extern/se-scraper
1
0
Fork 1
mirror of https://github.com/NikolaiT/se-scraper.git synced 2025-06-28 21:41:27 +02:00
Code Issues Packages Projects Releases Wiki Activity

fix(google): parser for one shopping product right panel

Browse Source
This commit is contained in:
HugoPoi 2020-06-17 16:24:11 +02:00
parent b9a9227f14
commit ad8903b001
3 changed files with 274 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
src/modules/google.js
View File

@ -29,7 +29,7 @@ class GoogleScraper extends Scraper {
if (n) { if (n) {
return n.getAttribute(attr); return n.getAttribute(attr);
} else { } else {
return null; return;
} }
}; };
@ -111,14 +111,14 @@ class GoogleScraper extends Scraper {
// parse right side product information // parse right side product information
results.right_info.review = _attr(document, '#rhs .cu-container g-review-stars span', 'aria-label'); results.right_info.review = _attr(document, '#rhs .cu-container g-review-stars span', 'aria-label');
let title_el = document.querySelector('#rhs .cu-container g-review-stars'); let title_el = document.querySelector('#rhs .cu-container .Q7Oxbd');
if (title_el) { if (title_el) {
results.right_info.review.title = title_el.parentNode.querySelector('div:first-child').innerText; results.right_info.title = title_el.innerText;
} }
let num_reviews_el = document.querySelector('#rhs .cu-container g-review-stars'); let num_reviews_el = document.querySelector('#rhs .cu-container .PGDKUd');
if (num_reviews_el) { if (num_reviews_el) {
results.right_info.num_reviews = num_reviews_el.parentNode.querySelector('div:nth-of-type(2)').innerText; results.right_info.num_reviews = num_reviews_el.innerText;
} }
results.right_info.vendors = []; results.right_info.vendors = [];
@ -127,20 +127,16 @@ class GoogleScraper extends Scraper {
document.querySelectorAll('#rhs .cu-container .rhsvw > div > div:nth-child(4) > div > div:nth-child(3) > div').forEach((el) => { document.querySelectorAll('#rhs .cu-container .rhsvw > div > div:nth-child(4) > div > div:nth-child(3) > div').forEach((el) => {
results.right_info.vendors.push({ results.right_info.vendors.push({
price: _text(el, 'span:nth-of-type(1)'), price: _text(el, 'span:nth-of-type(1)'),
merchant_name: _text(el, 'span:nth-child(3) a:nth-child(2)'), merchant_name: _text(el, '.doUe3s0oL2B__jackpot-merchant a'),
merchant_ad_link: _attr(el, 'span:nth-child(3) a:first-child', 'href'), merchant_ad_link: _attr(el, 'span:nth-child(3) a:first-child', 'href'),
merchant_link: _attr(el, 'span:nth-child(3) a:nth-child(2)', 'href'), merchant_link: _attr(el, 'span:nth-child(3) a:nth-child(2)', 'href'), // TODO this is not working anymore
source_name: _text(el, 'span:nth-child(4) a'), source_name: _text(el, 'span:nth-child(4) a'),
source_link: _attr(el, 'span:nth-child(4) a', 'href'), source_link: _attr(el, 'span:nth-child(4) a', 'href'),
info: _text(el, 'div span'), info: _text(el, '.SdBHnc.e2CF7c'),
shipping: _text(el, 'span:last-child > span'), shipping: _text(el, '.JfwJme'),
}) })
}); });
if (!results.right_info.title) {
results.right_info = {};
}
let right_side_info_el = document.getElementById('rhs'); let right_side_info_el = document.getElementById('rhs');
if (right_side_info_el) { if (right_side_info_el) {

220
test/mocks/google/shopping right product review_page1.html Normal file
View File

File diff suppressed because one or more lines are too long

46
test/modules/google.js
View File

@ -143,4 +143,48 @@ describe('Module Google', function(){
}); });
}); });
}); it('shopping extract right one product', function () {
const googleScraper = new GoogleScraper({
config: {
search_engine_name: 'google',
throw_on_detection: true,
keywords: ['shopping right product review'],
logger: testLogger,
scrape_from_file: '',
num_pages: 1,
}
});
googleScraper.STANDARD_TIMEOUT = 500;
return googleScraper.run({page}).then(({results, metadata, num_requests}) => {
assert.strictEqual(num_requests, 1, 'One request should be done');
assert.strictEqual(results['shopping right product review']['1'].results.length, 9, 'Must have 9 organic results parsed on page 1');
assert.deepEqual(results['shopping right product review']['1'].right_info, {
title: 'Lacoste Lunettes',
'info': '',
'num_reviews': '146 avis',
'review': 'Note : 4,6 sur 5',
'vendors': [
{
'info': '317 · 2807',
'merchant_ad_link': 'https://www.googleadservices.com/pagead/aclk?sa=L&ai=DChcSEwihq9C82ojqAhUIyrIKHbIHAx8YABACGgJscg&ohost=www.google.com&cid=CAASE-Roz5UHMJg95vk99OwXQnKbUG0&sig=AOD64_0Wfsw3t3eO_yEtq8lWRIjiF6EqZw&ctype=5&q=&ved=2ahUKEwjsqsi82ojqAhVFPBoKHY38DAIQ9A56BAgNEH0&adurl=',
'merchant_name': 'Edel-Optics FR',
'price': '102,75 €',
'shipping': 'Livraison gratuite',
'source_link': 'https://www.google.com/search?tbm=shop&q=lacoste%20317',
'source_name': 'Par Google',
},
{
'info': '317 · 2805',
'merchant_ad_link': 'https://www.googleadservices.com/pagead/aclk?sa=L&ai=DChcSEwihq9C82ojqAhUIyrIKHbIHAx8YABADGgJscg&ohost=www.google.com&cid=CAASE-Roz5UHMJg95vk99OwXQnKbUG0&sig=AOD64_2R4Idoiqc783K8OLyv9W9YQTJfog&ctype=5&q=&ved=2ahUKEwjsqsi82ojqAhVFPBoKHY38DAIQ9A56BQgNEIEB&adurl=',
'merchant_name': 'EasyLunettes.fr',
'price': '75,00 €',
'shipping': 'Livraison gratuite',
'source_link': 'https://producthero.com/?utm_source=google&utm_medium=css&q=lacoste%20317',
'source_name': 'Par Producthero',
}
]
});
});
});
});