small fixes

This commit is contained in:
Nick Leeman 2023-07-15 20:39:09 +02:00
parent b67999e3bd
commit f5d85297b9
2 changed files with 47 additions and 8 deletions

View File

@ -5,7 +5,7 @@ import fs from "fs";
import { getDomain } from "tldts"; import { getDomain } from "tldts";
import { Browser, launch } from "puppeteer"; import { Browser, launch } from "puppeteer";
import axios from "axios"; import axios from "axios";
import { Dobey, EP, LDJsonParser, Macrovet } from "./modules/websites"; import { CoolBlue, Dobey, EP, LDJsonParser, Macrovet } from "./modules/websites";
import { createObjectCsvWriter } from "csv-writer"; import { createObjectCsvWriter } from "csv-writer";
import { CsvWriter } from "csv-writer/src/lib/csv-writer"; import { CsvWriter } from "csv-writer/src/lib/csv-writer";
import { ObjectMap } from "csv-writer/src/lib/lang/object"; import { ObjectMap } from "csv-writer/src/lib/lang/object";
@ -28,7 +28,9 @@ start();
async function debugStart() { async function debugStart() {
browser = await launch({ headless: false }); browser = await launch({ headless: false });
console.log(await crawlProductStock("https://www.hondenbed.nl/hondenkussen-taupe-bruin.html?utm_medium=affiliate&utm_source=tradetracker")); 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.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"));
} }
async function start() { async function start() {
@ -134,7 +136,7 @@ async function crawlUrl(url: string) {
await page.goto(url, { waitUntil: 'networkidle2' }); await page.goto(url, { waitUntil: 'networkidle2' });
// Get html from page // Get html from page
const html = await page.evaluate(() => document.body.innerHTML); const html = await page.evaluate(() => document.documentElement.outerHTML);
// Close page // Close page
await page.close(); await page.close();
@ -202,6 +204,8 @@ async function crawlUrl(url: string) {
in_stock: products[index].in_stock, in_stock: products[index].in_stock,
blog_url: products[index].blog_url, blog_url: products[index].blog_url,
}]); }]);
await wait(1000);
} catch (error) { } catch (error) {
console.log(` [ERROR] ${products[index].product_name} - ${products[index].domain} - ${products[index].product_url}`, error); console.log(` [ERROR] ${products[index].product_name} - ${products[index].domain} - ${products[index].product_url}`, error);
continue; continue;
@ -219,13 +223,11 @@ async function crawlProductStock(url: string): Promise<[string, string, boolean]
try { try {
// Open new page and goto url // Open new page and goto url
const page = await browser.newPage(); const page = await browser.newPage();
page.setUserAgent("Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/72.0.3626.121 Safari/537.36");
await page.goto(url, { waitUntil: 'networkidle2' }); await page.goto(url, { waitUntil: 'networkidle2' });
// Get html from page // Get html from page
const html = await page.evaluate(() => document.body.innerHTML); const html = await page.evaluate(() => document.documentElement.outerHTML);
// Load html in cheerio object
const $ = cheerio.load(html);
// Get domain from url // Get domain from url
const domain = getDomain(page.url()); const domain = getDomain(page.url());
@ -253,7 +255,7 @@ async function crawlProductStock(url: string): Promise<[string, string, boolean]
return [domain, page.url(), await LDJsonParser.check(html)]; return [domain, page.url(), await LDJsonParser.check(html)];
case "coolblue.nl": case "coolblue.nl":
return [domain, page.url(), await LDJsonParser.check(html)]; return [domain, page.url(), await CoolBlue.check(html)];
case "bcc.nl": case "bcc.nl":
return [domain, page.url(), await LDJsonParser.check(html)]; return [domain, page.url(), await LDJsonParser.check(html)];
@ -284,4 +286,8 @@ async function crawlProductStock(url: string): Promise<[string, string, boolean]
console.error(error); console.error(error);
console.log("-- Error while trying to crawl page! Skipping..."); console.log("-- Error while trying to crawl page! Skipping...");
} }
}
function wait(ms: number) {
return new Promise(resolve => setTimeout(resolve, ms));
} }

View File

@ -21,6 +21,7 @@ export namespace LDJsonParser {
const $ = cheerio.load(html); const $ = cheerio.load(html);
$('script[type="application/ld+json"]').each((index, element) => { $('script[type="application/ld+json"]').each((index, element) => {
snippets.push($(element).html()); snippets.push($(element).html());
}); });
for (const snippet of snippets) { for (const snippet of snippets) {
@ -43,6 +44,38 @@ export namespace LDJsonParser {
} }
} }
} catch (error) { } catch (error) {
console.log(error);
continue;
}
}
return false;
} catch (error) {
console.log(error);
console.error(`Error occured during stock check!`);
return false;
}
}
}
export namespace CoolBlue {
export async function check(html: string) {
try {
const snippets: Array<string> = [];
const $ = cheerio.load(html);
$('script[type="application/ld+json"]').each((index, element) => {
snippets.push($(element).html());
});
for (const snippet of snippets) {
try {
if (snippet.includes(`"@type": "Product"`)) {
if (snippet.includes(`"availability": "https://schema.org/InStock"`)) {
return true;
}
}
} catch (error) {
console.log(error);
continue; continue;
} }
} }