Gistify
Back to Home
Gistify Snippet
kyiov
@kyiov
3d ago·1 file(s)
View profile
Public

Otakudesu Anime Scraper & Detail Fetcher

Scrapes Otakudesu.fit for anime search results and detailed information including title, image, status, studio, genres, and episode lists using Cheerio.

#scraper#anime#otakudesu#cheerio#nodejs#javascript
6
Raw
otakudesu.jsjavascript
240 lines
1import * as cheerio from 'cheerio'
2import { pathToFileURL } from 'node:url'
3import dns from 'node:dns'
4
5dns.setDefaultResultOrder('ipv4first')
6
7/**
8 * Otakudesu Scraper (otakudesu.fit)
9 * Sumber: kyiov/mithra-ex (system/lib/scraper/otakudesu.js)
10 */
11
12const BASE = 'https://otakudesu.fit'
13const UA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/126.0.0.0 Safari/537.36'
14
15async function get(url) {
16 const r = await fetch(url, {
17 headers: {
18 'User-Agent': UA,
19 Accept: 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
20 'Accept-Language': 'id-ID,id;q=0.9,en;q=0.8',
21 Referer: BASE + '/',
22 Cookie: 'sg=1', // Bypass searchguard check
23 },
24 redirect: 'follow',
25 signal: AbortSignal.timeout(25000),
26 })
27 const body = await r.text()
28 if (r.status !== 200) throw new Error(`HTTP ${r.status} untuk ${url}`)
29 return body
30}
31
32function toAbs(u) {
33 if (!u) return ''
34 if (u.startsWith('http')) return u
35 return BASE + (u.startsWith('/') ? u : '/' + u)
36}
37
38/**
39 * Cari anime via ?s=<query>&post_type=anime
40 */
41export async function searchAnime(query) {
42 const q = String(query || '').trim()
43 if (!q) return []
44 try {
45 const html = await get(`${BASE}/?s=${encodeURIComponent(q)}&post_type=anime`)
46 const $ = cheerio.load(html)
47 const results = []
48 $('article.bs').each((i, el) => {
49 const a = $(el).find('a.tip[href*="/series/"]').first()
50 const href = a.attr('href') || ''
51 const title = (a.attr('title') || a.text().trim() || '').replace(/\s+/g, ' ')
52 const img = $(el).find('img').first().attr('src') || ''
53 if (!href || !title) return
54 results.push({
55 source: 'otakudesu',
56 title,
57 url: toAbs(href),
58 image: img,
59 type: $(el).find('.typez').first().text().trim() || 'Anime',
60 rating: $(el).find('.rating').first().text().trim() || '',
61 status: $(el).find('.epx, .status').first().text().trim() || '',
62 })
63 })
64 return results
65 } catch (e) {
66 console.error('[otakudesu] search error:', e.message)
67 return []
68 }
69}
70
71/**
72 * Ambil detail anime: info, genre, sinopsis, dan daftar episode
73 */
74export async function getAnimeDetail(url) {
75 const html = await get(url)
76 const $ = cheerio.load(html)
77
78 const title = $('h1').first().text().trim() || $('title').text().replace(/ - Otaku Desu.*$/i, '').trim()
79 const img = $('.single-info .thumb img, .info-left img, .bixbox .thumb img').first().attr('src') || ''
80
81 let status = ''
82 let studio = ''
83 const spe = $('.spe').first().text().replace(/\s+/g, ' ')
84 const stM = spe.match(/Status:\s*([^,]+?)(?:\s+(?:Studio|Released|Season|Type|Duration|Score|Producer|Total|Rating):|$)/i)
85 if (stM) status = stM[1].trim()
86 const stuM = spe.match(/Studio:\s*([^,]+?)(?:\s+(?:Released|Season|Type|Duration|Score|Producer|Total|Rating):|$)/i)
87 if (stuM) studio = stuM[1].trim()
88
89 const genres = []
90 $('a[href*="/genres/"]').each((i, el) => {
91 const t = $(el).text().trim()
92 if (t && !genres.includes(t)) genres.push(t)
93 })
94
95 const episodes = []
96 const seen = new Set()
97
98 // Parsing dari daftar .eplister jika tersedia
99 $('.eplister ul li').each((i, el) => {
100 const a = $(el).find('a').first()
101 const href = a.attr('href') || ''
102 const epNumStr = $(el).find('.epl-num').first().text().trim()
103 const epTitle = $(el).find('.epl-title').first().text().trim() || a.text().trim()
104 const date = $(el).find('.epl-date').first().text().trim()
105 const num = parseInt(epNumStr, 10) || i + 1
106
107 if (href && !seen.has(href)) {
108 seen.add(href)
109 episodes.push({
110 number: num,
111 title: epTitle || `Episode ${num}`,
112 url: toAbs(href),
113 date: date || '',
114 })
115 }
116 })
117
118 // Fallback ke pola regex jika eplister kosong
119 if (episodes.length === 0) {
120 $('a[href*="-episode-"]').each((i, el) => {
121 const href = $(el).attr('href') || ''
122 const m = href.match(/-episode-(\d+)(?:-subtitle-indonesia)?/i)
123 if (!m) return
124 const num = parseInt(m[1], 10)
125 if (seen.has(href)) return
126 seen.add(href)
127 episodes.push({
128 number: num,
129 title: `Episode ${num}`,
130 url: toAbs(href),
131 date: '',
132 })
133 })
134 }
135
136 episodes.sort((a, b) => a.number - b.number)
137 const synopsis = $('.entry-content[itemprop="description"], .sinopc').first().text().trim() || ''
138
139 return { source: 'otakudesu', title, image: img, status, studio, genres, synopsis, episodes }
140}
141
142/**
143 * Ambil episode: iframe video embed dan link download
144 */
145export async function getEpisode(url) {
146 const html = await get(url)
147 const $ = cheerio.load(html)
148
149 const title = $('h1').first().text().trim() || $('title').text().replace(/ - Otaku Desu.*$/i, '').trim()
150
151 let iframe = ''
152 $('iframe').each((i, el) => {
153 const src = $(el).attr('src') || $(el).attr('data-src') || ''
154 if (src) { iframe = src; return false }
155 })
156
157 const downloadLinks = []
158 const seenDl = new Set()
159 $('.dlbox a, .downloadbox a').each((i, el) => {
160 const href = $(el).attr('href') || ''
161 if (!href || href === '#' || seenDl.has(href)) return
162 seenDl.add(href)
163 downloadLinks.push({
164 label: $(el).text().trim().replace(/\s+/g, ' ').slice(0, 60) || 'link',
165 url: href
166 })
167 })
168
169 return { source: 'otakudesu', title, iframe, downloadLinks }
170}
171
172export default { searchAnime, getAnimeDetail, getEpisode }
173
174// ─────────────────────────────────────────────────────────────────────
175// Direct Execution / CLI Runner
176// ─────────────────────────────────────────────────────────────────────
177if (process.argv[1] && import.meta.url === pathToFileURL(process.argv[1]).href) {
178 const args = process.argv.slice(2)
179
180 ;(async () => {
181 try {
182 if (args.length === 0) {
183 console.log('[+] Menjalankan default search: "one piece"...\n')
184 const results = await searchAnime('one piece')
185 console.log(JSON.stringify(results, null, 2))
186 return
187 }
188
189 const cmd = args[0].toLowerCase()
190
191 if (cmd === 'search' || cmd === 'cari') {
192 const query = args.slice(1).join(' ') || 'one piece'
193 console.log(`[+] Mencari: "${query}"...`)
194 const results = await searchAnime(query)
195 console.log(JSON.stringify(results, null, 2))
196 } else if (cmd === 'detail') {
197 const url = args[1]
198 if (!url) {
199 console.error('Format: node otakudesu.js detail <url>')
200 process.exit(1)
201 }
202 console.log(`[+] Detail: ${url}`)
203 const detail = await getAnimeDetail(url)
204 console.log(JSON.stringify(detail, null, 2))
205 } else if (cmd === 'episode' || cmd === 'ep') {
206 const url = args[1]
207 if (!url) {
208 console.error('Format: node otakudesu.js episode <url>')
209 process.exit(1)
210 }
211 console.log(`[+] Episode: ${url}`)
212 const ep = await getEpisode(url)
213 console.log(JSON.stringify(ep, null, 2))
214 } else if (args[0].startsWith('http://') || args[0].startsWith('https://')) {
215 const url = args[0]
216 if (url.includes('/series/')) {
217 console.log(`[+] Terdeteksi Series: ${url}`)
218 const detail = await getAnimeDetail(url)
219 console.log(JSON.stringify(detail, null, 2))
220 } else if (url.includes('-episode-')) {
221 console.log(`[+] Terdeteksi Episode: ${url}`)
222 const ep = await getEpisode(url)
223 console.log(JSON.stringify(ep, null, 2))
224 } else {
225 const detail = await getAnimeDetail(url)
226 console.log(JSON.stringify(detail, null, 2))
227 }
228 } else {
229 const query = args.join(' ')
230 console.log(`[+] Mencari: "${query}"...`)
231 const results = await searchAnime(query)
232 console.log(JSON.stringify(results, null, 2))
233 }
234 } catch (err) {
235 console.error('Error:', err.message || err)
236 process.exitCode = 1
237 }
238 })()
239}
240
Discussion & Comments0

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

Sign in with your GitHub account to write comments and join the discussion.