forked from DebaucheryLibrarian/traxxx
Fixed Nubiles base poster query, handling trailing commas in qu source set. Added profile scene scraper to Dogfart. Added tag photo.
This commit is contained in:
parent
e38922f372
commit
db4e74fb99
|
@ -0,0 +1,5 @@
|
|||
<!-- Generated by IcoMoon.io -->
|
||||
<svg version="1.1" xmlns="http://www.w3.org/2000/svg" width="16" height="16" viewBox="0 0 16 16">
|
||||
<title>feed</title>
|
||||
<path d="M6 8c0-1.105 0.895-2 2-2s2 0.895 2 2c0 1.105-0.895 2-2 2s-2-0.895-2-2zM10.38 3.602c1.56 0.846 2.62 2.498 2.62 4.398s-1.059 3.552-2.62 4.398c0.689-1.096 1.12-2.66 1.12-4.398s-0.431-3.302-1.12-4.398zM4.5 8c0 1.738 0.431 3.302 1.12 4.398-1.56-0.846-2.62-2.498-2.62-4.398s1.059-3.552 2.62-4.398c-0.689 1.096-1.12 2.66-1.12 4.398zM1.5 8c0 2.686 0.85 5.097 2.198 6.746-2.223-1.421-3.698-3.911-3.698-6.746s1.474-5.325 3.698-6.746c-1.348 1.649-2.198 4.060-2.198 6.746zM12.302 1.254c2.223 1.421 3.698 3.911 3.698 6.746s-1.474 5.325-3.698 6.746c1.348-1.649 2.198-4.060 2.198-6.746s-0.85-5.097-2.198-6.746z"></path>
|
||||
</svg>
|
After Width: | Height: | Size: 772 B |
Binary file not shown.
After Width: | Height: | Size: 2.2 MiB |
Binary file not shown.
After Width: | Height: | Size: 5.9 KiB |
Binary file not shown.
After Width: | Height: | Size: 1.4 MiB |
Binary file not shown.
After Width: | Height: | Size: 26 KiB |
|
@ -792,8 +792,9 @@ const tagPhotos = [
|
|||
['dv-tp', 1, 'Adriana Chechik in "Adriana\'s Triple Anal Penetration!"'],
|
||||
['dv-tp', 0, 'Luna Rival in LegalPorno SZ1490'],
|
||||
['ebony', 1, 'Ana Foxxx in "DP Me 4" for HardX'],
|
||||
['facial', 2, 'Ashly Anderson for Hookup Hotshot'],
|
||||
['facial', 3, 'Paige Owens in "Oral Restraint" for Babes'],
|
||||
['facial', 'poster', 'Jynx Maze'],
|
||||
['facial', 2, 'Ashly Anderson for Hookup Hotshot'],
|
||||
['facefucking', 6, 'Halle Hayes in "Towering Temptress" for 5K Porn'],
|
||||
['facefucking', 1, 'Paige Owens in "Dark Meat 12" for Evil Angel'],
|
||||
['facefucking', 0, 'Ashly Anderson in "Rough Love" for Hookup Hotshot'],
|
||||
|
|
|
@ -6,6 +6,8 @@ const { JSDOM } = require('jsdom');
|
|||
const moment = require('moment');
|
||||
|
||||
const http = require('../utils/http');
|
||||
const slugify = require('../utils/slugify');
|
||||
const qu = require('../utils/qu');
|
||||
|
||||
async function getPhotos(albumUrl) {
|
||||
const res = await http.get(albumUrl);
|
||||
|
@ -20,39 +22,39 @@ async function getPhotos(albumUrl) {
|
|||
|
||||
return {
|
||||
url: pageUrl,
|
||||
extract: ({ qu }) => qu.q('.scenes-module img', 'src'),
|
||||
extract: ({ query }) => query.img('.scenes-module img'),
|
||||
};
|
||||
});
|
||||
|
||||
return photoUrls;
|
||||
}
|
||||
|
||||
function scrapeLatest(html, site) {
|
||||
function scrapeLatest(html, site, filter = true) {
|
||||
const { document } = new JSDOM(html).window;
|
||||
const sceneElements = Array.from(document.querySelectorAll('.recent-updates'));
|
||||
|
||||
return sceneElements.reduce((acc, element) => {
|
||||
const siteUrl = element.querySelector('.help-block').textContent;
|
||||
return sceneElements.map((element) => {
|
||||
const siteUrl = element.querySelector('.recent-details-title .help-block, .model-details-title .site-name').textContent;
|
||||
|
||||
if (`www.${siteUrl.toLowerCase()}` !== new URL(site.url).host) {
|
||||
if (filter && `www.${siteUrl.toLowerCase()}` !== new URL(site.url).host) {
|
||||
// different dogfart site
|
||||
return acc;
|
||||
return null;
|
||||
}
|
||||
|
||||
const sceneLinkElement = element.querySelector('.thumbnail');
|
||||
const url = `https://dogfartnetwork.com${sceneLinkElement.href}`;
|
||||
const url = qu.prefixUrl(sceneLinkElement.href, 'https://dogfartnetwork.com');
|
||||
const { pathname } = new URL(url);
|
||||
const entryId = `${site.slug}_${pathname.split('/')[4]}`;
|
||||
|
||||
const title = element.querySelector('.scene-title').textContent;
|
||||
const actors = title.split(/[,&]|\band\b/).map(actor => actor.trim());
|
||||
const actors = title.split(/[,&]|\band\b/).map(actor => actor.replace(/BTS/i, '').trim());
|
||||
|
||||
const poster = `https:${element.querySelector('img').src}`;
|
||||
const teaser = sceneLinkElement.dataset.preview_clip_url;
|
||||
|
||||
return [
|
||||
...acc,
|
||||
{
|
||||
const channel = siteUrl?.match(/(.*).com/)?.[1].toLowerCase();
|
||||
|
||||
return {
|
||||
url,
|
||||
entryId,
|
||||
title,
|
||||
|
@ -62,9 +64,9 @@ function scrapeLatest(html, site) {
|
|||
src: teaser,
|
||||
},
|
||||
site,
|
||||
},
|
||||
];
|
||||
}, []);
|
||||
channel,
|
||||
};
|
||||
}).filter(Boolean);
|
||||
}
|
||||
|
||||
async function scrapeScene(html, url, site) {
|
||||
|
@ -97,8 +99,8 @@ async function scrapeScene(html, url, site) {
|
|||
const poster = `https:${trailerElement.dataset.poster}`;
|
||||
const { trailer } = trailerElement.dataset;
|
||||
|
||||
const lastPhotosUrl = Array.from(document.querySelectorAll('.pagination a')).slice(-1)[0].href;
|
||||
const photos = await getPhotos(`${origin}${pathname}${lastPhotosUrl}`, site, url);
|
||||
const lastPhotosUrl = Array.from(document.querySelectorAll('.pagination a')).slice(-1)[0]?.href;
|
||||
const photos = lastPhotosUrl ? await getPhotos(`${origin}${pathname}${lastPhotosUrl}`, site, url) : [];
|
||||
|
||||
const stars = Math.floor(Number(document.querySelector('span[itemprop="average"]')?.textContent || document.querySelector('span[itemprop="ratingValue"]')?.textContent) / 2);
|
||||
const tags = Array.from(document.querySelectorAll('.scene-details .categories a')).map(({ textContent }) => textContent);
|
||||
|
@ -137,7 +139,23 @@ async function fetchScene(url, site) {
|
|||
return scrapeScene(res.body.toString(), url, site);
|
||||
}
|
||||
|
||||
async function fetchProfile(baseActor, entity) {
|
||||
const slug = slugify(baseActor.name, '+');
|
||||
const url = `https://www.dogfartnetwork.com/tour/girls/${slug}/`;
|
||||
|
||||
const res = await http.get(url);
|
||||
|
||||
if (res.ok) {
|
||||
const scenes = scrapeLatest(res.body, entity, false);
|
||||
|
||||
return { scenes };
|
||||
}
|
||||
|
||||
return res.status;
|
||||
}
|
||||
|
||||
module.exports = {
|
||||
fetchLatest,
|
||||
fetchScene,
|
||||
fetchProfile,
|
||||
};
|
||||
|
|
|
@ -45,8 +45,9 @@ function scrapeAll(scenes, site, origin) {
|
|||
release.date = qu.date('.date', 'MMM D, YYYY');
|
||||
release.actors = qu.all('.models a.model', true);
|
||||
|
||||
const poster = qu.q('img').dataset.original;
|
||||
release.poster = [
|
||||
const poster = qu.sourceSet('img', 'data-srcset')?.[0];
|
||||
|
||||
release.poster = poster && [
|
||||
poster.replace('_640', '_1280'),
|
||||
poster,
|
||||
];
|
||||
|
|
|
@ -197,6 +197,7 @@ const scrapers = {
|
|||
devilsfilm: famedigital,
|
||||
digitalplayground,
|
||||
dtfsluts: fullpornnetwork,
|
||||
dogfartnetwork: dogfart,
|
||||
dorcelclub: dorcel,
|
||||
doubleviewcasting: firstanalquest,
|
||||
elegantangel,
|
||||
|
|
|
@ -253,7 +253,7 @@ function urls(context, selector = 'a', attr = 'href', { origin, protocol = 'http
|
|||
return attr ? urlEls.map(urlEl => prefixUrl(urlEl, origin, protocol)) : urlEls;
|
||||
}
|
||||
|
||||
function sourceSet(context, selector, attr, options = {}) {
|
||||
function sourceSet(context, selector, attr = 'srcset', options = {}) {
|
||||
const srcset = q(context, selector, attr);
|
||||
|
||||
if (!srcset) {
|
||||
|
@ -265,11 +265,16 @@ function sourceSet(context, selector, attr, options = {}) {
|
|||
.map((source) => {
|
||||
const [link, descriptor] = source.split(' ');
|
||||
|
||||
if (link) {
|
||||
return {
|
||||
descriptor: descriptor || 'fallback',
|
||||
url: prefixUrl(link, options.origin, options.protocol),
|
||||
};
|
||||
}
|
||||
|
||||
return null;
|
||||
})
|
||||
.filter(Boolean)
|
||||
.sort((sourceA, sourceB) => {
|
||||
if (sourceB.descriptor === 'fallback' || parseInt(sourceA.descriptor, 10) > parseInt(sourceB.descriptor, 10)) {
|
||||
return -1;
|
||||
|
|
|
@ -79,6 +79,7 @@ module.exports = {
|
|||
theme: path.join(__dirname, 'assets/css/_theme.scss'),
|
||||
breakpoints: path.join(__dirname, 'assets/css/_breakpoints.scss'),
|
||||
config: path.join(__dirname, `assets/js/config/${process.env.NODE_ENV || 'default'}.js`),
|
||||
vue: 'vue/dist/vue.esm-bundler.js',
|
||||
},
|
||||
},
|
||||
};
|
||||
|
|
Loading…
Reference in New Issue