Perf: 1-pass domain parse

This commit is contained in:
SukkaW
2024-06-07 10:18:31 +08:00
parent 2eb48289b7
commit fba420ee4c
2 changed files with 10 additions and 13 deletions

View File

@@ -1,7 +1,7 @@
// tldts-experimental is way faster than tldts, but very little bit inaccurate // tldts-experimental is way faster than tldts, but very little bit inaccurate
// (since it is hashes based). But the result is still deterministic, which is // (since it is hashes based). But the result is still deterministic, which is
// enough when sorting. // enough when sorting.
import { getDomain, getSubdomain } from 'tldts-experimental'; import * as tldts from 'tldts-experimental';
import { sort } from './timsort'; import { sort } from './timsort';
import { looseTldtsOpt } from '../constants/loose-tldts-opt'; import { looseTldtsOpt } from '../constants/loose-tldts-opt';
@@ -17,11 +17,10 @@ export const buildParseDomainMap = (inputs: string[]) => {
for (let i = 0, len = inputs.length; i < len; i++) { for (let i = 0, len = inputs.length; i < len; i++) {
const cur = inputs[i]; const cur = inputs[i];
if (!domainMap.has(cur)) { if (!domainMap.has(cur)) {
const topD = getDomain(cur, looseTldtsOpt); const parsed = tldts.parse(cur, looseTldtsOpt);
domainMap.set(cur, topD ?? cur); domainMap.set(cur, parsed.domain ?? cur);
// if (!subdomainMap.has(cur)) { // if (!subdomainMap.has(cur)) {
const subD = getSubdomain(cur, looseTldtsOpt); subdomainMap.set(cur, parsed.subdomain ?? cur);
subdomainMap.set(cur, subD ?? cur);
} }
} }
@@ -45,13 +44,11 @@ export const sortDomains = (
const main_domain_a = domainMap.get(a)!; const main_domain_a = domainMap.get(a)!;
const main_domain_b = domainMap.get(b)!; const main_domain_b = domainMap.get(b)!;
let t = compare( let t = compare(main_domain_a, main_domain_b)
main_domain_a, || compare(
main_domain_b /** subdomain_a */ subdomainMap.get(a)!,
) || compare( /** subdomain_b */ subdomainMap.get(b)!
/** subdomain_a */ subdomainMap.get(a)!, );
/** subdomain_b */ subdomainMap.get(b)!
);
if (t !== 0) return t; if (t !== 0) return t;
if (a !== main_domain_a || b !== main_domain_b) { if (a !== main_domain_a || b !== main_domain_b) {

View File

@@ -19,7 +19,7 @@ import { getGorhillPublicSuffixPromise } from './get-gorhill-publicsuffix';
mixedInputs: false mixedInputs: false
}; };
(['getDomain', 'getPublicSuffix', 'getSubdomain'] as const).forEach(methodName => { (['getDomain', 'getPublicSuffix', 'getSubdomain', 'parse'] as const).forEach(methodName => {
group(methodName, () => { group(methodName, () => {
if (methodName in gorhill) { if (methodName in gorhill) {
bench('gorhill', () => { bench('gorhill', () => {