remove .includes and add a set to make it O(n) (#2875)

Author: Rahban1

CHANGELOG.md

@@ -11,6 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Changed
 
+* reduced time complexity from O(n^2) to O(n) to improve the initial load time for search ([#2875])
 * Git no longer displays a message about the default branch name when calling `deploydocs()`. ([#2854])
 * Don't escape characters (such as ~) in URLs when writing LaTeX. ([#2210])
 

assets/html/js/search.js

@@ -206,6 +206,7 @@ function worker_function(documenterSearchIndex, documenterBaseURL, filters) {
     // find anything if searching for "add!", only for the entire qualification
     tokenize: (string) => {
       const tokens = [];
+      const tokenSet = new Set();
       let remaining = string;
 
       // julia specific patterns
@@ -232,8 +233,9 @@ function worker_function(documenterSearchIndex, documenterBaseURL, filters) {
         let match;
         while ((match = pattern.exec(remaining)) != null) {
           const token = match[0].trim();
-          if (token && !tokens.includes(token)) {
+          if (token && !tokenSet.has(token)) {
             tokens.push(token);
+            tokenSet.add(token);
           }
         }
       }
@@ -243,8 +245,9 @@ function worker_function(documenterSearchIndex, documenterBaseURL, filters) {
         .split(/[\s\-,;()[\]{}]+/)
         .filter((t) => t.trim());
       for (const token of basicTokens) {
-        if (token && !tokens.includes(token)) {
+        if (token && !tokenSet.has(token)) {
           tokens.push(token);
+          tokenSet.add(token);
         }
       }
 

test/search/wrapper.js

@@ -33,6 +33,7 @@ const index = new MiniSearch({
     },
     tokenize: (string) => {
         const tokens = [];
+        const tokenSet = new Set();
         let remaining = string;
 
         // julia specific patterns
@@ -59,8 +60,9 @@ const index = new MiniSearch({
           let match;
           while ((match = pattern.exec(remaining)) != null) {
             const token = match[0].trim();
-            if (token && !tokens.includes(token)) {
+            if (token && !tokenSet.has(token)) {
               tokens.push(token);
+              tokenSet.add(token);
             }
           }
         }
@@ -70,8 +72,9 @@ const index = new MiniSearch({
           .split(/[\s\-,;()[\]{}]+/)
           .filter((t) => t.trim());
         for (const token of basicTokens) {
-          if (token && !tokens.includes(token)) {
+          if (token && !tokenSet.has(token)) {
             tokens.push(token);
+            tokenSet.add(token);
           }
         }
 
@@ -89,4 +92,4 @@ const results = index.search(__QUERY__, {
 
 // Extract unique page names from results (same logic as search.js)
 const pages = [...new Set(results.map(r => r.title))];
-console.log(JSON.stringify(pages.slice(0,5))); 
+console.log(JSON.stringify(pages.slice(0,5)));