Provider for immobilien.de (#65)

* Add provider for immobilien.de
This commit is contained in:
Quoc Duong Bui
2022-12-19 19:29:13 +01:00
committed by GitHub
parent c5f4333878
commit 47e4230b39
3 changed files with 115 additions and 1 deletions

View File

@@ -0,0 +1,59 @@
const utils = require('../utils');
let appliedBlackList = [];
function shortenLink(link) {
return link.substring(0, link.indexOf('?'));
}
function parseId(shortenedLink) {
return shortenedLink.substring(shortenedLink.lastIndexOf('/') + 1);
}
function normalize(o) {
const id = parseId(shortenLink(o.link));
const size = o.size || 'N/A m²';
const price = o.price || 'N/A €';
const title = o.title || 'No title available';
const address = o.address || 'No address available';
const link = shortenLink(o.link);
return Object.assign(o, { id, price, size, title, address, link });
}
function applyBlacklist(o) {
const titleNotBlacklisted = !utils.isOneOf(o.title, appliedBlackList);
const descNotBlacklisted = !utils.isOneOf(o.description, appliedBlackList);
return titleNotBlacklisted && descNotBlacklisted;
}
const config = {
url: null,
crawlContainer: '.estates_list .list_immo a._ref',
sortByDateParam: 'sort_col=*created_ts&sort_dir=desc',
crawlFields: {
price: '.list_entry .immo_preis .label_info',
size: '.list_entry .flaeche .label_info | removeNewline | trim',
title: '.list_entry .part_text h3 span',
description: '.list_entry .description | trim',
link: '@href',
address: '.list_entry .place',
},
paginate: '.list_immo .blocknav .blocknav_list li.next a@href',
normalize: normalize,
filter: applyBlacklist,
};
exports.init = (sourceConfig, blacklist) => {
config.enabled = sourceConfig.enabled;
config.url = sourceConfig.url;
appliedBlackList = blacklist || [];
};
exports.metaInformation = {
name: 'Immobilien.de',
baseUrl: 'https://www.immobilien.de/',
id: __filename.slice(__dirname.length + 1, -3),
};
exports.config = config;

View File

@@ -0,0 +1,51 @@
const similarityCache = require('../../lib/services/similarity-check/similarityCache');
const mockNotification = require('../mocks/mockNotification');
const providerConfig = require('./testProvider.json');
const mockStore = require('../mocks/mockStore');
const proxyquire = require('proxyquire').noCallThru();
const expect = require('chai').expect;
const provider = require('../../lib/provider/immobilienDe');
describe('#immobilien.de testsuite()', () => {
after(() => {
similarityCache.stopCacheCleanup();
});
provider.init(providerConfig.immobilienDe, [], []);
const Fredy = proxyquire('../../lib/FredyRuntime', {
'./services/storage/listingsStorage': {
...mockStore,
},
'./notification/notify': mockNotification,
});
it('should test immobilien.de provider', async () => {
return await new Promise((resolve) => {
const fredy = new Fredy(provider.config, null, provider.metaInformation.id, 'test1', similarityCache);
fredy.execute().then((listing) => {
expect(listing).to.be.a('array');
const notificationObj = mockNotification.get();
expect(notificationObj).to.be.a('object');
expect(notificationObj.serviceName).to.equal('immobilienDe');
notificationObj.payload.forEach((notify) => {
/** check the actual structure **/
expect(notify.id).to.be.a('string');
expect(notify.price).to.be.a('string');
expect(notify.size).to.be.a('string');
expect(notify.title).to.be.a('string');
expect(notify.link).to.be.a('string');
expect(notify.address).to.be.a('string');
/** check the values if possible **/
expect(notify.price).that.does.include('€');
expect(notify.size).that.does.include('m²');
expect(notify.title).to.be.not.empty;
expect(notify.link).that.does.include('https://www.immobilien.de');
expect(notify.address).to.be.not.empty;
});
resolve();
});
});
});
});

View File

@@ -4,6 +4,10 @@
"enabled": true,
"id": "einsAImmobilien"
},
"immobilienDe": {
"url": "https://www.immobilien.de/Wohnen/Suchergebnisse-51797.html?search._digest=true&search._filter=wohnen&search.flaeche_von=50&search.objektart=wohnung&search.preis_bis=1200&search.typ=mieten&search.umkreis=15&search.wo=district%3A2434%2C2695%2C2621%2C2700%2C2967%2C2734%2C2909%2C2955%2C2392%2C2746%2C2767%2C2982%2C2904%2C2612%2C2892%2C2587%2C2871%2C2975%2C2591%2C2887%2C2569%2C2640%2C2735&sort_col=*created_ts&sort_dir=desc",
"enabled": true
},
"immonet": {
"url": "https://www.immonet.de/immobiliensuche/sel.do?pageoffset=1&listsize=100&objecttype=1&locationname=Düsseldorf&acid=&actype=&district=8717&district=8718&district=8719&district=8720&district=8721&district=8723&district=8724&district=8725&district=8727&district=8728&district=8729&district=8730&district=8731&district=8732&district=8733&district=8737&district=8738&district=8741&district=8745&district=8747&district=8750&district=8752&district=8754&district=8755&district=8756&district=8759&district=8760&district=8761&district=8763&district=8764&district=8765&ajaxIsRadiusActive=false&sortby=19&suchart=1&radius=0&pcatmtypes=1_1&pCatMTypeStoragefield=&parentcat=1&marketingtype=1&fromprice=&toprice=420000&fromarea=90&toarea=&fromplotarea=&toplotarea=&fromrooms=3&torooms=&objectcat=225&objectcat=18&objectcat=17&objectcat=12&objectcat=16&objectcat=181&objectcat=14&objectcat=15&objectcat=226&objectcat=13&wbs=-1&fromyear=&toyear=",
"enabled": true
@@ -36,4 +40,4 @@
"url": "https://www.wg-gesucht.de/wg-zimmer-in-Duesseldorf.30.0.1.0.html?offer_filter=1&noDeact=1&city_id=30&category=0&rent_type=0&rMax=5000",
"enabled": true
}
}
}