@easyops-cn/docusaurus-search-local 0.44.6 → 0.45.0

package/CHANGELOG.md

@@ -2,6 +2,19 @@
 
 All notable changes to this project will be documented in this file. See [standard-version](https://github.com/conventional-changelog/standard-version) for commit guidelines.
 
+## [0.45.0](https://github.com/easyops-cn/docusaurus-search-local/compare/v0.44.6...v0.45.0) (2024-10-09)
+
+
+### Features
+
+* fix a perf issue when searching with a large number of Chinese words ([abe720c](https://github.com/easyops-cn/docusaurus-search-local/commit/abe720cd3d521122e3aaa8b2d8ba7d0967fe046d)), closes [#312](https://github.com/easyops-cn/docusaurus-search-local/issues/312)
+
+
+### Bug Fixes
+
+* fix a perf issue that cut zh words unnecessary repeatedly ([870dc88](https://github.com/easyops-cn/docusaurus-search-local/commit/870dc887f85ded4cbf000c2480e9e92c6fa0cad6))
+* limit the number of terms, closes [#312](https://github.com/easyops-cn/docusaurus-search-local/issues/312) ([c889047](https://github.com/easyops-cn/docusaurus-search-local/commit/c8890479914e89f7dea075f5f191bbc76e99d249))
+
 ## [0.44.6](https://github.com/easyops-cn/docusaurus-search-local/compare/v0.44.5...v0.44.6) (2024-10-08)
 
 

package/dist/client/client/utils/smartTerms.js

@@ -1,4 +1,6 @@
 import { cutZhWords } from "./cutZhWords";
+const MAX_TERMS = 12;
+const HALF_MAX_TERMS = MAX_TERMS / 2;
 /**
  * Get all possible terms for a list of tokens consists of words mixed in Chinese and non-Chinese,
  * by a Chinese words dictionary.
@@ -9,27 +11,57 @@ import { cutZhWords } from "./cutZhWords";
  * @returns A smart term list.
  */
 export function smartTerms(tokens, zhDictionary) {
-    const terms = [];
-    function cutMixedWords(subTokens, carry) {
-        if (subTokens.length === 0) {
-            terms.push(carry);
-            return;
-        }
-        const token = subTokens[0];
+    const tokenTerms = tokens
+        .map((token) => {
         if (/\p{Unified_Ideograph}/u.test(token)) {
-            const terms = cutZhWords(token, zhDictionary);
-            for (const term of terms) {
-                const nextCarry = carry.concat(...term);
-                cutMixedWords(subTokens.slice(1), nextCarry);
-            }
+            return cutZhWords(token, zhDictionary);
         }
         else {
-            const nextCarry = carry.concat({
-                value: token,
-            });
-            cutMixedWords(subTokens.slice(1), nextCarry);
+            return [{ value: token }];
+        }
+    })
+        .slice(0, MAX_TERMS);
+    const tokenTermsThatAreMultiple = tokenTerms.filter((tokenTerm) => tokenTerm.length > 1);
+    let termsProduct = 1;
+    let overflowed = false;
+    for (const tokenTerm of tokenTermsThatAreMultiple) {
+        if (overflowed) {
+            tokenTerm.splice(1, tokenTerm.length - 1);
+        }
+        else {
+            if (tokenTerm.length > HALF_MAX_TERMS) {
+                tokenTerm.splice(HALF_MAX_TERMS, tokenTerm.length - HALF_MAX_TERMS);
+            }
+            const product = termsProduct * tokenTerm.length;
+            if (product >= MAX_TERMS) {
+                if (product > MAX_TERMS) {
+                    const max = Math.floor(MAX_TERMS / termsProduct);
+                    tokenTerm.splice(max, tokenTerm.length - max);
+                    termsProduct = max * termsProduct;
+                }
+                else {
+                    termsProduct = product;
+                }
+                if (termsProduct > HALF_MAX_TERMS) {
+                    overflowed = true;
+                }
+            }
+            else {
+                termsProduct = product;
+            }
+        }
+    }
+    // Get all possible combinations of terms.
+    const terms = [];
+    function combine(index, carry) {
+        if (index === tokenTerms.length || carry.length >= MAX_TERMS) {
+            terms.push(carry.slice(0, MAX_TERMS));
+            return;
+        }
+        for (const term of tokenTerms[index]) {
+            combine(index + 1, carry.concat(term));
         }
     }
-    cutMixedWords(tokens, []);
+    combine(0, []);
     return terms;
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@easyops-cn/docusaurus-search-local",
-  "version": "0.44.6",
+  "version": "0.45.0",
   "description": "An offline/local search plugin for Docusaurus v3",
   "repository": {
     "type": "git",