zerochan.js

Imported from GitHub Gist: zerochan.js

#github-gist#javascript#scraper#gist:45e5174c082516452b529d6191627e85

6 / kyiov

Raw

zerochan.jsjavascript

79 lines

1	`import axios from 'axios'`
2	`import * as cheerio from 'cheerio'`
3
4	`const BASE_URL = 'https://www.zerochan.net'`
5	`const UA = 'MithraEX - kyiov'`
6
7	`function absolutize(url) {`
8	`if (!url) return null`
9	`if (/^https?:\/\//i.test(url)) return url`
10	return `${BASE_URL}${url.startsWith('/') ? '' : '/'}${url}`
11	`}`
12
13	`export async function zerochanSearch(query, page = 1) {`
14	`if (!query) throw new Error('Query is required.')`
15
16	const url = `${BASE_URL}/search?q=${encodeURIComponent(query)}&p=${page}`
17	`const { data } = await axios.get(url, {`
18	`headers: {`
19	`'User-Agent': UA,`
20	`'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,webp,/;q=0.8'`
21	`},`
22	`validateStatus: (status) => status === 200 \|\| status === 404,`
23	`timeout: 20000`
24	`})`
25
26	`const $ = cheerio.load(data)`
27	`const results = []`
28
29	`$('#thumbs2 li, #thumbs li').each((_, el) => {`
30	`const a = $(el).find('a.thumb').first()`
31	`const img = $(el).find('img').first()`
32	`const title = img.attr('alt') \|\| a.attr('title') \|\| $(el).find('p a').text().trim()`
33	`const href = a.attr('href')`
34	`const cover = absolutize(img.attr('data-src') \|\| img.attr('src'))`
35	`const id = $(el).attr('data-id') \|\| (href ? href.replace(/[^\d]/g, '') : null)`
36
37	`if (title && href) {`
38	`results.push({`
39	`id,`
40	`title,`
41	`url: absolutize(href),`
42	`cover,`
43	`meta: img.attr('title') \|\| null`
44	`})`
45	`}`
46	`})`
47
48	`return results`
49	`}`
50
51	`export async function zerochanDetail(urlOrId) {`
52	`if (!urlOrId) throw new Error('URL or ID is required.')`
53
54	const url = /^https?:\/\//i.test(urlOrId) ? urlOrId : `${BASE_URL}/${urlOrId}`
55	`const { data } = await axios.get(url, {`
56	`headers: {`
57	`'User-Agent': UA,`
58	`'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,webp,/;q=0.8'`
59	`},`
60	`validateStatus: (status) => status === 200 \|\| status === 404,`
61	`timeout: 20000`
62	`})`
63
64	`const $ = cheerio.load(data)`
65	`const title = $('h1').first().text().trim() \|\| $('.content h1').first().text().trim() \|\| 'Zerochan Image'`
66	`const cover = absolutize($('img').first().attr('data-src') \|\| $('img').first().attr('src'))`
67	`const metaText = $('img').first().attr('title') \|\| $('.meta').text().trim() \|\| ''`
68
69	`return {`
70	`title,`
71	`cover,`
72	`source: url,`
73	`metaText`
74	`}`
75	`}`
76
77	`export default { zerochanSearch, zerochanDetail }`
78
79

Discussion & Comments0

No comments yet — be the first to leave feedback or start a code discussion.