feat(plugin-slimsearch): improve default querySplitter, close #299

plugins/search/plugin-slimsearch/src/client/helpers/search.ts

@@ -5,6 +5,7 @@ import { useRouteLocale } from 'vuepress/client'
 
 import { isFunction } from 'vuepress/shared'
 import type { SearchResult, WorkerSearchOptions } from '../typings/index.js'
+import { defaultQuerySplitter } from '../utils/index.js'
 
 declare const __VUEPRESS_DEV__: boolean
 
@@ -33,7 +34,9 @@ export interface SearchOptions extends SearchLocaleOptions {
   locales?: Record<string, SearchLocaleOptions>
 }
 
-const searchOptions: Ref<SearchOptions> = ref({})
+const searchOptions: Ref<SearchOptions> = ref({
+  querySplitter: (query) => Promise.resolve(defaultQuerySplitter(query)),
+})
 
 const slimsearchSymbol: InjectionKey<Ref<SearchOptions>> = Symbol(
   __VUEPRESS_DEV__ ? 'slimsearch' : '',

plugins/search/plugin-slimsearch/src/client/utils/index.ts

@@ -2,3 +2,4 @@ export * from './createSearchWorker.js'
 export * from './getResultPath.js'
 export * from './isFocusingTextControl.js'
 export * from './isKeyMatched.js'
+export * from './querySplitter.js'

plugins/search/plugin-slimsearch/src/client/utils/querySplitter.ts

@@ -0,0 +1,16 @@
+const CJK_REGEXP =
+  /[\u4e00-\u9fff\u3400-\u4dbf\u3040-\u309f\u30a0-\u30ff\uac00-\ud7af]/
+
+export const defaultQuerySplitter = (query: string): string[] =>
+  query
+    .split(/\s+/)
+    .map((word) => {
+      if (word.length > 3) {
+        const chars = word.split('')
+
+        if (chars.every((char) => CJK_REGEXP.test(char))) return chars
+      }
+
+      return word
+    })
+    .flat()

plugins/search/plugin-slimsearch/tests/querySliptter.spec.ts

@@ -0,0 +1,18 @@
+import { expect, it } from 'vitest'
+import { defaultQuerySplitter } from '../src/client/utils/querySplitter.js'
+
+it('defaultQuerySplitter', () => {
+  const testCases: [string, string[]][] = [
+    ['你好世界', ['你', '好', '世', '界']],
+    ['你好 世界', ['你好', '世界']],
+    ['hello world', ['hello', 'world']],
+    ['hello 你好 world 世界', ['hello', '你好', 'world', '世界']],
+    ['hi hello 你好 世界', ['hi', 'hello', '你好', '世界']],
+    ['', ['']],
+  ]
+
+  for (const [query, expected] of testCases) {
+    const result = defaultQuerySplitter(query)
+    expect(result).toEqual(expected)
+  }
+})