Cheerio Scraper

1import { ApifyClient } from 'apify-client';
2
3// Initialize the ApifyClient with API token
4const client = new ApifyClient({
5    token: '<YOUR_API_TOKEN>',
6});
7
8// Prepare Actor input
9const input = {
10    "startUrls": [
11        {
12            "url": "https://apify.com"
13        }
14    ],
15    "pseudoUrls": [
16        {
17            "purl": "https://apify.com[(/[\\w-]+)?]"
18        }
19    ],
20    "linkSelector": "a",
21    "pageFunction": async function pageFunction(context) {
22        const { $, request, log } = context;
23        const title = $('title').text();
24        log.info(`URL: ${request.url} TITLE: ${title}`);
25        return {
26            url: request.url,
27            title
28        };
29    },
30    "proxyConfiguration": {
31        "useApifyProxy": false
32    },
33    "initialCookies": []
34};
35
36(async () => {
37    // Run the Actor and wait for it to finish
38    const run = await client.actor("apify/cheerio-scraper").call(input);
39
40    // Fetch and print Actor results from the run's dataset (if any)
41    console.log('Results from dataset');
42    const { items } = await client.dataset(run.defaultDatasetId).listItems();
43    items.forEach((item) => {
44        console.dir(item);
45    });
46})();