fix: stabilize AI-aware page translation cache title

[frogGuaGuaGuaGua]

Type of Changes

• 🐛 Bug fix (fix)
• ✅ Tests

Description

This fixes the unstable same-page cache misses reported in #1254 when AI Content Aware is enabled.

Root cause

getOrFetchArticleData() cached article textContent by URL, but it kept re-reading the live document.title on every request.
During page translation, the tab title itself can be translated, so the same page could produce two different article-title contexts:

1. first translation pass → original source title
2. second pass after toggling back on → translated browser title

That drift changed the AI-aware cache key for otherwise identical body text, causing needless cache misses and repeated API calls.

Fix

• Cache the article title together with the article text content for a given URL
• Reuse that stable source title while the URL stays the same
• Add regression tests covering:
  • stable title reuse on the same URL
  • unchanged live-title behavior when AI Content Aware is off
  • cache refresh when the URL changes

Fixes #1254.

Testing

• vitest run src/utils/host/translate/__tests__/article-context.test.ts src/utils/host/__tests__/translate-text.test.tsx src/entrypoints/host.content/translation-control/__tests__/page-translation-title.test.ts
• eslint src/utils/host/translate/article-context.ts src/utils/host/translate/__tests__/article-context.test.ts

---

Summary by cubic

Stabilizes page translation caching by storing the source title with article text per URL, preventing cache key drift when the browser title changes during translation. Fixes #1254 and reduces duplicate API calls.

• Bug Fixes
  • Cache { url, title, textContent } in getOrFetchArticleData() and reuse the cached title while the URL stays the same.
  • Keep live document.title behavior when Content Aware is off.
  • Refresh cached data when the URL changes.
  • Added tests for stable title reuse, live-title behavior, and URL-change refresh.

^{Written for commit 63235c2. Summary will update on new commits.}

[changeset-bot]

🦋 Changeset detected

Latest commit: 63235c2

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@read-frog/extension	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[github-actions]

Contributor trust score

40/100 — Moderate

This score estimates contributor familiarity with mengxi-ream/read-frog using public GitHub signals. It is advisory only and does not block merges automatically.

Outcome

• PR: fix: stabilize AI-aware page translation cache title #1280 — fix: stabilize AI-aware page translation cache title
• Author: @frogGuaGuaGuaGua
• Trust label: contrib-trust:moderate
• Maintainer review: not required
• Override label: add trust-check:skip to stop future trust automation updates

Score breakdown

Dimension	Score	Signals
Repo familiarity	21/35	commits in repo, merged PRs, reviews
Community standing	11/25	account age, followers, repo role
OSS influence	0/20	stars on owned non-fork repositories
PR track record	8/20	merge rate across resolved PRs in this repo

Signals used

• Repo commits: 6 (author commits reachable from the repository default branch)
• Repo PR history: merged 6, open 2, closed-unmerged 9
• Repo reviews: 0
• PR changed lines: 110 (+106 / -4)
• Repo permission: write
• Followers: 1
• Account age: 0 months
• Owned non-fork repos considered: max 0, total 0 (none)

Policy

• Low-score review threshold: < 30
• Auto-close: score < 20 and changed lines > 1000
• Policy version: v1.1

Updated automatically when the PR changes or when a maintainer reruns the workflow.

[dosubot]

Related Documentation

1 document(s) may need updating based on files changed in this PR:

Read Frog - Open Source Immersive Translate

how translation cache works

View Suggested Changes

@@ -52,13 +52,13 @@
 This means that the same text will have different cache entries when:
 1. It appears on different pages (different article titles)
 2. It appears in articles with different content
-3. The page title changes
 
 **Article Context Extraction:**
 - Article context is extracted using the `article-context.ts` module
 - Uses Mozilla Readability to extract clean article content
 - Falls back to `document.body.textContent` if Readability fails
-- Context is cached per URL to avoid repeated extraction
+- Both the article title and text content are cached together per URL to ensure cache stability
+- The cached title is reused for the same URL even if the browser tab title is translated during page translation
 
 **Impact on Cache Behavior:**
 - Cache entries are more specific when AI Content Aware is enabled with LLM providers

[Accept] [Decline]

Note: You must be authenticated to accept/decline updates.

^{How did I do? Any feedback?}  

[cubic-dev-ai]

No issues found across 3 files