1import { ApifyClient } from 'apify-client';
2
3
4
5const client = new ApifyClient({
6 token: '<YOUR_API_TOKEN>',
7});
8
9
10const input = {
11 "start_urls": [
12 {
13 "url": "https://docs.python.org/3/"
14 }
15 ],
16 "documents": [
17 {
18 "text": "Your document text here...",
19 "source_id": "doc_001"
20 }
21 ],
22 "content_selectors": [
23 "article",
24 "main",
25 ".content"
26 ],
27 "exclude_selectors": [
28 "nav",
29 "footer",
30 "aside",
31 ".sidebar"
32 ],
33 "language_filter": [
34 "en"
35 ]
36};
37
38
39const run = await client.actor("mea/ai-training-data-curator").call(input);
40
41
42console.log('Results from dataset');
43console.log(`๐พ Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`);
44const { items } = await client.dataset(run.defaultDatasetId).listItems();
45items.forEach((item) => {
46 console.dir(item);
47});
48
49