Added Bang! scraper. Showing studio on tile. Added favicons to tiles.

This commit is contained in:
2020-01-06 05:19:38 +01:00
parent 3a7aa58abb
commit bb5b8085f6
90 changed files with 430 additions and 45 deletions

View File

@@ -180,6 +180,23 @@ async function attachChannelSite(release) {
};
}
async function attachStudio(release) {
if (!release.studio) {
return release;
}
const studio = await knex('studios')
.where('name', release.studio)
.orWhere('slug', release.studio)
.orWhere('url', release.studio)
.first();
return {
...release,
studio,
};
}
async function curateReleaseEntry(release) {
const curatedRelease = {
site_id: release.site.id,
@@ -295,27 +312,31 @@ async function storeReleaseAssets(release, releaseId) {
await createMediaDirectory('releases', subpath);
try {
await Promise.all([
storePhotos(release.photos, {
targetId: releaseId,
subpath,
primaryRole: release.poster ? null : 'poster',
}, identifier),
release.poster && storePhotos([release.poster], {
// don't use Promise.all to prevent concurrency issues with duplicate detection
if (release.poster) {
await storePhotos([release.poster], {
role: 'poster',
targetId: releaseId,
subpath,
}, identifier),
storePhotos(release.covers, {
role: 'cover',
targetId: releaseId,
subpath,
}, identifier),
storeTrailer(release.trailer, {
targetId: releaseId,
subpath,
}, identifier),
]);
}, identifier);
}
await storePhotos(release.photos, {
targetId: releaseId,
subpath,
primaryRole: release.poster ? null : 'poster',
}, identifier);
await storePhotos(release.covers, {
role: 'cover',
targetId: releaseId,
subpath,
}, identifier);
await storeTrailer(release.trailer, {
targetId: releaseId,
subpath,
}, identifier);
} catch (error) {
console.log(release.url, error);
}
@@ -363,7 +384,8 @@ async function storeReleases(releases) {
const storedReleases = await Promise.map(releases, async (release) => {
try {
const releaseWithChannelSite = await attachChannelSite(release);
const releaseId = await storeRelease(releaseWithChannelSite);
const releaseWithStudio = await attachStudio(release);
const releaseId = await storeRelease(releaseWithStudio);
return {
id: releaseId,

127
src/scrapers/bang.js Normal file
View File

@@ -0,0 +1,127 @@
'use strict';
const bhttp = require('bhttp');
function encodeId(id) {
return Buffer
.from(id, 'hex')
.toString('base64')
.replace(/\+/g, '-')
.replace(/\//g, '_')
.replace(/=/g, ',');
}
function scrapeLatest(scenes, site) {
return scenes.map(({ _source: scene }) => {
const release = {
site,
entryId: encodeId(scene.id),
title: scene.name,
description: scene.description,
actors: scene.actors.map(actor => actor.name),
tags: scene.genres.concat(scene.actions).map(genre => genre.name),
duration: scene.duration,
};
const slug = release.title.toLowerCase().trim().replace(/\s+/g, '-');
release.url = `https://www.bang.com/video/${release.entryId}/${slug}`;
const date = new Date(scene.releaseDate);
release.date = new Date(Date.UTC(date.getUTCFullYear(), date.getUTCMonth(), date.getUTCDate()));
if (scene.is4k) release.tags.push('4k');
if (scene.gay) release.tags.push('gay');
const defaultPoster = scene.screenshots.find(photo => photo.default === true);
const photoset = scene.screenshots.filter(photo => photo.default === false);
const photos = defaultPoster ? photoset : photoset.slice(1);
const poster = defaultPoster || photoset[0];
release.poster = `https://i.bang.com/screenshots/${scene.dvd.id}/movie/1/${poster.screenId}.jpg`;
release.photos = photos.map(photo => `https://i.bang.com/screenshots/${scene.dvd.id}/movie/1/${photo.screenId}.jpg`);
release.trailer = {
src: `https://i.bang.com/v/${scene.dvd.id}/${scene.identifier}/preview.mp4`,
};
release.studio = scene.series.name
.replace(/[! .]/g, '')
.replace('&', 'and');
return release;
});
}
async function fetchLatest(site, page = 1) {
const clusterId = '617fb597b659459bafe6472470d9073a';
const authKey = 'YmFuZy1yZWFkOktqVDN0RzJacmQ1TFNRazI=';
const res = await bhttp.post(`https://${clusterId}.us-east-1.aws.found.io/videos/video/_search`, {
size: 50,
from: (page - 1) * 50,
query: {
bool: {
must: [
{
match: {
status: 'ok',
},
},
{
range: {
releaseDate: {
lte: 'now',
},
},
},
{
nested: {
path: 'studio',
query: {
bool: {
must: [
{
match: {
'studio.name': {
operator: 'AND',
query: 'bang! originals',
},
},
},
],
},
},
},
},
],
must_not: [
{
match: {
type: 'trailer',
},
},
],
},
},
sort: [
{
releaseDate: {
order: 'desc',
},
},
],
}, {
encodeJSON: true,
headers: {
Authorization: `Basic ${authKey}`,
},
});
return scrapeLatest(res.body.hits.hits, site);
}
module.exports = {
fetchLatest,
// fetchScene,
};

View File

@@ -4,9 +4,6 @@ const bhttp = require('bhttp');
const { JSDOM } = require('jsdom');
const cheerio = require('cheerio');
const moment = require('moment');
const knex = require('../knex');
const { matchTags } = require('../tags');
function extractTitle(originalTitle) {
const titleComponents = originalTitle.split(' ');
@@ -102,16 +99,8 @@ async function scrapeScene(html, url, site, useGallery) {
const trailer = data.clip.qualities.find(clip => clip.quality === 'vga' || clip.quality === 'hd');
const studioName = $('.watchpage-studioname').first().text().trim();
const studioSlug = studioName.replace(/\s+/g, '').toLowerCase();
const rawTags = $(tagsElement).find('a').map((tagIndex, tagElement) => $(tagElement).text()).toArray();
const [studio, tags] = await Promise.all([
knex('studios')
.where({ name: studioName })
.orWhere({ slug: studioSlug })
.first(),
matchTags(rawTags),
]);
const studio = studioName.replace(/[\s.']+/g, '').toLowerCase();
const tags = $(tagsElement).find('a').map((tagIndex, tagElement) => $(tagElement).text()).toArray();
return {
url,

View File

@@ -2,6 +2,7 @@
// releases
const twentyonesextury = require('./21sextury');
const bang = require('./bang');
const bangbros = require('./bangbros');
const blowpass = require('./blowpass');
const dogfart = require('./dogfart');
@@ -31,6 +32,7 @@ const pornhub = require('./pornhub');
module.exports = {
releases: {
'21sextury': twentyonesextury,
bang,
bangbros,
blowpass,
brazzers,

View File

@@ -3,7 +3,7 @@
const path = require('path');
const Promise = require('bluebird');
const fs = require('fs-extra');
const fetchScene = require('../scrape-release');
const fetchScene = require('../scrape-releases');
const argv = require('../argv');