43 lines
1.1 KiB
JavaScript
43 lines
1.1 KiB
JavaScript
let fetch = require("node-fetch");
|
|
let cheerio = require("cheerio");
|
|
const areThereAnyNewItems = require("./areThereAnyNewItems");
|
|
|
|
async function scrapTheItems(url, controlDate, noNewItems = false) {
|
|
let items = [];
|
|
let response = await fetch(url);
|
|
const body = await response.text();
|
|
const $ = cheerio.load(body);
|
|
$("#rezultatipretrage")
|
|
.find(".listitem")
|
|
.each(async (index, elem) => {
|
|
if (noNewItems) return;
|
|
const itemDate = $(elem)
|
|
.find(".cijena > .datum > div")
|
|
.first()
|
|
.attr("data-cijelidatum");
|
|
|
|
if (controlDate && !areThereAnyNewItems(itemDate, controlDate)) {
|
|
noNewItems = true;
|
|
return;
|
|
}
|
|
|
|
const id = $(elem)
|
|
.find("a")
|
|
.first()
|
|
.attr("href");
|
|
const cijena = $(elem)
|
|
.find(".cijena > .datum > span")
|
|
.first()
|
|
.text();
|
|
const image = $(elem)
|
|
.find("a > .slika > img")
|
|
.first()
|
|
.attr("src");
|
|
|
|
items.push({ url: id, price: cijena, image, date: itemDate });
|
|
});
|
|
return items;
|
|
}
|
|
|
|
module.exports = scrapTheItems;
|