fix(deepseek): fall back to reasoning "搜索到 XX 个网页" panel for citations

When the standard search-results panel never opens, click the reasoning-chain "搜索到 N 个网页" affordance to reveal the citation source list, and prefer the revealed panel links over stale snapshot links so citations point at the right URLs. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-26 10:18:44 +08:00
parent 05aa982f76
commit 2fd6267632
2 changed files with 401 additions and 6 deletions
@@ -1,4 +1,48 @@
-import { describe, expect, it } from "vitest";
+import { describe, expect, it, vi } from "vitest";
+
+vi.mock("electron/main", () => ({
+  app: {
+    getPath: () => "/tmp/geo-rankly-deepseek-test",
+    getVersion: () => "0.0.0-test",
+    isPackaged: false,
+    on: vi.fn(),
+    whenReady: () => Promise.resolve(),
+  },
+  BrowserWindow: vi.fn(),
+  WebContentsView: vi.fn(),
+  ipcMain: {
+    handle: vi.fn(),
+    on: vi.fn(),
+    removeHandler: vi.fn(),
+  },
+  Menu: {
+    buildFromTemplate: vi.fn(() => ({ popup: vi.fn() })),
+    setApplicationMenu: vi.fn(),
+  },
+  Tray: vi.fn(),
+  nativeTheme: {
+    on: vi.fn(),
+    shouldUseDarkColors: false,
+  },
+  safeStorage: {
+    decryptString: vi.fn(),
+    encryptString: vi.fn(),
+    isEncryptionAvailable: () => false,
+  },
+  session: {
+    fromPartition: vi.fn(() => ({
+      cookies: {
+        get: vi.fn(),
+        remove: vi.fn(),
+        set: vi.fn(),
+      },
+      webRequest: {
+        onBeforeSendHeaders: vi.fn(),
+        onHeadersReceived: vi.fn(),
+      },
+    })),
+  },
+}));

 import { __deepseekTestUtils } from "./deepseek";
 import { getMonitorAdapter } from "./index";
@@ -11,9 +55,11 @@ const {
  dedupeSourceItems,
  classifyDeepseekSources,
  mergeDeepSeekRawSourceLinks,
+  resolveDeepSeekCitationPanelLinks,
  isObservationComplete,
  isDeepSeekToggleSelected,
  isDeepSeekWebpagesTriggerText,
+  isDeepSeekReasoningSearchTriggerText,
 } = __deepseekTestUtils;

 describe("deepseek adapter helpers", () => {
@@ -67,6 +113,74 @@ describe("deepseek adapter helpers", () => {
    expect(classified.searchResults[0]?.url).toBe("https://example.com/search");
  });

+  it("uses revealed xx webpages panel links instead of stale snapshot links for citation sources", () => {
+    const citationPanelLinks = resolveDeepSeekCitationPanelLinks(
+      [
+        {
+          url: "https://stale.example/from-snapshot",
+          title: "Stale snapshot result",
+          text: "Wrong result from the page snapshot",
+          siteName: "Stale",
+          kind: "search",
+        },
+      ],
+      [
+        {
+          url: "https://right.example/from-xx-webpages",
+          title: "Right side panel result",
+          text: "Correct result from DeepSeek search results",
+          siteName: "Right",
+          kind: "search",
+        },
+      ],
+    );
+    const classified = classifyDeepseekSources({
+      inlineLinks: [
+        {
+          url: "https://inline.example/body-link",
+          title: "Inline body link",
+          text: "[1]",
+          siteName: "Inline",
+          kind: "inline",
+        },
+      ],
+      sourcePanelLinks: [],
+      searchPanelLinks: citationPanelLinks,
+    });
+
+    expect(citationPanelLinks.map((item) => item.url)).toEqual([
+      "https://right.example/from-xx-webpages",
+    ]);
+    expect(classified.citations.map((item) => item.url)).toEqual([
+      "https://right.example/from-xx-webpages",
+    ]);
+    expect(classified.searchResults.map((item) => item.url)).toEqual([
+      "https://right.example/from-xx-webpages",
+    ]);
+    expect(classified.inlineCitationCandidates.map((item) => item.url)).toEqual([
+      "https://inline.example/body-link",
+    ]);
+  });
+
+  it("falls back to snapshot search links when the revealed xx webpages panel is unavailable", () => {
+    const citationPanelLinks = resolveDeepSeekCitationPanelLinks(
+      [
+        {
+          url: "https://snapshot.example/result",
+          title: "Snapshot result",
+          text: "Fallback result",
+          siteName: "Snapshot",
+          kind: "search",
+        },
+      ],
+      [],
+    );
+
+    expect(citationPanelLinks.map((item) => item.url)).toEqual([
+      "https://snapshot.example/result",
+    ]);
+  });
+
  it("enriches numeric inline citations with source metadata from revealed search cards", () => {
    const classified = classifyDeepseekSources({
      inlineLinks: [
@@ -267,6 +381,12 @@ describe("deepseek adapter helpers", () => {
    expect(isDeepSeekWebpagesTriggerText("搜索到 39 个网页")).toBe(false);
  });

+  it("treats reasoning 搜索到 xx 个网页 as a separate citation-source fallback trigger", () => {
+    expect(isDeepSeekReasoningSearchTriggerText("搜索到 30 个网页")).toBe(true);
+    expect(isDeepSeekReasoningSearchTriggerText("57 个网页")).toBe(false);
+    expect(isDeepSeekReasoningSearchTriggerText("已阅读 57 个网页")).toBe(false);
+  });
+
  it("normalizes DeepSeek model labels without mistaking search toggles for models", () => {
    const snapshot = {
      url: "https://chat.deepseek.com/",
@@ -18,6 +18,9 @@ const DEEPSEEK_SEARCH_RESULTS_HEADING_PATTERN = /^(?:搜索结果|search results
 const DEEPSEEK_WEBPAGES_TRIGGER_ZH_PATTERN = /^(?:已阅读\s*)?\d+\s*个网页$/i
 const DEEPSEEK_WEBPAGES_TRIGGER_EN_PATTERN =
  /^(?:(?:read|viewed)\s*)?\d+\s*(?:web\s*pages?|webpages?)$/i
+const DEEPSEEK_REASONING_SEARCH_TRIGGER_ZH_PATTERN = /^搜索到\s*\d+\s*个网页$/i
+const DEEPSEEK_REASONING_SEARCH_TRIGGER_EN_PATTERN =
+  /^(?:found|searched)\s*\d+\s*(?:web\s*pages?|webpages?)$/i
 const DEEPSEEK_BUSY_TEXT_PATTERN =
  /(searching|thinking|generating|please wait until the message is generated|搜索中|思考中|生成中|停止生成|stop generating)/i
 const DEEPSEEK_BUSY_DESCRIPTOR_PATTERN =
@@ -354,6 +357,18 @@ function isDeepSeekWebpagesTriggerText(value: string | null | undefined): boolea
  )
 }

+function isDeepSeekReasoningSearchTriggerText(value: string | null | undefined): boolean {
+  const normalized = normalizeText(value)
+  if (!normalized) {
+    return false
+  }
+
+  return (
+    DEEPSEEK_REASONING_SEARCH_TRIGGER_ZH_PATTERN.test(normalized) ||
+    DEEPSEEK_REASONING_SEARCH_TRIGGER_EN_PATTERN.test(normalized)
+  )
+}
+
 function buildSourceItem(source: DeepSeekSourceInput): MonitoringSourceItem | null {
  const url = normalizeUrl(getSourceNormalizedUrl(source) ?? source.url)
  if (!url) {
@@ -466,6 +481,16 @@ function classifyDeepseekSources(input: {
  }
 }

+function resolveDeepSeekCitationPanelLinks(
+  snapshotSearchPanelLinks: DeepSeekRawSourceLink[],
+  revealedSearchPanelLinks: DeepSeekRawSourceLink[],
+): DeepSeekRawSourceLink[] {
+  if (revealedSearchPanelLinks.length > 0) {
+    return mergeDeepSeekRawSourceLinks(revealedSearchPanelLinks)
+  }
+  return mergeDeepSeekRawSourceLinks(snapshotSearchPanelLinks)
+}
+
 function extractQuestionText(payload: Record<string, unknown>): string {
  const candidates = [
    payload.question_text,
@@ -1127,6 +1152,247 @@ async function maybeRevealDeepSeekSources(page: PlaywrightPage): Promise<void> {
  }
 }

+async function maybeRevealDeepSeekReasoningSearchSources(page: PlaywrightPage): Promise<void> {
+  if (await hasDeepSeekSearchResultsPanel(page)) {
+    return
+  }
+
+  const reasoningSearchTrigger = await page
+    .evaluate(() => {
+      const normalize = (value: unknown): string | null => {
+        if (typeof value !== 'string') {
+          return null
+        }
+        const trimmed = value.trim().replace(/\s+/g, ' ')
+        return trimmed || null
+      }
+      const isVisible = (node: Element | null): node is HTMLElement => {
+        if (!(node instanceof HTMLElement)) {
+          return false
+        }
+        const style = window.getComputedStyle(node)
+        if (style.display === 'none' || style.visibility === 'hidden' || style.opacity === '0') {
+          return false
+        }
+        const rect = node.getBoundingClientRect()
+        return rect.width > 0 && rect.height > 0
+      }
+      const isTriggerText = (value: unknown): boolean => {
+        const text = normalize(value)
+        return Boolean(
+          text &&
+            (/^搜索到\s*\d+\s*个网页$/i.test(text) ||
+              /^(?:found|searched)\s*\d+\s*(?:web\s*pages?|webpages?)$/i.test(text)),
+        )
+      }
+      const isReasoningNode = (node: HTMLElement): boolean => {
+        if (
+          node.closest(
+            ".ds-think-content, [class*='think-content'], [class*='reasoning-content']",
+          )
+        ) {
+          return true
+        }
+
+        let current: HTMLElement | null = node
+        for (let depth = 0; current && depth < 8; depth += 1) {
+          const text = normalize(current.textContent || '')
+          const className = String(current.className || '')
+          if (
+            /think|reason|analysis|思考/i.test(className) ||
+            Boolean(text && /已思考|思考过程|搜索到\s*\d+\s*个网页/.test(text))
+          ) {
+            return true
+          }
+          current = current.parentElement
+        }
+
+        return false
+      }
+      const isClickable = (node: HTMLElement): boolean => {
+        const tagName = node.tagName.toLowerCase()
+        const role = normalize(node.getAttribute('role'))
+        const style = window.getComputedStyle(node)
+        return (
+          tagName === 'button' ||
+          tagName === 'a' ||
+          role === 'button' ||
+          node.tabIndex >= 0 ||
+          style.cursor === 'pointer' ||
+          typeof (node as { onclick?: unknown }).onclick === 'function'
+        )
+      }
+      const findClickTarget = (node: HTMLElement): HTMLElement | null => {
+        let fallback: HTMLElement | null = null
+        const nextParent = (element: HTMLElement): HTMLElement | null =>
+          element.parentNode instanceof HTMLElement ? element.parentNode : null
+        let current: HTMLElement | null = node
+        for (let depth = 0; current !== null && depth < 10; depth += 1) {
+          const element: HTMLElement = current
+          if (!isVisible(element)) {
+            current = nextParent(element)
+            continue
+          }
+          const currentText = normalize(
+            element.textContent || element.getAttribute('aria-label') || '',
+          )
+          if (!fallback && currentText && isTriggerText(currentText)) {
+            fallback = element
+          }
+          if (isClickable(element)) {
+            return element
+          }
+          current = nextParent(element)
+        }
+        return fallback
+      }
+
+      let bestTarget: HTMLElement | null = null
+      let bestScore = -1
+      for (const node of Array.from(
+        document.querySelectorAll("span, div, button, a, [role='button']"),
+      ).slice(0, 1600)) {
+        if (!(node instanceof HTMLElement) || !isVisible(node)) {
+          continue
+        }
+        const text = normalize(node.textContent || node.getAttribute('aria-label') || '')
+        if (!isTriggerText(text) || !isReasoningNode(node)) {
+          continue
+        }
+
+        const target = findClickTarget(node)
+        if (!(target instanceof HTMLElement)) {
+          continue
+        }
+
+        const rect = target.getBoundingClientRect()
+        let score = 0
+        if (isClickable(target)) {
+          score += 32
+        }
+        if (rect.left >= window.innerWidth * 0.18 && rect.left <= window.innerWidth * 0.75) {
+          score += 18
+        }
+        if (rect.top >= window.innerHeight * 0.2) {
+          score += 12
+        }
+        if (rect.width <= 360) {
+          score += 8
+        }
+        if (target.closest(".ds-think-content, [class*='think-content']")) {
+          score += 16
+        }
+        if (score > bestScore) {
+          bestScore = score
+          bestTarget = target
+        }
+      }
+
+      if (!bestTarget) {
+        return null
+      }
+
+      bestTarget.scrollIntoView({ block: 'center', inline: 'center' })
+      const rect = bestTarget.getBoundingClientRect()
+      return {
+        x: rect.left + rect.width / 2,
+        y: rect.top + rect.height / 2,
+      }
+    })
+    .catch(() => null)
+
+  if (!reasoningSearchTrigger || typeof reasoningSearchTrigger !== 'object') {
+    return
+  }
+
+  await page.mouse.click(reasoningSearchTrigger.x, reasoningSearchTrigger.y).catch(() => undefined)
+  await page.waitForTimeout(500).catch(() => undefined)
+  if (await waitForDeepSeekSearchResultsPanel(page, 5_000)) {
+    return
+  }
+
+  await page
+    .evaluate(() => {
+      const normalize = (value: unknown): string | null => {
+        if (typeof value !== 'string') {
+          return null
+        }
+        const trimmed = value.trim().replace(/\s+/g, ' ')
+        return trimmed || null
+      }
+      const isVisible = (node: Element | null): node is HTMLElement => {
+        if (!(node instanceof HTMLElement)) {
+          return false
+        }
+        const style = window.getComputedStyle(node)
+        if (style.display === 'none' || style.visibility === 'hidden' || style.opacity === '0') {
+          return false
+        }
+        const rect = node.getBoundingClientRect()
+        return rect.width > 0 && rect.height > 0
+      }
+      const isTriggerText = (value: unknown): boolean => {
+        const text = normalize(value)
+        return Boolean(text && /^搜索到\s*\d+\s*个网页$/i.test(text))
+      }
+      const isClickable = (node: HTMLElement): boolean => {
+        const tagName = node.tagName.toLowerCase()
+        const role = normalize(node.getAttribute('role'))
+        const style = window.getComputedStyle(node)
+        return (
+          tagName === 'button' ||
+          tagName === 'a' ||
+          role === 'button' ||
+          node.tabIndex >= 0 ||
+          style.cursor === 'pointer' ||
+          typeof (node as { onclick?: unknown }).onclick === 'function'
+        )
+      }
+      const dispatchClick = (target: HTMLElement): void => {
+        target.dispatchEvent(
+          new PointerEvent('pointerdown', { bubbles: true, cancelable: true, pointerId: 1 }),
+        )
+        target.dispatchEvent(new MouseEvent('mousedown', { bubbles: true, cancelable: true }))
+        target.dispatchEvent(
+          new PointerEvent('pointerup', { bubbles: true, cancelable: true, pointerId: 1 }),
+        )
+        target.dispatchEvent(new MouseEvent('mouseup', { bubbles: true, cancelable: true }))
+        target.dispatchEvent(new MouseEvent('click', { bubbles: true, cancelable: true }))
+      }
+
+      for (const node of Array.from(
+        document.querySelectorAll("span, div, button, a, [role='button']"),
+      ).slice(0, 1600)) {
+        if (!(node instanceof HTMLElement) || !isVisible(node)) {
+          continue
+        }
+        const text = normalize(node.textContent || node.getAttribute('aria-label') || '')
+        if (!isTriggerText(text)) {
+          continue
+        }
+        let target: HTMLElement | null = node
+        let fallback: HTMLElement | null = node
+        for (let depth = 0; target && depth < 10; depth += 1) {
+          if (isVisible(target) && isClickable(target)) {
+            dispatchClick(target)
+            return
+          }
+          if (isVisible(target)) {
+            fallback = target
+          }
+          target = target.parentElement
+        }
+        if (fallback) {
+          dispatchClick(fallback)
+          return
+        }
+      }
+    })
+    .catch(() => undefined)
+  await page.waitForTimeout(500).catch(() => undefined)
+  await waitForDeepSeekSearchResultsPanel(page).catch(() => undefined)
+}
+
 async function maybeExpandDeepSeekReasoning(page: PlaywrightPage): Promise<void> {
  await page
    .evaluate(() => {
@@ -2561,19 +2827,26 @@ export const deepseekAdapter: MonitorAdapter = {
    context.reportProgress('deepseek.reveal_sources')
    await maybeExpandDeepSeekReasoning(page).catch(() => undefined)
    await maybeRevealDeepSeekSources(page).catch(() => undefined)
-    const revealedSearchPanelLinks = await collectDeepSeekSearchResultsPanelLinks(page).catch(
+    let revealedSearchPanelLinks = await collectDeepSeekSearchResultsPanelLinks(page).catch(
      () => [],
    )
+    if (revealedSearchPanelLinks.length === 0) {
+      await maybeRevealDeepSeekReasoningSearchSources(page).catch(() => undefined)
+      revealedSearchPanelLinks = await collectDeepSeekSearchResultsPanelLinks(page).catch(
+        () => [],
+      )
+    }
    const finalSnapshot = await readDeepSeekPageSnapshot(page, questionText).catch(
      () => waitResult.finalSnapshot,
    )
    const bestSnapshot = selectBetterSnapshot(waitResult.finalSnapshot, finalSnapshot)
+    const citationPanelLinks = resolveDeepSeekCitationPanelLinks(
+      bestSnapshot.searchPanelLinks,
+      revealedSearchPanelLinks,
+    )
    const mergedSnapshot: DeepSeekPageSnapshot = {
      ...bestSnapshot,
-      searchPanelLinks: mergeDeepSeekRawSourceLinks(
-        bestSnapshot.searchPanelLinks,
-        revealedSearchPanelLinks,
-      ),
+      searchPanelLinks: citationPanelLinks,
    }
    const classifiedSources = classifyDeepseekSources({
      inlineLinks: mergedSnapshot.inlineLinks,
@@ -2639,7 +2912,9 @@ export const __deepseekTestUtils = {
  dedupeSourceItems,
  classifyDeepseekSources,
  mergeDeepSeekRawSourceLinks,
+  resolveDeepSeekCitationPanelLinks,
  isObservationComplete,
  isDeepSeekToggleSelected,
  isDeepSeekWebpagesTriggerText,
+  isDeepSeekReasoningSearchTriggerText,
 }