-describedby={
+                settingsDirty
+                  ? 'document-parsing-unsaved-notice'
+                  : undefined
+              }
+              className="secondary-button document-parsing-workflow-test"
+              disabled={saving || testing || settingsDirty}
+              onClick={() =>
+                void testParsing('knowledge-index')
+              }
+              type="button"
+            >
+              <FileSearch aria-hidden="true" size={14} />
+              {testingPurpose === 'knowledge-index'
+                ? t('actions.testingParsing')
+                : t('documentParsing.workflows.testKnowledge')}
+            </button>
+          </div>
         </div>
       </section>
 
@@ -604,40 +657,6 @@ export function DocumentParsingSettingsSection({
           </button>
         </div>
 
-        <div className="document-ocr-provider">
-          <div>
-            <strong>{t('documentParsing.ocr.provider.title')}</strong>
-            <small>
-              {t('documentParsing.ocr.provider.description')}
-            </small>
-          </div>
-          <SegmentedControl
-            ariaLabel={t('documentParsing.ocr.provider.title')}
-            onChange={(value) => {
-              if (value === 'local') {
-                updateDraft('ocrProvider', value)
-              }
-            }}
-            options={[
-              {
-                value: 'local',
-                label: t('documentParsing.ocr.provider.local')
-              },
-              {
-                value: 'remote',
-                label: t('documentParsing.ocr.provider.remote'),
-                disabled: true
-              }
-            ]}
-            value={draft.ocrProvider}
-          />
-          <small>
-            {t('documentParsing.ocr.provider.remoteDescription')}
-          </small>
-        </div>
-
-        {draft.ocrProvider === 'local' && (
-          <>
         <label className="field document-ocr-model-selector">
           <span>{t('documentParsing.ocr.modelSelector')}</span>
           <select
@@ -647,6 +666,12 @@ export function DocumentParsingSettingsSection({
             }
             value={draft.localOcrModelId}
           >
+            {!selectedModelInCatalog && (
+              <option value={draft.localOcrModelId}>
+                {draft.localOcrModelId} ·{' '}
+                {t('documentParsing.ocr.unavailableOption')}
+              </option>
+            )}
             {snapshot.ocrModels.catalog.map((entry) => {
               const installed = snapshot.ocrModels.installed.some(
                 (candidate) => candidate.id === entry.id
@@ -662,7 +687,9 @@ export function DocumentParsingSettingsSection({
             })}
           </select>
           <small>
-            {pendingModelSelection
+            {invalidPendingModel
+              ? t('documentParsing.ocr.installBeforeSelecting')
+              : pendingModelSelection
               ? t('documentParsing.ocr.pendingSelection')
               : t('documentParsing.ocr.modelSelectorDescription')}
           </small>
@@ -835,11 +862,20 @@ export function DocumentParsingSettingsSection({
                     onClick={() =>
                       void runModelOperation(
                         model.id,
-                        () =>
+                        async () => {
-                          window.goodbuddy.documentParsing!
+                          const installed =
-                            .installOcrModel(model.id),
+                            await window.goodbuddy.documentParsing!
+                              .installOcrModel(model.id)
+                          if (!pendingModelSelection) {
+                            return installed
+                          }
+                          return window.goodbuddy.documentParsing!
+                            .update(draft)
+                        },
                         t(
-                          'documentParsing.ocr.notifications.installed',
+                          pendingModelSelection
+                            ? 'documentParsing.ocr.notifications.installedAndSelected'
+                            : 'documentParsing.ocr.notifications.installed',
                           { name: model.displayName }
                         )
                       )
@@ -847,7 +883,9 @@ export function DocumentParsingSettingsSection({
                     type="button"
                   >
                     <Download aria-hidden="true" size={13} />
-                    {t('documentParsing.ocr.download')}
+                    {pendingModelSelection
+                      ? t('documentParsing.ocr.downloadAndSelect')
+                      : t('documentParsing.ocr.download')}
                   </button>
                   <button
                     aria-label={t(
@@ -859,11 +897,20 @@ export function DocumentParsingSettingsSection({
                     onClick={() =>
                       void runModelOperation(
                         model.id,
-                        () =>
+                        async () => {
-                          window.goodbuddy.documentParsing!
+                          const imported =
-                            .importOcrModelArchive(model.id),
+                            await window.goodbuddy.documentParsing!
+                              .importOcrModelArchive(model.id)
+                          if (!imported || !pendingModelSelection) {
+                            return imported
+                          }
+                          return window.goodbuddy.documentParsing!
+                            .update(draft)
+                        },
                         t(
-                          'documentParsing.ocr.notifications.importedZip',
+                          pendingModelSelection
+                            ? 'documentParsing.ocr.notifications.importedAndSelected'
+                            : 'documentParsing.ocr.notifications.importedZip',
                           { name: model.displayName }
                         )
                       )
@@ -904,55 +951,17 @@ export function DocumentParsingSettingsSection({
           </article>
         ) : (
           <p className="settings-warning">
-            {t('documentParsing.ocr.catalogUnavailable')}
+            {t(
+              snapshot.ocrModels.catalog.length === 0
+                ? 'documentParsing.ocr.catalogUnavailable'
+                : 'documentParsing.ocr.selectedModelUnavailable'
+            )}
           </p>
         )}
 
-        <div className="document-ocr-settings__options">
+        <p className="settings-notice">
-          <label className="toggle-row">
+          {t('documentParsing.ocr.privacyNotice')}
-            <input
+        </p>
-              checked={draft.localOcrEnabled}
-              onChange={(event) =>
-                updateDraft('localOcrEnabled', event.target.checked)
-              }
-              role="switch"
-              type="checkbox"
-            />
-            <span className="field">
-              <strong>{t('documentParsing.ocr.enabled')}</strong>
-              <small>
-                {t('documentParsing.ocr.enabledDescription')}
-              </small>
-            </span>
-          </label>
-          <label className="field">
-            <span>{t('documentParsing.ocr.mode')}</span>
-            <select
-              aria-label={t('documentParsing.ocr.mode')}
-              disabled={!draft.localOcrEnabled}
-              onChange={(event) =>
-                updateDraft(
-                  'pdfOcrMode',
-                  event.target
-                    .value as DocumentParsingSettings['pdfOcrMode']
-                )
-              }
-              value={draft.pdfOcrMode}
-            >
-              <option value="auto">
-                {t('documentParsing.ocr.modes.auto')}
-              </option>
-              <option value="always">
-                {t('documentParsing.ocr.modes.always')}
-              </option>
-              <option value="disabled">
-                {t('documentParsing.ocr.modes.disabled')}
-              </option>
-            </select>
-          </label>
-        </div>
-          </>
-        )}
       </section>
 
       <details className="settings-section">
@@ -973,21 +982,6 @@ export function DocumentParsingSettingsSection({
               value={draft.maximumPages}
             />
           </label>
-          <label className="field">
-            <span>{t('documentParsing.advanced.concurrency')}</span>
-            <input
-              max={4}
-              min={1}
-              onChange={(event) =>
-                updateDraft(
-                  'ocrConcurrency',
-                  Number(event.target.value)
-                )
-              }
-              type="number"
-              value={draft.ocrConcurrency}
-            />
-          </label>
           <label className="field">
             <span>{t('documentParsing.advanced.timeout')}</span>
             <input
@@ -1004,7 +998,7 @@ export function DocumentParsingSettingsSection({
             />
           </label>
         </div>
-        <small>{t('documentParsing.advanced.concurrencyHint')}</small>
+        <small>{t('documentParsing.advanced.description')}</small>
       </details>
 
       {diagnostic && (

src/renderer/src/document-ocr-bridge.ts

@@ -6,6 +6,7 @@ import type {
 
 type WorkerOutput =
   | { type: 'ready' }
+  | { type: 'progress'; requestId: string; pageNumber?: number }
   | { type: 'result'; result: DocumentOcrResult }
   | { type: 'error'; requestId?: string; error: string }
 
@@ -13,6 +14,7 @@ type PendingWorkerRequest = {
   resolve: (result: DocumentOcrResult) => void
   reject: (error: Error) => void
   timer: number
+  timeoutMs: number
 }
 
 let worker: Worker | undefined
@@ -46,6 +48,18 @@ function terminateWorker(error: Error): void {
   pending.clear()
 }
 
+function armPageTimeout(
+  requestId: string,
+  request: PendingWorkerRequest
+): void {
+  window.clearTimeout(request.timer)
+  request.timer = window.setTimeout(() => {
+    pending.delete(requestId)
+    terminateWorker(new Error('单页 OCR 解析超时'))
+    request.reject(new Error('单页 OCR 解析超时'))
+  }, request.timeoutMs)
+}
+
 async function ensureWorker(modelId: string): Promise<Worker> {
   if (worker && workerReady && workerModelId === modelId) {
     await workerReady
@@ -74,6 +88,13 @@ async function ensureWorker(modelId: string): Promise<Worker> {
         resolveWorkerReady?.()
         return
       }
+      if (output.type === 'progress') {
+        const request = pending.get(output.requestId)
+        if (request) {
+          armPageTimeout(output.requestId, request)
+        }
+        return
+      }
       if (output.type === 'error' && !output.requestId) {
         rejectWorkerReady?.(new Error(output.error))
         return
@@ -134,21 +155,16 @@ async function recognize(
   if (cancelledRequestIds.has(request.requestId)) {
     throw new Error('本地 OCR 解析已取消')
   }
-  const pageCount = request.pageNumbers?.length ?? request.maximumPages
+  const timeoutMs = request.pageTimeoutSeconds * 1_000
-  const timeoutMs = Math.min(
-    10 * 60 * 1_000,
-    Math.max(
-      request.pageTimeoutSeconds * 1_000,
-      request.pageTimeoutSeconds * pageCount * 1_000
-    )
-  )
   return new Promise<DocumentOcrResult>((resolve, reject) => {
-    const timer = window.setTimeout(() => {
+    const workerRequest = {
-      pending.delete(request.requestId)
+      resolve,
-      terminateWorker(new Error('本地 OCR 解析超时'))
+      reject,
-      reject(new Error('本地 OCR 解析超时'))
+      timer: 0,
-    }, timeoutMs)
+      timeoutMs
-    pending.set(request.requestId, { resolve, reject, timer })
+    }
+    pending.set(request.requestId, workerRequest)
+    armPageTimeout(request.requestId, workerRequest)
     activeWorker.postMessage(
       { type: 'recognize', request },
       [request.data]

src/renderer/src/document-ocr-worker.ts

@@ -26,6 +26,7 @@ type WorkerInput = InitializeMessage | RecognizeMessage
 
 type WorkerOutput =
   | { type: 'ready' }
+  | { type: 'progress'; requestId: string; pageNumber?: number }
   | { type: 'result'; result: DocumentOcrResult }
   | { type: 'error'; requestId?: string; error: string }
 
@@ -141,6 +142,10 @@ async function renderPdfPage(
 async function recognizePdf(
   request: DocumentOcrRequest
 ): Promise<DocumentOcrResult> {
+  worker.postMessage({
+    type: 'progress',
+    requestId: request.requestId
+  } satisfies WorkerOutput)
   const pdfjs = await import('pdfjs-dist')
   pdfjs.GlobalWorkerOptions.workerSrc = pdfWorkerUrl
   const loadingTask = pdfjs.getDocument(
@@ -150,14 +155,21 @@ async function recognizePdf(
   const selectedPages = new Set(
     request.pageNumbers ??
       Array.from(
-        { length: Math.min(document.numPages, request.maximumPages) },
+        { length: document.numPages },
         (_, index) => index + 1
       )
   )
-  if (document.numPages > request.maximumPages) {
+  const invalidPage = [...selectedPages].find(
+    (pageNumber) => pageNumber > document.numPages
+  )
+  if (invalidPage !== undefined) {
+    await loadingTask.destroy()
+    throw new Error(`PDF 不包含第 ${invalidPage} 页`)
+  }
+  if (selectedPages.size > request.maximumPages) {
     await loadingTask.destroy()
     throw new Error(
-      `PDF 共 ${document.numPages} 页，超过 ${request.maximumPages} 页限制`
+      `PDF 有 ${selectedPages.size} 页需要 OCR，超过 ${request.maximumPages} 页限制`
     )
   }
   const sections: DocumentOcrResult['sections'] = []
@@ -171,6 +183,11 @@ async function recognizePdf(
       if (!selectedPages.has(pageNumber)) {
         continue
       }
+      worker.postMessage({
+        type: 'progress',
+        requestId: request.requestId,
+        pageNumber
+      } satisfies WorkerOutput)
       const page = await document.getPage(pageNumber)
       try {
         const section = await recognizeImage(
@@ -201,6 +218,11 @@ async function recognize(request: DocumentOcrRequest): Promise<DocumentOcrResult
   if (request.mimeType === 'application/pdf') {
     return recognizePdf(request)
   }
+  worker.postMessage({
+    type: 'progress',
+    requestId: request.requestId,
+    pageNumber: 1
+  } satisfies WorkerOutput)
   const section = await recognizeImage(request.data, '图片')
   return {
     requestId: request.requestId,

src/renderer/src/i18n/locales/en-US/settings.ts

@@ -83,7 +83,6 @@ export const settings = {
     saveAndTestRuntime: 'Save and test {{runtime}}',
     saving: 'Saving…',
     saveSettings: 'Save settings',
-    testParsing: 'Test parsing',
     testingParsing: 'Parsing…',
     select: 'Select',
     selectFile: 'Select file',
@@ -238,52 +237,55 @@ export const settings = {
       conversionUnavailable:
         'Not implemented yet; DOC, XLS, and PPT are currently unavailable',
       localOcr: 'Local OCR',
+      localOcrModel: 'Current OCR: {{name}}',
       ocrReady:
         'The model is installed, SHA-256 verified, and available offline',
       ocrUnavailable:
         'The model is not installed or failed verification. Download it from ModelScope.',
       partialNotice:
-        'Basic document parsing is available. Legacy Office conversion is not implemented yet; scanned PDFs use local OCR.'
+        'Basic document parsing is available. Legacy Office conversion is not implemented yet; scenario modes can use local OCR for scanned PDFs.'
     },
     workflows: {
-      title: 'Usage scenarios',
+      title: 'PDF parsing modes',
       description:
-        'Choose different parsing depth for chat attachments and knowledge imports',
+        'Choose how each scenario handles PDF text layers and scanned pages',
-      chat: 'Chat attachments',
+      chat: 'Chat and artifact files',
-      chatDescription:
-        'Controls parsing before an attachment is added to the current request',
       knowledge: 'Knowledge imports',
-      knowledgeDescription:
+      testChat: 'Test chat and artifact mode',
-        'Controls parsing before chunking, indexing, and source location',
+      testKnowledge: 'Test knowledge mode',
+      unsavedNotice:
+        'There are unsaved changes. Save them before testing the active mode.',
       chatOptions: {
-        auto: 'Automatic parsing (recommended)',
+        auto: 'Automatic recognition (recommended)',
-        fastText: 'Fast text',
+        fastText: 'Text layer only',
-        highFidelity: 'High-fidelity parsing'
+        highFidelity: 'OCR every page'
+      },
+      chatDescriptions: {
+        auto:
+          'Chat attachments and artifact PDFs prefer the text layer and use OCR only on pages without useful text.',
+        fastText:
+          'Chat attachments and artifact PDFs use only the text layer. Scanned documents may be unreadable.',
+        highFidelity:
+          'Run OCR on every PDF page. This is slower.'
       },
       knowledgeOptions: {
-        completeIndex: 'Complete indexing (recommended)',
+        completeIndex: 'Automatic recognition (recommended)',
-        fastIndex: 'Fast indexing',
+        fastIndex: 'Text layer only',
-        highFidelity: 'High-fidelity indexing'
+        highFidelity: 'OCR every page'
+      },
+      knowledgeDescriptions: {
+        'complete-index':
+          'Prefer the PDF text layer and use OCR only on pages without useful text.',
+        'fast-index':
+          'Use only the PDF text layer. Scanned pages are not indexed.',
+        'high-fidelity':
+          'Run OCR on every PDF page before chunking and indexing. This is slower.'
       }
     },
     ocr: {
       title: 'OCR recognition',
       description:
         'Install a local model on demand to recognize scanned PDFs on this device',
-      enabled: 'Enable local OCR',
-      enabledDescription:
-        'After installation, the model runs only on this device through ONNX Runtime WebAssembly. Documents are not uploaded for recognition.',
-      model: 'Local model',
-      runtime: 'Runtime',
-      provider: {
-        title: 'OCR source',
-        description:
-          'Choose either a local model or a remote service, then save settings to switch.',
-        local: 'Local model',
-        remote: 'Remote service (coming soon)',
-        remoteDescription:
-          'Remote integrations will support services such as MinerU and PaddleOCR-VL. They are disabled in this version.'
-      },
       modelSelector: 'Current OCR model',
       modelSelectorDescription:
         'This saved model is used for chat attachments and knowledge imports.',
@@ -291,6 +293,7 @@ export const settings = {
         'This model selection is not active yet. Save settings to switch.',
       installedOption: 'Installed',
       downloadableOption: 'Available to download',
+      unavailableOption: 'Unavailable in this version',
       openModelsDirectory: 'Open model folder',
       storagePrefix: 'Models are installed on demand in',
       storageSuffix:
@@ -314,6 +317,7 @@ export const settings = {
       },
       installed: 'Installed and verified',
       download: 'Download',
+      downloadAndSelect: 'Download and enable',
       importZip: 'Import ZIP',
       exportZip: 'Export ZIP',
       delete: 'Delete',
@@ -322,14 +326,12 @@ export const settings = {
       openRepository: 'Open ModelScope',
       catalogUnavailable:
         'No OCR model catalog is available in this version.',
-      mode: 'PDF OCR strategy',
+      selectedModelUnavailable:
-      modes: {
+        'The saved OCR model is unavailable in this version. Select and install another model above.',
-        auto: 'Automatic; recognize only pages without useful text',
+      installBeforeSelecting:
-        always: 'Always recognize every page',
+        'Download this model first. It will become the current model after installation.',
-        disabled: 'Use only the PDF text layer'
+      privacyNotice:
-      },
+        'OCR is enabled only when required by the scenario modes above. It always runs locally through ONNX Runtime WebAssembly and never uploads documents.',
-      modelLicense:
-        'The model uses Apache License 2.0 and is SHA-256 verified before loading.',
       operations: {
         preparing: 'Preparing model files',
         downloading: 'Downloading from ModelScope',
@@ -347,7 +349,11 @@ export const settings = {
       },
       notifications: {
         installed: '{{name}} installed',
+        installedAndSelected:
+          '{{name}} installed and selected as the current model',
         importedZip: '{{name}} imported from ZIP',
+        importedAndSelected:
+          '{{name}} imported and selected as the current model',
         exportedZip: '{{name}} exported as ZIP',
         removed: 'OCR model deleted'
       }
@@ -355,10 +361,9 @@ export const settings = {
     advanced: {
       title: 'Advanced parsing settings',
       maximumPages: 'Maximum OCR pages per document',
-      concurrency: 'OCR concurrency',
       timeout: 'OCR time budget per page (seconds)',
-      concurrencyHint:
+      description:
-        'The WASM baseline currently processes pages serially; this value is reserved for batching and hardware acceleration.'
+        'The page limit counts only pages actually sent to OCR. Parsing stops if one page exceeds its time budget.'
     },
     diagnostic: {
       title: 'Parsing test result',

src/renderer/src/i18n/locales/zh-CN/settings.ts

@@ -74,7 +74,6 @@ export const settings = {
     saveAndTestRuntime: '保存并测试 {{runtime}}',
     saving: '保存中…',
     saveSettings: '保存设置',
-    testParsing: '测试解析',
     testingParsing: '正在解析…',
     select: '选择',
     selectFile: '选择文件',
@@ -217,50 +216,55 @@ export const settings = {
       conversion: '旧版 Office 转换',
       conversionUnavailable: '尚未实现，DOC、XLS、PPT 暂不可用',
       localOcr: '本地 OCR',
+      localOcrModel: '当前 OCR：{{name}}',
       ocrReady: '模型已安装并通过 SHA-256 校验，可离线使用',
       ocrUnavailable: '模型尚未安装或校验失败，请从 ModelScope 下载',
       partialNotice:
-        '基础文档解析可用。旧版 Office 转换尚未实现；扫描 PDF 使用本地 OCR。'
+        '基础文档解析可用。旧版 Office 转换尚未实现；扫描 PDF 可按场景模式使用本地 OCR。'
     },
     workflows: {
-      title: '使用场景',
+      title: 'PDF 解析模式',
-      description: '为聊天附件和知识库选择不同的解析深度',
+      description: '直接选择各场景处理 PDF 文本层与扫描页面的方式',
-      chat: '聊天附件',
+      chat: '聊天与成果文件',
-      chatDescription: '控制附件加入当前请求前的解析方式',
       knowledge: '知识库导入',
-      knowledgeDescription: '控制文档分块、索引和来源定位前的解析方式',
+      testChat: '测试聊天与成果模式',
+      testKnowledge: '测试知识库模式',
+      unsavedNotice: '当前有未保存修改；保存后可测试实际生效的模式。',
       chatOptions: {
-        auto: '自动解析（推荐）',
+        auto: '自动识别（推荐）',
-        fastText: '快速文本',
+        fastText: '仅使用文本层',
-        highFidelity: '高保真解析'
+        highFidelity: '全页 OCR'
+      },
+      chatDescriptions: {
+        auto:
+          '聊天附件和成果 PDF 优先使用文本层，仅对无有效文本的页面使用 OCR。',
+        fastText:
+          '聊天附件和成果 PDF 仅使用文本层，不运行 OCR；扫描件可能无法读取。',
+        highFidelity: '对 PDF 的每一页运行 OCR，速度较慢。'
       },
       knowledgeOptions: {
-        completeIndex: '完整索引（推荐）',
+        completeIndex: '自动识别（推荐）',
-        fastIndex: '快速索引',
+        fastIndex: '仅使用文本层',
-        highFidelity: '高保真索引'
+        highFidelity: '全页 OCR'
+      },
+      knowledgeDescriptions: {
+        'complete-index':
+          '优先使用 PDF 文本层，仅对无有效文本的页面使用 OCR。',
+        'fast-index':
+          '仅使用 PDF 文本层，不运行 OCR；扫描页面不会进入索引。',
+        'high-fidelity':
+          '对 PDF 的每一页运行 OCR 后再分块和建立索引，速度较慢。'
       }
     },
     ocr: {
       title: 'OCR 识别',
       description: '按需安装本地模型，在设备上识别扫描 PDF',
-      enabled: '启用本地 OCR',
-      enabledDescription:
-        '模型安装后仅在本机通过 ONNX Runtime WebAssembly 运行，识别时不会上传文档。',
-      model: '本地模型',
-      runtime: '运行时',
-      provider: {
-        title: 'OCR 来源',
-        description: '本地模型与远程服务二选一，切换后保存设置生效。',
-        local: '本地模型',
-        remote: '远程服务（即将支持）',
-        remoteDescription:
-          '远程服务将支持 MinerU、PaddleOCR-VL 等接口，当前版本暂不可选。'
-      },
       modelSelector: '当前 OCR 模型',
       modelSelectorDescription: '选择已保存，聊天附件和知识库将使用此模型。',
       pendingSelection: '模型选择尚未生效，点击“保存设置”后切换。',
       installedOption: '已安装',
       downloadableOption: '可下载',
+      unavailableOption: '当前版本不可用',
       openModelsDirectory: '打开模型目录',
       storagePrefix: '模型按需安装到',
       storageSuffix: '。可导出 ZIP，并在内网设备直接导入。',
@@ -283,6 +287,7 @@ export const settings = {
       },
       installed: '已安装并校验',
       download: '下载',
+      downloadAndSelect: '下载并启用',
       importZip: '导入 ZIP',
       exportZip: '导出 ZIP',
       delete: '删除',
@@ -290,14 +295,11 @@ export const settings = {
       cancel: '取消',
       openRepository: '打开 ModelScope',
       catalogUnavailable: '当前版本没有可用的 OCR 模型目录。',
-      mode: 'PDF OCR 策略',
+      selectedModelUnavailable:
-      modes: {
+        '已保存的 OCR 模型在当前版本不可用，请从上方选择并安装其他模型。',
-        auto: '自动，仅识别无有效文本的页面',
+      installBeforeSelecting: '请先下载该模型；下载完成后会自动设为当前模型。',
-        always: '始终识别所有页面',
+      privacyNotice:
-        disabled: '仅使用 PDF 文本层'
+        'OCR 只在需要时由上方场景模式启用，并始终在本机通过 ONNX Runtime WebAssembly 运行，不会上传文档。',
-      },
-      modelLicense:
-        '模型采用 Apache License 2.0，并在加载前校验 SHA-256。',
       operations: {
         preparing: '正在准备模型文件',
         downloading: '正在从 ModelScope 下载',
@@ -315,7 +317,9 @@ export const settings = {
       },
       notifications: {
         installed: '{{name}} 已安装',
+        installedAndSelected: '{{name}} 已安装并设为当前模型',
         importedZip: '{{name}} 已从 ZIP 导入',
+        importedAndSelected: '{{name}} 已导入并设为当前模型',
         exportedZip: '{{name}} 已导出为 ZIP',
         removed: 'OCR 模型已删除'
       }
@@ -323,10 +327,9 @@ export const settings = {
     advanced: {
       title: '高级解析设置',
       maximumPages: '单文档最大 OCR 页数',
-      concurrency: 'OCR 并发数',
       timeout: '每页 OCR 时间预算（秒）',
-      concurrencyHint:
+      description:
-        '当前 WASM 基线按页串行执行；该值为后续批处理和硬件加速保留。'
+        '页数限制只计算实际进入 OCR 的页面；单页超过时间预算时会终止本次解析。'
     },
     diagnostic: {
       title: '解析测试结果',

src/renderer/src/styles.css

@@ -5530,6 +5530,11 @@ details.settings-section > :not(summary) + :not(summary) {
   line-height: 1.5;
 }
 
+.document-parsing-workflows .field > label {
+  font-size: var(--font-caption);
+  font-weight: 650;
+}
+
 .document-parsing-workflows .field {
   padding: var(--space-3);
   border: 1px solid var(--border-default);
@@ -5559,37 +5564,6 @@ details.settings-section > :not(summary) + :not(summary) {
   overflow-wrap: anywhere;
 }
 
-.document-ocr-provider {
-  display: grid;
-  align-items: center;
-  padding: var(--space-3);
-  border: 1px solid var(--border-default);
-  border-radius: var(--radius-card);
-  background: var(--surface-raised);
-  grid-template-columns: minmax(0, 1fr) auto;
-  gap: var(--space-2) var(--space-4);
-}
-
-.document-ocr-provider > div:first-child {
-  display: grid;
-  gap: var(--space-1);
-}
-
-.document-ocr-provider strong {
-  color: var(--text-primary);
-  font-size: var(--font-body);
-}
-
-.document-ocr-provider small {
-  color: var(--text-muted);
-  font-size: var(--font-caption);
-  line-height: 1.5;
-}
-
-.document-ocr-provider > small {
-  grid-column: 1 / -1;
-}
-
 .document-ocr-model-selector {
   padding: var(--space-3);
   border: 1px solid var(--border-default);
@@ -5724,49 +5698,9 @@ details.settings-section > :not(summary) + :not(summary) {
   font-size: var(--font-caption);
 }
 
-.document-ocr-settings__options {
+.document-parsing-workflow-test {
-  display: grid;
+  width: fit-content;
-  align-items: start;
+  margin-top: auto;
-  grid-template-columns: repeat(2, minmax(0, 1fr));
-  gap: var(--space-4);
-}
-
-.document-ocr-settings__options > * {
-  min-height: 100%;
-  padding: var(--space-3);
-  border: 1px solid var(--border-default);
-  border-radius: var(--radius-card);
-  background: var(--surface-raised);
-}
-
-.document-ocr-settings__options .settings-checkbox {
-  display: grid;
-  align-items: start;
-  grid-template-columns: auto minmax(0, 1fr);
-  gap: var(--space-3);
-}
-
-.document-ocr-settings__options .settings-checkbox > input {
-  width: 16px;
-  height: 16px;
-  margin: 2px 0 0;
-  accent-color: var(--accent-solid);
-}
-
-.document-ocr-settings__options .settings-checkbox > span {
-  display: grid;
-  gap: var(--space-1);
-}
-
-.document-ocr-settings__options .settings-checkbox strong {
-  color: var(--text-primary);
-  font-size: var(--font-body);
-}
-
-.document-ocr-settings__options .settings-checkbox small {
-  color: var(--text-muted);
-  font-size: var(--font-caption);
-  line-height: 1.5;
 }
 
 .document-parsing-diagnostic-backdrop {
@@ -5866,8 +5800,7 @@ details.settings-section > :not(summary) + :not(summary) {
 @media (max-width: 720px) {
   .document-parsing-status__list,
   .document-parsing-grid,
-  .document-parsing-diagnostic dl,
+  .document-parsing-diagnostic dl {
-  .document-ocr-settings__options {
     grid-template-columns: minmax(0, 1fr);
   }
 

src/shared/contracts.ts

@@ -83,7 +83,8 @@ import type {
   DocumentOcrResult,
   DocumentParsingDiagnostic,
   DocumentParsingSettings,
-  DocumentParsingSnapshot
+  DocumentParsingSnapshot,
+  DocumentParsingTestPurpose
 } from './document-parsing-contracts'
 import type {
   KnowledgeChunkDeleteInput,
@@ -1159,7 +1160,9 @@ export type DesktopApi = {
     update: (
       input: DocumentParsingSettings
     ) => Promise<DocumentParsingSnapshot>
-    test: () => Promise<DocumentParsingDiagnostic | undefined>
+    test: (
+      purpose: DocumentParsingTestPurpose
+    ) => Promise<DocumentParsingDiagnostic | undefined>
     installOcrModel: (
       modelId: string
     ) => Promise<DocumentParsingSnapshot>

src/shared/document-parsing-contracts.ts

@@ -3,9 +3,15 @@ import { z } from 'zod'
 export const documentParsingPurposeSchema = z.enum([
   'chat-attachment',
   'knowledge-index',
+  'artifact-import',
   'diagnostic'
 ])
 
+export const documentParsingTestPurposeSchema = z.enum([
+  'chat-attachment',
+  'knowledge-index'
+])
+
 export const chatDocumentWorkflowSchema = z.enum([
   'auto',
   'fast-text',
@@ -18,14 +24,6 @@ export const knowledgeDocumentWorkflowSchema = z.enum([
   'high-fidelity'
 ])
 
-export const pdfOcrModeSchema = z.enum([
-  'auto',
-  'always',
-  'disabled'
-])
-
-export const documentOcrProviderSchema = z.literal('local')
-
 export const localOcrModelIdSchema = z
   .string()
   .min(1)
@@ -157,12 +155,8 @@ export const documentParsingSettingsSchema = z
   .object({
     chatWorkflow: chatDocumentWorkflowSchema,
     knowledgeWorkflow: knowledgeDocumentWorkflowSchema,
-    pdfOcrMode: pdfOcrModeSchema,
-    ocrProvider: documentOcrProviderSchema,
-    localOcrEnabled: z.boolean(),
     localOcrModelId: localOcrModelIdSchema,
     maximumPages: z.number().int().min(1).max(500),
-    ocrConcurrency: z.number().int().min(1).max(4),
     pageTimeoutSeconds: z.number().int().min(10).max(300)
   })
   .strict()
@@ -199,7 +193,7 @@ export const documentParsingSnapshotSchema = z
 
 export const documentParsingTestInputSchema = z
   .object({
-    purpose: documentParsingPurposeSchema.default('diagnostic')
+    purpose: documentParsingTestPurposeSchema
   })
   .strict()
 
@@ -252,6 +246,27 @@ export const documentOcrRequestSchema = z
     pageTimeoutSeconds: z.number().int().min(10).max(300)
   })
   .strict()
+  .superRefine((request, context) => {
+    if (!request.pageNumbers) {
+      return
+    }
+    if (
+      new Set(request.pageNumbers).size !== request.pageNumbers.length
+    ) {
+      context.addIssue({
+        code: 'custom',
+        path: ['pageNumbers'],
+        message: 'OCR 页码不能重复'
+      })
+    }
+    if (request.pageNumbers.length > request.maximumPages) {
+      context.addIssue({
+        code: 'custom',
+        path: ['pageNumbers'],
+        message: 'OCR 页数超过当前文档限制'
+      })
+    }
+  })
 
 export const documentOcrSectionSchema = z
   .object({
@@ -280,6 +295,9 @@ export const documentOcrFailureSchema = z
 export type DocumentParsingPurpose = z.infer<
   typeof documentParsingPurposeSchema
 >
+export type DocumentParsingTestPurpose = z.infer<
+  typeof documentParsingTestPurposeSchema
+>
 export type DocumentParsingSettings = z.infer<
   typeof documentParsingSettingsSchema
 >