tools/knightcrawler
1
0
Fork 0
mirror of https://github.com/knightcrawler-stremio/knightcrawler.git synced 2024-12-20 03:29:51 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Updated scrapers to latest available commit

Browse Source
This commit is contained in:
Gabisonfire
2024-01-17 16:43:58 -05:00
parent cab7f38c66
commit 909ade0d8e
39 changed files with 311 additions and 299 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
scraper/scrapers/1337x/1337x_api.js
View File

@@ -1,10 +1,11 @@
const axios = require('axios');
const cheerio = require('cheerio');
const needle = require('needle');
const Sugar = require('sugar-date');
const decode = require('magnet-uri');
const Promises = require('../../lib/promises');
const { escapeHTML } = require('../../lib/metadata');
const { getRandomUserAgent } = require('../../lib/requestHelper');
const { parseSize } = require("../scraperHelper");
const defaultProxies = [
'https://1337x.to'
@@ -32,7 +33,7 @@ function torrent(torrentId, config = {}, retries = 2) {
const slug = torrentId.startsWith('/torrent/') ? torrentId.replace('/torrent/', '') : torrentId;
return Promises.first(proxyList
.map((proxyUrl) => singleRequest(`${proxyUrl}/torrent/${slug}`, config)))
.map((proxyUrl) => singleRequest(`${proxyUrl}/torrent/${slug}`, config)))
.then((body) => parseTorrentPage(body))
.then((torrent) => ({ torrentId: slug, ...torrent }))
.catch((err) => torrent(slug, config, retries - 1));
@@ -51,7 +52,7 @@ function search(keyword, config = {}, retries = 2) {
: `${proxyUrl}/search/${keyword}/${page}/`;
return Promises.first(proxyList
.map(proxyUrl => singleRequest(requestUrl(proxyUrl), config)))
.map(proxyUrl => singleRequest(requestUrl(proxyUrl), config)))
.then(body => parseTableBody(body))
.then(torrents => torrents.length === 40 && page < extendToPage
? search(keyword, { ...config, page: page + 1 }).catch(() => [])
@@ -73,18 +74,18 @@ function browse(config = {}, retries = 2) {
: `${proxyUrl}/cat/${category}/${page}/`;
return Promises.first(proxyList
.map((proxyUrl) => singleRequest(requestUrl(proxyUrl), config)))
.map((proxyUrl) => singleRequest(requestUrl(proxyUrl), config)))
.then((body) => parseTableBody(body))
.catch((err) => browse(config, retries - 1));
}
function singleRequest(requestUrl, config = {}) {
const timeout = config.timeout || defaultTimeout;
const options = { userAgent: getRandomUserAgent(), open_timeout: timeout, follow: 2 };
const options = { headers: { 'User-Agent': getRandomUserAgent() }, timeout: timeout };
return needle('get', requestUrl, options)
return axios.get(requestUrl, options)
.then((response) => {
const body = response.body;
const body = response.data;
if (!body) {
throw new Error(`No body: ${requestUrl}`);
} else if (body.includes('502: Bad gateway') ||
@@ -164,19 +165,4 @@ function parseDate(dateString) {
return Sugar.Date.create(dateString);
}
function parseSize(sizeText) {
if (!sizeText) {
return undefined;
}
let scale = 1;
if (sizeText.includes('GB')) {
scale = 1024 * 1024 * 1024
} else if (sizeText.includes('MB')) {
scale = 1024 * 1024;
} else if (sizeText.includes('KB')) {
scale = 1024;
}
return Math.floor(parseFloat(sizeText.replace(/,/g, '')) * scale);
}
module.exports = { torrent, search, browse, Categories };

2
scraper/scrapers/1337x/1337x_scraper.js
View File

@@ -92,7 +92,7 @@ async function processTorrentRecord(record) {
function typeMapping() {
const mapping = {};
mapping[leetx.Categories.MOVIE] = Type.MOVIE;
mapping[leetx.Categories.DOCUMENTARIES] = Type.SERIES;
mapping[leetx.Categories.DOCUMENTARIES] = Type.MOVIE;
mapping[leetx.Categories.TV] = Type.SERIES;
mapping[leetx.Categories.ANIME] = Type.ANIME;
return mapping;