perf: reduce regex usage in getDomainFromOriginOrURL (#241)

patrickhulce · Feb 6, 2025 · 021299a · 021299a
1 parent 26bbd54
commit 021299a
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 2 deletions.
diff --git a/lib/create-entity-finder-api.js b/lib/create-entity-finder-api.js
@@ -6,8 +6,10 @@ const ROOT_DOMAIN_REGEX = /[^.]+\.([^.]+|(gov|com|co|ne)\.\w{2})$/i
 function getDomainFromOriginOrURL(originOrURL) {
   if (typeof originOrURL !== 'string') return null
   if (originOrURL.length > 10000 || originOrURL.startsWith('data:')) return null
-  if (DOMAIN_IN_URL_REGEX.test(originOrURL)) return originOrURL.match(DOMAIN_IN_URL_REGEX)[1]
-  if (DOMAIN_CHARACTERS.test(originOrURL)) return originOrURL.match(DOMAIN_CHARACTERS)[0]
+  let m = originOrURL.match(DOMAIN_IN_URL_REGEX)
+  if (m) return m[1]
+  m = originOrURL.match(DOMAIN_CHARACTERS)
+  if (m) return m[0]
   return null
 }
 

diff --git a/lib/create-entity-finder-api.test.js b/lib/create-entity-finder-api.test.js
@@ -1,3 +1,5 @@
+const fs = require('fs')
+const path = require('path')
 const {createAPIFromDataset} = require('./create-entity-finder-api.js')
 
 describe('getEntity', () => {
@@ -41,4 +43,16 @@ describe('getEntity', () => {
     expect(api.getEntity('https://bar.example.co.uk/path').name).toEqual('Domain')
     expect(api.getEntity('https://baz.bar.example.co.uk/path').name).toEqual('Domain')
   })
+
+  it.skip('stress test', () => {
+    const urls = fs
+      .readFileSync(path.join(__dirname, '../data/random-urls.txt'), 'utf8')
+      .split('\n')
+      .filter(Boolean)
+    console.time('getEntity')
+    for (let i = 0; i < 1_000_000; i++) {
+      api.getEntity(urls[i % urls.length])
+    }
+    console.timeEnd('getEntity')
+  })
 })