Scraper wallpaper - unsplash.com (anti-bot)

#github-gist#javascript#gist:0efd29a97ad0740db33bda14777a215d#gistsync:2026-09-02T10:46:06Z

1 / kyiov

Raw

unsplash.jsjavascript

121 lines

1
2
3	`import fs from 'node:fs/promises'`
4	`import os from 'node:os'`
5	`import path from 'node:path'`
6	`import { execFile } from 'node:child_process'`
7	`import { promisify } from 'node:util'`
8	`import { fileURLToPath } from 'node:url'`
9
10	`const execFileAsync = promisify(execFile)`
11
12	`const BASE = 'https://unsplash.com'`
13	`const UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/130.0.0.0 Safari/537.36'`
14
15	`const BRIDGE = fileURLToPath(new URL('./scrapling_fetch.py', import.meta.url))`
16
17	`const uniq = (arr) => [...new Set(arr)]`
18
19	`async function fetchPageHtml(url) {`
20	`const outputDir = await fs.mkdtemp(path.join(os.tmpdir(), 'unsplash-scrape-'))`
21	`const opts = JSON.stringify({`
22	`mode: 'auto',`
23	`referer: 'https://www.google.com/',`
24	`timeout: 60000,`
25	`solve_cloudflare: true,`
26	`headless: true,`
27	`})`
28	`try {`
29	`const { stdout } = await execFileAsync(process.env.SCRAPLING_PYTHON \|\| "/tmp/scrapping-venv/bin/python", [BRIDGE, url, outputDir, opts], {`
30	`timeout: 90000,`
31	`maxBuffer: 10 * 1024 * 1024,`
32	`})`
33	`const lastLine = stdout.trim().split('\n').filter((l) => l.trim().startsWith('{')).at(-1) \|\| '{}'`
34	`const result = JSON.parse(lastLine)`
35	`if (!result.success \|\| !result.file) {`
36	`throw new Error(result.error \|\| 'Scrapling gagal mengambil halaman.')`
37	`}`
38	`return await fs.readFile(result.file, 'utf8')`
39	`} finally {`
40	`await fs.rm(outputDir, { recursive: true, force: true }).catch(() => {})`
41	`}`
42	`}`
43
44	`export function parseWallpapers(html) {`
45	`const ids = []`
46	`const slugs = []`
47	`const alts = []`
48
49	`const imgRe = /<img[^>]src="https:\/\/images\.unsplash\.com\/((?:photo\|flagged\/photo)-[^"?]+)[^"]"[^>]*>/g`
50	`let m`
51	`while ((m = imgRe.exec(html)) !== null) {`
52	`ids.push(m[1])`
53	`const altMatch = /alt="([^"]*)"/.exec(m[0])`
54	`const alt = altMatch ? altMatch[1].replace(/'/g, "'").trim() : ''`
55	`if (alt && !/^Go to /.test(alt) && alt.length >= 4) alts.push(alt)`
56	`}`
57
58	`// 2) ID dari <link rel="preload" imageSrcSet="...photo-XXX..."> (halaman browse)`
59	`const preloadRe = /imageSrcSet="https:\/\/images\.unsplash\.com\/((?:photo\|flagged\/photo)-[^"?]+)/g`
60	`while ((m = preloadRe.exec(html)) !== null) ids.push(m[1])`
61
62	// 3) Slug dari link download `/photos/<slug>/download`
63	`const slugRe = /\/photos\/([a-zA-Z0-9_-]{6,})\/download/g`
64	`while ((m = slugRe.exec(html)) !== null) slugs.push(m[1])`
65
66	`const uniqIds = uniq(ids)`
67	`const uniqSlugs = uniq(slugs)`
68
69	`return uniqIds.map((id, i) => {`
70	`const slug = uniqSlugs[i] \|\| ''`
71	`return {`
72	`id,`
73	`slug,`
74	alt: alts[i] \|\| (slug ? `Unsplash Wallpaper ${slug}` : 'Unsplash Wallpaper'),
75	pageUrl: slug ? `${BASE}/photos/${slug}` : '',
76	`}`
77	`})`
78	`}`
79
80	`/**`
81	`* Ambil daftar wallpaper dari URL halaman Unsplash.`
82	`*/`
83	`export async function fetchWallpapers(url, options = {}) {`
84	`const html = await fetchPageHtml(url)`
85	`const list = parseWallpapers(html)`
86	`if (list.length === 0) throw new Error('Tidak ada wallpaper ditemukan di halaman tersebut.')`
87	`return list`
88	`}`
89
90	`/**`
91	`* Wallpaper acak dari halaman browse /id/wallpapers.`
92	`*/`
93	`export async function browseWallpapers(options = {}) {`
94	return fetchWallpapers(`${BASE}/id/wallpapers`, options)
95	`}`
96
97	`/**`
98	`* Cari wallpaper dengan kata kunci via halaman search.`
99	`*/`
100	`export async function searchWallpapers(query, options = {}) {`
101	`const q = String(query).trim()`
102	`if (!q) throw new Error('Query pencarian kosong.')`
103	return fetchWallpapers(`${BASE}/s/photos/${encodeURIComponent(q)}`, options)
104	`}`
105
106	`/**`
107	`* Bangun URL CDN dengan ukuran tertentu. Default 2560px q90 (HD, ~1-3 MB),`
108	* `hd` true -> 3840px (4K).
109	`*/`
110	`export function buildImageUrl(entry, { w = 2560, q = 90, hd = false } = {}) {`
111	`const width = hd ? 3840 : w`
112	return `https://images.unsplash.com/${entry.id}?w=${width}&q=${q}&fm=jpg&auto=format&fit=max&ixlib=rb-4.1.0`
113	`}`
114
115	`/**`
116	`* Download buffer wallpaper (selalu JPEG).`
117	`*/`
118	`export async function downloadWallpaper(entry, options = {}) {`
119	`const url = buildImageUrl(entry, options)`
120	`const res = await fetch(url, {`
121	`headers: { 'User-Agent': UA, Accept: 'image`

Discussion & Comments0

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