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

improve baixa title parsing

Browse Source
This commit is contained in:
TheBeastLT
2021-09-20 22:27:51 +02:00
parent ccecf68ecb
commit 807f53587e
2 changed files with 6 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
scraper/scrapers/comoeubaixo/comoeubaixo_api.js
View File

@@ -113,13 +113,12 @@ function parseTorrentPage(body) {
const category = details.find('strong:contains(\'Gêneros: \')').next().attr('href').split('/')[0] const category = details.find('strong:contains(\'Gêneros: \')').next().attr('href').split('/')[0]
const torrents = magnets.map(magnetLink => { const torrents = magnets.map(magnetLink => {
const decodedMagnet = decode(magnetLink); const decodedMagnet = decode(magnetLink);
const name = escapeHTML(decodedMagnet.name || '').replace(/\+/g, ' '); const name = sanitizePtName(escapeHTML(decodedMagnet.name || '').replace(/\+/g, ' '));
const sanitizedTitle = sanitizePtName(name);
const originalTitle = details.find('strong:contains(\'Baixar\')')[0].nextSibling.nodeValue.split('-')[0]; const originalTitle = details.find('strong:contains(\'Baixar\')')[0].nextSibling.nodeValue.split('-')[0];
const year = details.find('strong:contains(\'Data de Lançamento: \')').next().text().trim(); const year = details.find('strong:contains(\'Data de Lançamento: \')').next().text().trim();
const fallBackTitle = `${originalTitle.trim()} ${year.trim()} ${sanitizedTitle.trim()}`; const fallBackTitle = `${originalTitle.trim()} ${year.trim()} ${name.trim()}`;
return { return {
title: sanitizedTitle.length > 5 ? sanitizedTitle : fallBackTitle, title: name.length > 5 ? name : fallBackTitle,
infoHash: decodedMagnet.infoHash, infoHash: decodedMagnet.infoHash,
magnetLink: magnetLink, magnetLink: magnetLink,
category: category, category: category,

6
scraper/scrapers/ondebaixa/ondebaixa_api.js
View File

@@ -112,12 +112,12 @@ function parseTorrentPage(body) {
const category = details.find('span:contains(\'Gêneros: \')').next().html() const category = details.find('span:contains(\'Gêneros: \')').next().html()
const torrents = magnets.map(magnetLink => { const torrents = magnets.map(magnetLink => {
const decodedMagnet = decode(magnetLink); const decodedMagnet = decode(magnetLink);
const name = escapeHTML(decodedMagnet.name || '').replace(/\+/g, ' '); const name = sanitizePtName(escapeHTML(decodedMagnet.name || '').replace(/\+/g, ' '));
const originalTitle = details.find('span:contains(\'Título Original: \')').next().text().trim(); const originalTitle = details.find('span:contains(\'Título Original: \')').next().text().trim();
const year = details.find('span:contains(\'Ano de Lançamento: \')').next().text().trim(); const year = details.find('span:contains(\'Ano de Lançamento: \')').next().text().trim();
const fallbackTitle = `${originalTitle} ${year}`; const fallBackTitle = `${originalTitle.trim()} ${year.trim()} ${name.trim()}`;
return { return {
title: name ? sanitizePtName(name) : fallbackTitle, title: name.length > 5 ? name : fallBackTitle,
originalName: sanitizePtOriginalName(originalTitle), originalName: sanitizePtOriginalName(originalTitle),
year: year, year: year,
infoHash: decodedMagnet.infoHash, infoHash: decodedMagnet.infoHash,