R524809
92 lines
3.3 KiB
TypeScript
92 lines
3.3 KiB
TypeScript
import assert from 'node:assert/strict';
|
|
import { readFile } from 'node:fs/promises';
|
|
import test from 'node:test';
|
|
|
|
import {
|
|
extractOzonCharacteristics,
|
|
extractOzonDescription,
|
|
extractOzonGalleryImageUrls,
|
|
} from '../lib/ozon-parser.ts';
|
|
import { dedupeImageUrl, normaliseImageUrl } from '../lib/product-media.ts';
|
|
|
|
test('Ozon gallery fixture only returns product-gallery images', async () => {
|
|
const fixture = await readFile(
|
|
new URL('../../../packages/reference/ozon1.html', import.meta.url),
|
|
'utf8',
|
|
);
|
|
const state = fixture.match(
|
|
/id="state-webGallery-[^"]+"\s+data-state='([^']+)'/,
|
|
)?.[1];
|
|
assert.ok(state, 'Ozon gallery widget state must exist in the fixture');
|
|
|
|
const urls = extractOzonGalleryImageUrls(
|
|
JSON.parse(state),
|
|
'https://www.ozon.ru/product/ryukzak-5387005639/',
|
|
);
|
|
assert.equal(urls.length, 11);
|
|
assert.ok(urls.every((url) => /\/s3\/multimedia-/.test(url)));
|
|
assert.ok(
|
|
urls.every(
|
|
(url) => !/logo_ozon|payments-cdn|recommend|review/i.test(url),
|
|
),
|
|
);
|
|
});
|
|
|
|
test('Ozon short characteristics fixture returns its product parameters', async () => {
|
|
const fixture = await readFile(
|
|
new URL('../../../packages/reference/ozon1.html', import.meta.url),
|
|
'utf8',
|
|
);
|
|
const state = fixture.match(
|
|
/id="state-webShortCharacteristics-[^"]+"\s+data-state='([^']+)'/,
|
|
)?.[1];
|
|
assert.ok(state, 'Ozon short characteristics widget state must exist');
|
|
|
|
const attributes = extractOzonCharacteristics([
|
|
{
|
|
...JSON.parse(state),
|
|
__atlasWidgetName: 'webShortCharacteristics-fixture',
|
|
},
|
|
]);
|
|
assert.equal(attributes.length, 5);
|
|
assert.deepEqual(
|
|
attributes.slice(0, 3).map(({ name, value }) => ({ name, value })),
|
|
[
|
|
{ name: '类型', value: '背包' },
|
|
{ name: '原产国', value: '中国' },
|
|
{ name: '材料', value: 'Cordura 1000D' },
|
|
],
|
|
);
|
|
});
|
|
|
|
test('Ozon rich annotation converts to description text without images', () => {
|
|
const description = extractOzonDescription([
|
|
{
|
|
__atlasWidgetName: 'webDescription-fixture',
|
|
richAnnotationJson: JSON.stringify([
|
|
{ type: 'text', content: '耐用的背包' },
|
|
{ type: 'image', src: 'https://ir.ozone.ru/s3/example.jpg' },
|
|
{ type: 'text', text: '适合旅行' },
|
|
]),
|
|
},
|
|
]);
|
|
assert.equal(description?.value, '耐用的背包\n适合旅行');
|
|
});
|
|
|
|
test('Ozon thumbnail variants normalise to one original URL', () => {
|
|
const original = 'https://ir.ozone.ru/s3/multimedia-1-9/13597160877.jpg';
|
|
const thumbnail =
|
|
'https://ir.ozone.ru/s3/multimedia-1-9/wc100/13597160877.jpg?width=100&quality=80';
|
|
assert.equal(normaliseImageUrl(thumbnail, original, 'ozon'), original);
|
|
assert.equal(
|
|
dedupeImageUrl(thumbnail, 'ozon'),
|
|
dedupeImageUrl(original, 'ozon'),
|
|
);
|
|
});
|
|
|
|
test('Alibaba transformed image URLs normalise before de-duplication', () => {
|
|
const original = 'https://cbu01.alicdn.com/img/ibank/O1CN01.jpg';
|
|
const transformed = `${original}_400x400.jpg?width=400`;
|
|
assert.equal(normaliseImageUrl(transformed, original, '1688'), original);
|
|
});
|