chore: bump to v2.2.27, 修复维护工具模型检测与中文显示

package.json

@@ -1,7 +1,7 @@
 {
   "name": "llm-wiki",
   "private": true,
-  "version": "2.2.26",
+  "version": "2.2.27",
   "type": "module",
   "scripts": {
     "dev": "vite",

src-tauri/tauri.conf.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://schema.tauri.app/config/2",
   "productName": "QMaiWrite",
-  "version": "2.2.26",
+  "version": "2.2.27",
   "identifier": "com.qingmuai.writer",
   "build": {
     "beforeDevCommand": "npm run dev",

src/components/settings/sections/maintenance-section.tsx

@@ -20,6 +20,7 @@ import { Button } from "@/components/ui/button"
 import { Label } from "@/components/ui/label"
 import { useWikiStore } from "@/stores/wiki-store"
 import { hasUsableLlm } from "@/lib/has-usable-llm"
+import { resolveDefaultModel } from "@/lib/novel/model-resolver"
 import { runDuplicateDetection } from "@/lib/dedup-runner"
 import { addNotDuplicate } from "@/lib/dedup-storage"
 import {
@@ -100,7 +101,8 @@ export function MaintenanceSection() {
     return () => clearInterval(id)
   }, [])
 
-  const llmReady = hasUsableLlm(llmConfig, providerConfigs)
+  const resolvedLlmConfig = useMemo(() => resolveDefaultModel(llmConfig), [llmConfig])
+  const llmReady = hasUsableLlm(resolvedLlmConfig, providerConfigs)
   const projectReady = !!project
   const projectScanState = project && scanState.projectPath === project.path ? scanState : emptyScanState
   const { scanning, scanError, groups, scanCompleted } = projectScanState

src/lib/changelog.ts

@@ -78,6 +78,19 @@ const TWO_POINT_TWO_ELEVEN_CHANGELOG: ChangelogEntry = {
   },
 }
 
+const TWO_POINT_TWO_TWENTY_SEVEN_CHANGELOG: ChangelogEntry = {
+  version: "2.2.27",
+  date: "2026-06-28",
+  highlights: {
+    en: [],
+    zh: [
+      "修复维护工具「检测重复实体」模型检测逻辑：现在会正确解析 AI 会话模型和默认模型，不再误报「请先配置大模型」。",
+      "修复扫描结果展示英文的问题：重复检测和合并的 LLM 提示词改为中文，LLM 返回的重复原因描述现在显示中文。",
+      "修复队列错误信息为中文：合并任务失败时的错误提示改为中文。",
+    ],
+  },
+}
+
 const TWO_POINT_TWO_TWENTY_SIX_CHANGELOG: ChangelogEntry = {
   version: "2.2.26",
   date: "2026-06-28",
@@ -638,6 +651,7 @@ export const CHANGELOG: ChangelogEntry[] = [
 ]
 
 export function currentVersionChangelog(version: string): ChangelogEntry[] {
+  if (version === TWO_POINT_TWO_TWENTY_SEVEN_CHANGELOG.version) return [TWO_POINT_TWO_TWENTY_SEVEN_CHANGELOG]
   if (version === TWO_POINT_TWO_TWENTY_SIX_CHANGELOG.version) return [TWO_POINT_TWO_TWENTY_SIX_CHANGELOG]
   if (version === TWO_POINT_TWO_TWENTY_FIVE_CHANGELOG.version) return [TWO_POINT_TWO_TWENTY_FIVE_CHANGELOG]
   if (version === TWO_POINT_TWO_TWENTY_FOUR_CHANGELOG.version) return [TWO_POINT_TWO_TWENTY_FOUR_CHANGELOG]
@@ -669,6 +683,7 @@ export function currentVersionChangelog(version: string): ChangelogEntry[] {
 
 export function allChangelog(): ChangelogEntry[] {
   return [
+    TWO_POINT_TWO_TWENTY_SEVEN_CHANGELOG,
     TWO_POINT_TWO_TWENTY_SIX_CHANGELOG,
     TWO_POINT_TWO_TWENTY_FIVE_CHANGELOG,
     TWO_POINT_TWO_TWENTY_FOUR_CHANGELOG,

src/lib/dedup-queue.ts

@@ -302,7 +302,7 @@ async function processNext(projectId: string): Promise<void> {
 
   if (!pp) {
     next.status = "failed"
-    next.error = "Project not found in registry (was it deleted?)"
+    next.error = "项目未在注册表中找到（可能已被删除？）"
     await saveQueue(currentProjectPath)
     processNext(projectId)
     return
@@ -318,7 +318,7 @@ async function processNext(projectId: string): Promise<void> {
 
   if (!hasUsableLlm(llmConfig, state.providerConfigs)) {
     next.status = "failed"
-    next.error = "LLM not configured — set API key in Settings"
+    next.error = "LLM 未配置，请在设置中配置大模型提供方"
     processing = false
     await saveQueue(pp)
     return

src/lib/dedup.ts

@@ -168,34 +168,36 @@ function truncate(s: string, max: number): string {
 // Stage 2: LLM-driven duplicate detection
 // ──────────────────────────────────────────────────────────────────
 
-const DETECTOR_SYSTEM_PROMPT = `You are a wiki maintenance assistant. You will receive a list of entity / concept pages from a wiki. Identify groups of slugs that likely refer to the same underlying topic under different names — for example:
+const DETECTOR_SYSTEM_PROMPT = `你是一个维基维护助手。你将收到一个维基中的实体/概念页面列表。请找出那些很可能指向同一主题但名称不同的 slug 分组——例如：
 
-- Same name in two languages (English vs Chinese, etc.)
-- Plural vs singular form (e.g. "dpao" vs "dpaos")
-- Abbreviation vs full form (e.g. "vfa" vs "volatile-fatty-acids")
-- Synonyms in the same language
-- The same proper noun spelled differently
+- 同一名称的不同语言版本（中英文等）
+- 单复数形式（如 "dpao" 和 "dpaos"）
+- 缩写与全称（如 "vfa" 和 "volatile-fatty-acids"）
+- 同义词
+- 同一专有名词的不同拼写
 
-Output ONLY valid JSON. No prose, no markdown fences, no explanation outside the JSON. The schema is:
+只输出有效的 JSON。不要输出散文、markdown 代码块或 JSON 之外的任何解释。JSON 结构如下：
 
 {
   "groups": [
     {
       "slugs": ["slug-a", "slug-b"],
-      "reason": "Both refer to X; first is English, second is Chinese.",
+      "reason": "两个页面都指向 X；第一个是英文，第二个是中文。",
       "confidence": "high"
     }
   ]
 }
 
-Rules:
-- Only include groups of 2 or more slugs from the input list.
-- "high" = clearly the same entity, only naming differs.
-- "medium" = likely the same but context-dependent.
-- "low" = uncertain; user should review carefully.
-- Never invent slugs that aren't in the input.
-- If no duplicates exist, output {"groups": []}.
-- Pages of different \`type\` (e.g. an entity and a concept) usually should NOT be grouped — only group across types when they're unambiguously the same thing.`
+规则：
+- 只包含输入列表中 2 个或更多 slug 的分组。
+- "high" = 明显是同一实体，只是命名不同。
+- "medium" = 可能是同一实体，但需要结合上下文判断。
+- "low" = 不确定，需要用户仔细审查。
+- 不要编造输入列表中不存在的 slug。
+- 如果没有重复项，输出 {"groups": []}。
+- 不同 \`type\`（如 entity 和 concept）的页面通常不应分在一组——只有在明确是同一事物时才跨类型分组。
+
+重要：reason 字段必须使用中文描述。`
 
 /**
  * Run the LLM duplicate-detector. The caller hands in summaries
@@ -318,17 +320,17 @@ function normalizeGroupKey(slugs: string[]): string {
 // Stage 3: merge a confirmed duplicate group
 // ──────────────────────────────────────────────────────────────────
 
-const MERGER_SYSTEM_PROMPT = `You are a wiki maintenance assistant. You will be given several wiki pages that all describe the same entity or concept under different names. Merge them into a single coherent wiki page.
+const MERGER_SYSTEM_PROMPT = `你是一个维基维护助手。你将收到几个描述同一实体或概念但名称不同的维基页面。请将它们合并为一个连贯的维基页面。
 
-Output the COMPLETE merged file (frontmatter + body). The first character of your response MUST be "-" (the opening of "---"). No preamble, no explanation outside the file.
+输出完整的合并文件（frontmatter + 正文）。你回复的第一个字符必须是 "-"（"---" 的开头）。不要输出前言或文件之外的任何解释。
 
-Rules:
-- Preserve every distinct factual claim from every input page.
-- Eliminate redundancy (don't say the same thing twice across sections).
-- Reorganize sections so the structure is logical for the unified topic, not a concatenation of inputs.
-- Use [[wikilink]] syntax in the body where the inputs did.
-- Frontmatter: keep the standard fields (type, title, created, updated, tags, related, sources). The caller will overwrite sources / tags / related / updated with deterministic unions afterward — your job is to produce a sensible body and reasonable frontmatter shape.
-- Pick the most descriptive title. If the inputs use different languages, prefer the language that matches the majority of the body content.`
+规则：
+- 保留每个输入页面中所有不同的事实性陈述。
+- 消除冗余（不要在多个章节中重复相同的内容）。
+- 重新组织章节结构，使其对统一后的主题具有逻辑性，而不是简单拼接输入内容。
+- 在正文中使用 [[wikilink]] 语法（如果输入中使用了的话）。
+- Frontmatter：保留标准字段（type, title, created, updated, tags, related, sources）。调用方会在之后用确定性合并覆盖 sources / tags / related / updated 字段——你的任务是生成合理的正文和合理的 frontmatter 结构。
+- 选择最具描述性的标题。如果输入使用了不同语言，优先选择与正文内容多数语言匹配的语言。`
 
 const FIELDS_TO_UNION = ["sources", "tags", "related"] as const