bringing back immonet by using scrapingant

This commit is contained in:
weakmap@gmail.com
2023-04-15 18:24:51 +02:00
parent 6fbee3e7c6
commit 4ba098e0b6
10 changed files with 47 additions and 30 deletions

View File

@@ -1,14 +1,12 @@
import utils from '../utils.js';
let appliedBlackList = [];
function normalize(o) {
const id = parseInt(o.id.substring(o.id.indexOf('_') + 1, o.id.length));
const id = o.id.substring(o.id.lastIndexOf('/') + 1, o.id.length);
const size = o.size != null ? o.size.replace('Wohnfläche ', '') : 'N/A m²';
const price = o.price.replace('Kaufpreis ', '');
const address = o.address.split(' • ')[o.address.split(' • ').length - 1];
const title = o.title || 'No title available';
//normally we would just read the link from the source, but immonet decided to trick user by adding a click listener instead of
//a href to do some weird reporting. (Very user friendly for handicaped ppl... not)
const link = `https://www.immonet.de/angebot/${id}`;
const link = o.id;
return Object.assign(o, { id, address, price, size, title, link });
}
function applyBlacklist(o) {
@@ -18,14 +16,14 @@ function applyBlacklist(o) {
}
const config = {
url: null,
crawlContainer: '#result-list-stage .item',
crawlContainer: '.content-wrapper-tiles .ng-star-inserted',
sortByDateParam: 'sortby=19',
crawlFields: {
id: '@id',
price: 'div[id*="selPrice_"] | trim',
size: 'div[id*="selArea_"] | trim',
title: '.item a img@title',
address: '.item .box-25 .ellipsis .text-100 | removeNewline | trim',
id: '.card a@href',
title: '.card h3 |trim',
price: '.card .has-font-300 .is-bold | trim',
size: '.card .has-font-300 .ml-100 | trim',
address: '.card span:nth-child(2) | trim',
},
paginate: '#idResultList .margin-bottom-6.margin-bottom-sm-12 .panel a.pull-right@href',
normalize: normalize,