all extensions to new format :P
This commit is contained in:
168
novelbin.js
168
novelbin.js
@@ -1,129 +1,121 @@
|
||||
class NovelBin {
|
||||
constructor(fetchPath, cheerioPath, browser) {
|
||||
this.browser = browser;
|
||||
this.fetch = require(fetchPath);
|
||||
this.cheerio = require(cheerioPath);
|
||||
constructor() {
|
||||
this.baseUrl = "https://novelbin.me";
|
||||
this.type = "book-board";
|
||||
this.type = "book-board";
|
||||
this.mediaType = "ln";
|
||||
}
|
||||
|
||||
async fetchSearchResult(query = "", page = 1) {
|
||||
const url = !query || query.trim() === ""
|
||||
? `${this.baseUrl}/sort/novelbin-hot?page=${page}`
|
||||
: `${this.baseUrl}/search?keyword=${encodeURIComponent(query)}`;
|
||||
async search(queryObj) {
|
||||
const query = queryObj.query || "";
|
||||
const url = `${this.baseUrl}/search?keyword=${encodeURIComponent(query)}`;
|
||||
|
||||
const res = await fetch(url, {
|
||||
headers: {
|
||||
"user-agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/139.0.0.0 Safari/537.36",
|
||||
"accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
|
||||
"referer": this.baseUrl + "/"
|
||||
}
|
||||
});
|
||||
|
||||
const res = await this.fetch(url);
|
||||
const html = await res.text();
|
||||
const $ = this.cheerio.load(html);
|
||||
|
||||
const results = [];
|
||||
|
||||
$(".list-novel .row, .col-novel-main .list-novel .row").each((i, el) => {
|
||||
const titleEl = $(el).find("h3.novel-title a");
|
||||
if (!titleEl.length) return;
|
||||
|
||||
const title = titleEl.text().trim();
|
||||
let href = titleEl.attr("href");
|
||||
if (!href) return;
|
||||
|
||||
if (!href.startsWith("http")) { href = `${this.baseUrl}${href}` }
|
||||
$('h3.novel-title a').each((i, el) => {
|
||||
const href = $(el).attr('href');
|
||||
const title = $(el).text().trim();
|
||||
|
||||
const idMatch = href.match(/novel-book\/([^/?]+)/);
|
||||
const id = idMatch ? idMatch[1] : null;
|
||||
if (!id) return;
|
||||
const coverUrl = `${this.baseUrl}/media/novel/${id}.jpg`;
|
||||
|
||||
const img = `${this.baseUrl}/media/novel/${id}.jpg`;
|
||||
|
||||
results.push({
|
||||
id,
|
||||
title,
|
||||
image: coverUrl,
|
||||
sampleImageUrl: coverUrl,
|
||||
tags: [],
|
||||
image: img,
|
||||
rating: null,
|
||||
type: "book"
|
||||
});
|
||||
});
|
||||
|
||||
const hasNextPage = $(".PagedList-skipToNext a").length > 0;
|
||||
return results;
|
||||
}
|
||||
|
||||
async getMetadata(id) {
|
||||
const res = await fetch(`${this.baseUrl}/novel-book/${id}`);
|
||||
const html = await res.text();
|
||||
const $ = this.cheerio.load(html);
|
||||
|
||||
const getMeta = (property) => $(`meta[property='${property}']`).attr('content') || "";
|
||||
|
||||
const title = getMeta("og:novel:novel_name") || $('title').text() || "";
|
||||
const summary = $('meta[name="description"]').attr('content') || "";
|
||||
const genresRaw = getMeta("og:novel:genre");
|
||||
const genres = genresRaw ? genresRaw.split(',').map(g => g.trim()) : [];
|
||||
const status = getMeta("og:novel:status") || "";
|
||||
const image = getMeta("og:image");
|
||||
|
||||
const lastChapterName = getMeta("og:novel:lastest_chapter_name");
|
||||
const chaptersMatch = lastChapterName.match(/Chapter\s+(\d+)/i);
|
||||
const chapters = chaptersMatch ? Number(chaptersMatch[1]) : 0;
|
||||
|
||||
return {
|
||||
results,
|
||||
hasNextPage,
|
||||
page
|
||||
id,
|
||||
title,
|
||||
format: "Light Novel",
|
||||
score: 0,
|
||||
genres,
|
||||
status,
|
||||
published: "???",
|
||||
summary,
|
||||
chapters,
|
||||
image
|
||||
};
|
||||
}
|
||||
|
||||
async findChapters(bookId) {
|
||||
const res = await this.fetch(`${this.baseUrl}/novel-book/${bookId}`);
|
||||
const res = await fetch(`${this.baseUrl}/ajax/chapter-archive?novelId=${bookId}`, {
|
||||
headers: {
|
||||
"user-agent": "Mozilla/5.0"
|
||||
}
|
||||
});
|
||||
|
||||
const html = await res.text();
|
||||
const $ = this.cheerio.load(html);
|
||||
|
||||
const chapters = [];
|
||||
|
||||
$("#chapter-archive ul.list-chapter li a").each((i, el) => {
|
||||
const a = $(el);
|
||||
const title = a.attr("title") || a.text().trim();
|
||||
let href = a.attr("href");
|
||||
|
||||
if (!href) return;
|
||||
|
||||
if (href.startsWith("https://novelbin.me")) { href = href.replace("https://novelbin.me", "") }
|
||||
const match = title.match(/chapter\s*([\d.]+)/i);
|
||||
const chapterNumber = match ? match[1] : "0";
|
||||
$('a[title]').each((i, el) => {
|
||||
const fullUrl = $(el).attr('href');
|
||||
const title = $(el).attr('title').trim();
|
||||
const numMatch = title.match(/chapter\s+(\d+(?:\.\d+)?)/i);
|
||||
|
||||
chapters.push({
|
||||
id: href,
|
||||
title: title.trim(),
|
||||
chapter: chapterNumber,
|
||||
language: "en"
|
||||
id: fullUrl,
|
||||
title,
|
||||
number: numMatch ? numMatch[1] : "0",
|
||||
releaseDate: null,
|
||||
index: i
|
||||
});
|
||||
});
|
||||
|
||||
return { chapters: chapters };
|
||||
return chapters;
|
||||
}
|
||||
|
||||
async findChapterPages(chapterId) {
|
||||
const url = chapterId.startsWith('http') ? chapterId : `${this.baseUrl}${chapterId}`;
|
||||
|
||||
const content = await this.browser.scrape(
|
||||
url,
|
||||
() => {
|
||||
const contentDiv = document.querySelector('#chr-content, .chr-c');
|
||||
if (!contentDiv) return "<p>Error: Could not find content.</p>";
|
||||
contentDiv.querySelectorAll('script, div[id^="pf-"], div[style*="text-align:center"], ins, div[align="center"], .ads, .adsbygoogle').forEach(el => el.remove());
|
||||
|
||||
const paragraphs = contentDiv.querySelectorAll('p');
|
||||
let cleanHtml = '';
|
||||
|
||||
paragraphs.forEach(p => {
|
||||
let text = p.textContent || '';
|
||||
|
||||
text = text.replace(/△▼△▼△▼△/g, '');
|
||||
text = text.replace(/[※\s]{2,}/g, '');
|
||||
|
||||
const html = p.innerHTML.trim();
|
||||
|
||||
const isAd = text.includes('Remove Ads From') || text.includes('Buy no ads experience');
|
||||
const isEmpty = html === '' || html === ' ' || text.trim() === '';
|
||||
|
||||
if (!isAd && !isEmpty) {
|
||||
if (p.textContent !== text) {
|
||||
p.textContent = text;
|
||||
}
|
||||
cleanHtml += p.outerHTML;
|
||||
}
|
||||
});
|
||||
|
||||
return cleanHtml || contentDiv.innerHTML;
|
||||
},
|
||||
{ waitSelector: '#chr-content', timeout: 2000 }
|
||||
);
|
||||
|
||||
return [{
|
||||
type: 'text',
|
||||
content: content,
|
||||
index: 0
|
||||
}];
|
||||
async findChapterPages(chapterUrl) {
|
||||
const {result} = await this.scrape(chapterUrl, async (page) => {
|
||||
return page.evaluate(() => {
|
||||
document.querySelectorAll('div[id^="pf-"]').forEach(e => e.remove());
|
||||
const ps = Array.from(document.querySelectorAll("p")).map(p => p.outerHTML.trim()).filter(p => p.length > 7);
|
||||
return ps.join("\n");
|
||||
});
|
||||
}, {
|
||||
waitUntil: "domcontentloaded",
|
||||
renderWaitTime: 300
|
||||
});
|
||||
return result || "<p>Error: chapter text not found</p>";
|
||||
}
|
||||
}
|
||||
|
||||
module.exports = { NovelBin };
|
||||
module.exports = NovelBin;
|
||||
Reference in New Issue
Block a user