AI/firecrawl
1
0
Fork 0
mirror of https://git.mirrors.martin98.com/https://github.com/mendableai/firecrawl synced 2025-08-18 02:35:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update canonical-url.test.ts

Browse Source
This commit is contained in:
Nicolas 2025-01-03 23:55:05 -03:00
parent f2e0bfbfe3
commit d48ddb8820
1 changed files with 53 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

54
apps/api/src/lib/canonical-url.test.ts
View File

@ -1,4 +1,44 @@
import { normalizeUrl } from './canonical-url';
import { normalizeUrl, normalizeUrlOnlyHostname } from './canonical-url';
describe('normalizeUrlOnlyHostname', () => {
it('should remove protocol and www from URL', () => {
const url = 'https://www.example.com';
const expected = 'example.com';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
it('should remove only protocol if www is not present', () => {
const url = 'https://example.com';
const expected = 'example.com';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
it('should handle URLs without protocol', () => {
const url = 'www.example.com';
const expected = 'example.com';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
it('should handle URLs without protocol and www', () => {
const url = 'example.com';
const expected = 'example.com';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
it('should handle URLs with paths', () => {
const url = 'https://www.example.com/path/to/resource';
const expected = 'example.com';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
it('should handle invalid URLs gracefully', () => {
const url = 'not a valid url';
const expected = 'not a valid url';
expect(normalizeUrlOnlyHostname(url)).toBe(expected);
});
});
describe('normalizeUrl', () => {
it('should remove protocol and www from URL', () => {
@ -27,10 +67,22 @@ describe('normalizeUrl', () => {
it('should handle URLs with paths', () => {
const url = 'https://www.example.com/path/to/resource';
const expected = 'example.com/path/to/resource';
expect(normalizeUrl(url)).toBe(expected);
});
it('should handle URLs with trailing slash', () => {
const url = 'https://www.example.com/';
const expected = 'example.com';
expect(normalizeUrl(url)).toBe(expected);
});
it('should handle URLs with trailing slash and path', () => {
const url = 'https://www.example.com/path/';
const expected = 'example.com/path';
expect(normalizeUrl(url)).toBe(expected);
});
it('should handle invalid URLs gracefully', () => {
const url = 'not a valid url';
const expected = 'not a valid url';