Frozenverse/ContentEggStockCrawler
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases 1 Wiki Activity

added mediamarkt support.

Browse Source
This commit is contained in:
Nick Leeman 2023-07-16 10:56:44 +02:00
parent f5d85297b9
commit 421155d3fa
2 changed files with 44 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/cli.ts
View File

@ -5,7 +5,7 @@ import fs from "fs";
import { getDomain } from "tldts";
import { Browser, launch } from "puppeteer";
import axios from "axios";
import { CoolBlue, Dobey, EP, LDJsonParser, Macrovet } from "./modules/websites";
import { CoolBlue, Dobey, EP, LDJsonParser, Macrovet, Mediamarkt } from "./modules/websites";
import { createObjectCsvWriter } from "csv-writer";
import { CsvWriter } from "csv-writer/src/lib/csv-writer";
import { ObjectMap } from "csv-writer/src/lib/lang/object";
@ -23,12 +23,13 @@ let browser: Browser = null;
let sitemapUrl: string = "";
let writer: CsvWriter<ObjectMap<any>> = null;
// debugStart();
start();
debugStart();
// start();
async function debugStart() {
browser = await launch({ headless: false });
console.log(await crawlProductStock("https://prf.hn/click/camref:1100leWsm/destination:https%3A%2F%2Fwww.coolblue.nl%2Fproduct%2F856992%2Fmio-cyclo-discover.html"));
console.log(await crawlProductStock("https://www.mediamarkt.nl/nl/product/_sitecom-ca-040-usb-c-naar-usb-c-gamingkabel-met-pd-1658247.html"));
console.log(await crawlProductStock("https://www.mediamarkt.nl/nl/product/_apple-iphone-14-128gb-midnight-1738479.html"));
// console.log(await crawlProductStock("https://www.coolblue.nl/product/923036/hp-deskjet-2720e-all-in-one.html?clickref=1101lwW9ebAE&utm_source=performancehorizon&utm_medium=affiliate&utm_campaign=Comparison%2FReview&utm_content=1101lwW9ebAE&utm_term=1100leWsm&ref=293530&PHGref=1101lwW9ebAE&cmt=c_ph%2Capm_Comparison%2FReview_%2Cacid1101l93%2Cacr_1100leWsm%2Caclr_1101lwW9ebAE"));
// console.log(await crawlProductStock("https://www.coolblue.nl/product/882996/canon-pixma-ts-3450-zwart.html?clickref=1011lwWDBFEL&utm_source=performancehorizon&utm_medium=affiliate&utm_campaign=Comparison%2FReview&utm_content=1011lwWDBFEL&utm_term=1100leWsm&ref=293530&PHGref=1011lwWDBFEL&cmt=c_ph%2Capm_Comparison%2FReview_%2Cacid1101l93%2Cacr_1100leWsm%2Caclr_1011lwWDBFEL"));
}
@ -278,6 +279,9 @@ async function crawlProductStock(url: string): Promise<[string, string, boolean]
case "ezydog.nl":
return [domain, page.url(), await LDJsonParser.check(html)];
case "mediamarkt.nl":
return [domain, page.url(), await Mediamarkt.check(html)];
default:
console.error(`-- ${domain} is not an supported website! Cannot check stock!`);
return [domain, page.url(), false];

41
src/modules/websites.ts
View File

@ -20,8 +20,14 @@ export namespace LDJsonParser {
const snippets: Array<string> = [];
const $ = cheerio.load(html);
$('script[type="application/ld+json"]').each((index, element) => {
snippets.push($(element).html());
let html = $(element).html();
// remove all spaces
html = html.replace(/\s/g, '');
// remove all newlines
html = html.replace(/\n/g, '');
snippets.push(html);
});
for (const snippet of snippets) {
@ -64,13 +70,20 @@ export namespace CoolBlue {
const snippets: Array<string> = [];
const $ = cheerio.load(html);
$('script[type="application/ld+json"]').each((index, element) => {
snippets.push($(element).html());
let html = $(element).html();
// remove all spaces
html = html.replace(/\s/g, '');
// remove all newlines
html = html.replace(/\n/g, '');
snippets.push(html);
});
for (const snippet of snippets) {
try {
if (snippet.includes(`"@type": "Product"`)) {
if (snippet.includes(`"availability": "https://schema.org/InStock"`)) {
if (snippet.includes(`"@type":"Product"`)) {
if (snippet.includes(`"availability":"http://schema.org/InStock"`) || snippet.includes(`"availability":"https://schema.org/InStock"`)) {
return true;
}
}
@ -89,6 +102,24 @@ export namespace CoolBlue {
}
}
export namespace Mediamarkt {
export async function check(html: string) {
try {
const $ = cheerio.load(html);
// Check if data-test="mms-delivery-online-availability_AVAILABLE" exists
if ($('[data-test="mms-delivery-online-availability_AVAILABLE"]').length > 0) {
return true;
}
return false;
} catch (error) {
console.log(error);
console.error(`Error occured during stock check!`);
return false;
}
}
}
export namespace Dobey {
export async function check(html: string) {