From e18e4da559120088d64c150c35d60f7492d4dccf Mon Sep 17 00:00:00 2001 From: Simon Date: Sun, 20 Jul 2025 09:10:07 +0000 Subject: [PATCH] rul34video --- src/api.rs | 45 +++++ src/providers/mod.rs | 15 +- src/providers/rule34video.rs | 312 +++++++++++++++++++++++++++++++++++ 3 files changed, 366 insertions(+), 6 deletions(-) create mode 100644 src/providers/rule34video.rs diff --git a/src/api.rs b/src/api.rs index ceea8db..fa3ae08 100644 --- a/src/api.rs +++ b/src/api.rs @@ -8,6 +8,7 @@ use crate::providers::hanime::HanimeProvider; use crate::providers::perverzija::PerverzijaProvider; use crate::providers::pmvhaven::PmvhavenProvider; use crate::providers::pornhub::PornhubProvider; +use crate::providers::rule34video::Rule34videoProvider; use crate::providers::spankbang::SpankbangProvider; use crate::util::cache::VideoCache; use crate::{DbPool, providers::*, status::*, videos::*}; @@ -387,6 +388,48 @@ async fn status(req: HttpRequest) -> Result { }], nsfw: true, }); + + // pronhub + status.add_channel(Channel { + id: "rule34video".to_string(), + name: "Rule34Video".to_string(), + description: "If it exists, there is porn".to_string(), + premium: false, + favicon: "https://www.google.com/s2/favicons?sz=64&domain=rule34video.com".to_string(), + status: "active".to_string(), + categories: vec![], + options: vec![ChannelOption { + id: "sort".to_string(), + title: "Sort".to_string(), + description: "Sort the Videos".to_string(), //"Sort the videos by Date or Name.".to_string(), + systemImage: "list.number".to_string(), + colorName: "blue".to_string(), + options: vec![ + FilterOption { + id: "post_date".to_string(), + title: "Newest".to_string(), + }, + FilterOption { + id: "video_viewed".to_string(), + title: "Most Viewed".to_string(), + }, + FilterOption { + id: "rating".to_string(), + title: "Top Rated".to_string(), + }, + FilterOption { + id: "duration".to_string(), + title: "Longest".to_string(), + }, + FilterOption { + id: "pseudo_random".to_string(), + title: "Random".to_string(), + }, + ], + multiSelect: false, + }], + nsfw: true, + }); status.iconUrl = format!("http://{}/favicon.ico", host).to_string(); Ok(web::HttpResponse::Ok().json(&status)) } @@ -501,6 +544,8 @@ pub fn get_provider(channel: &str) -> Option { "spankbang" => Some(AnyProvider::Spankbang(SpankbangProvider::new())), "pornhub" => Some(AnyProvider::Pornhub(PornhubProvider::new())), "pmvhaven" => Some(AnyProvider::Pmvhaven(PmvhavenProvider::new())), + "rule34video" => Some(AnyProvider::Rule34video(Rule34videoProvider::new())), + _ => Some(AnyProvider::Perverzija(PerverzijaProvider::new())), } } diff --git a/src/providers/mod.rs b/src/providers/mod.rs index fa2aed6..836e91b 100644 --- a/src/providers/mod.rs +++ b/src/providers/mod.rs @@ -1,11 +1,7 @@ use crate::{ - DbPool, providers::{ - hanime::HanimeProvider, perverzija::PerverzijaProvider, pmvhaven::PmvhavenProvider, - pornhub::PornhubProvider, spankbang::SpankbangProvider, - }, - util::cache::VideoCache, - videos::VideoItem, + hanime::HanimeProvider, perverzija::PerverzijaProvider, pmvhaven::PmvhavenProvider, pornhub::PornhubProvider, rule34video::Rule34videoProvider, spankbang::SpankbangProvider + }, util::cache::VideoCache, videos::VideoItem, DbPool }; pub mod hanime; @@ -13,6 +9,7 @@ pub mod perverzija; pub mod pmvhaven; pub mod pornhub; pub mod spankbang; +pub mod rule34video; pub trait Provider { async fn get_videos( @@ -36,6 +33,7 @@ pub enum AnyProvider { Spankbang(SpankbangProvider), Pornhub(PornhubProvider), Pmvhaven(PmvhavenProvider), + Rule34video(Rule34videoProvider), } impl Provider for AnyProvider { async fn get_videos( @@ -89,6 +87,11 @@ impl Provider for AnyProvider { category.clone(),) .await } + AnyProvider::Rule34video(p) => { + p.get_videos(cache, pool, channel, sort, query, page, per_page, featured, + category.clone(),) + .await + } } } } diff --git a/src/providers/rule34video.rs b/src/providers/rule34video.rs new file mode 100644 index 0000000..e3af978 --- /dev/null +++ b/src/providers/rule34video.rs @@ -0,0 +1,312 @@ +use crate::util::parse_abbreviated_number; +use crate::DbPool; +use crate::providers::Provider; +use crate::util::cache::VideoCache; +use crate::util::flaresolverr::{FlareSolverrRequest, Flaresolverr}; +use crate::util::time::parse_time_to_seconds; +use crate::videos::{VideoItem}; +use error_chain::error_chain; +use futures::stream::SplitSink; +use htmlentity::entity::{ICodedDataTrait, decode}; +use std::env; +use std::vec; +use wreq::{Client, Proxy}; +use wreq_util::Emulation; +use std::time::{SystemTime, UNIX_EPOCH}; + +error_chain! { + foreign_links { + Io(std::io::Error); + HttpRequest(wreq::Error); + } +} + +#[derive(Debug, Clone)] +pub struct Rule34videoProvider { + url: String, +} +impl Rule34videoProvider { + pub fn new() -> Self { + Rule34videoProvider { + url: "https://rule34video.com".to_string(), + } + } + async fn get( + &self, + cache: VideoCache, + page: u8, + sort: &str, + ) -> Result> { + let now = SystemTime::now() + .duration_since(UNIX_EPOCH) + .expect("Time went backwards"); + + let timestamp_millis = now.as_millis(); // u128 + let expected_sorts = vec!["post_date", "video_viewed", "rating", "duration", "pseudo_random"]; + let sort = if expected_sorts.contains(&sort) { + sort + } else { + "post_date" + }; + + let index = format!("rule34video:{}:{}", page, sort); + + let url = format!("{}/?mode=async&function=get_block&block_id=custom_list_videos_most_recent_videos&tag_ids=&sort_by={}&from={}&_={}", self.url, sort, page, timestamp_millis); + + let mut old_items: Vec = vec![]; + if !(sort == "pseudo_random") { + old_items = match cache.get(&index) { + Some((time, items)) => { + if time.elapsed().unwrap_or_default().as_secs() < 60 * 5 { + println!("Cache hit for URL: {}", url); + return Ok(items.clone()); + } else { + items.clone() + } + } + None => { + vec![] + } + }; + } + + let proxy = Proxy::all("http://192.168.0.103:8081").unwrap(); + let client = Client::builder().cert_verification(false).emulation(Emulation::Firefox136).build()?; + + let mut response = client.get(url.clone()).proxy(proxy.clone()).send().await?; + while response.status().is_redirection(){ + response = client.get(self.url.clone() + response.headers()["Location"].to_str().unwrap()).proxy(proxy.clone()).send().await?; + } + if response.status().is_success() { + let text = response.text().await?; + let video_items: Vec = self.get_video_items_from_html(text.clone()); + if !video_items.is_empty() { + cache.remove(&url); + cache.insert(url.clone(), video_items.clone()); + } else { + return Ok(old_items); + } + Ok(video_items) + } else { + let flare_url = env::var("FLARE_URL").expect("FLARE_URL not set"); + let flare = Flaresolverr::new(flare_url); + let result = flare + .solve(FlareSolverrRequest { + cmd: "request.get".to_string(), + url: url.clone(), + maxTimeout: 60000, + }) + .await; + let video_items = match result { + Ok(res) => { + // println!("FlareSolverr response: {}", res); + self.get_video_items_from_html(res.solution.response) + } + Err(e) => { + println!("Error solving FlareSolverr: {}", e); + return Err("Failed to solve FlareSolverr".into()); + } + }; + if !video_items.is_empty() { + cache.remove(&url); + cache.insert(url.clone(), video_items.clone()); + } else { + return Ok(old_items); + } + Ok(video_items) + } + } + async fn query( + &self, + cache: VideoCache, + page: u8, + query: &str, + sort: &str, + ) -> Result> { + let now = SystemTime::now() + .duration_since(UNIX_EPOCH) + .expect("Time went backwards"); + let timestamp_millis = now.as_millis(); // u128 + let expected_sorts = vec!["post_date", "video_viewed", "rating", "duration", "pseudo_random"]; + let sort = if expected_sorts.contains(&sort) { + sort + } else { + "post_date" + }; + + let index = format!("rule34video:{}:{}:{}", page, sort, query); + + let url = format!("{}/search/{}/?mode=async&function=get_block&block_id=custom_list_videos_videos_list_search&tag_ids=&sort_by={}&from={}&_={}", self.url, query, sort, page, timestamp_millis); + + // Check our Video Cache. If the result is younger than 1 hour, we return it. + let old_items = match cache.get(&index) { + Some((time, items)) => { + if time.elapsed().unwrap_or_default().as_secs() < 60 * 5 { + return Ok(items.clone()); + } else { + let _ = cache.check().await; + return Ok(items.clone()); + } + } + None => { + vec![] + } + }; + + let proxy = Proxy::all("http://192.168.0.103:8081").unwrap(); + let client = Client::builder().cert_verification(false).emulation(Emulation::Firefox136).build()?; + + let mut response = client.get(url.clone()).proxy(proxy.clone()).send().await?; + + if response.status().is_redirection(){ + + response = client.get(self.url.clone() + response.headers()["Location"].to_str().unwrap()).proxy(proxy).send().await?; + } + + if response.status().is_success() { + let text = response.text().await?; + let video_items: Vec = self.get_video_items_from_html(text.clone()); + if !video_items.is_empty() { + cache.remove(&url); + cache.insert(url.clone(), video_items.clone()); + } else { + return Ok(old_items); + } + Ok(video_items) + } else { + let flare_url = env::var("FLARE_URL").expect("FLARE_URL not set"); + let flare = Flaresolverr::new(flare_url); + let result = flare + .solve(FlareSolverrRequest { + cmd: "request.get".to_string(), + url: url.clone(), + maxTimeout: 60000, + }) + .await; + let video_items = match result { + Ok(res) => self.get_video_items_from_html(res.solution.response), + Err(e) => { + println!("Error solving FlareSolverr: {}", e); + return Err("Failed to solve FlareSolverr".into()); + } + }; + if !video_items.is_empty() { + cache.remove(&url); + cache.insert(url.clone(), video_items.clone()); + } else { + return Ok(old_items); + } + Ok(video_items) + } + } + + fn get_video_items_from_html(&self, html: String) -> Vec { + if html.is_empty() { + println!("HTML is empty"); + return vec![]; + } + let mut items: Vec = Vec::new(); + let video_listing_content = html.split("
>()[1].split("
>()[0].to_string(); + let raw_videos = video_listing_content + .split("
>()[1..] + .to_vec(); + for video_segment in &raw_videos { + // let vid = video_segment.split("\n").collect::>()[1] + // for (index, line) in vid.iter().enumerate() { + // println!("Line {}: {}", index, line); + // } + + if video_segment.contains("https://rule34video.com/images/advertisements"){ + continue; + } + + let mut title = video_segment.split("
").collect::>()[1] + .split("<") + .collect::>()[0] + .to_string(); + // html decode + title = decode(title.as_bytes()).to_string().unwrap_or(title); + let id = video_segment.split("https://rule34video.com/video/").collect::>()[1].split("/").collect::>()[0].to_string(); + let raw_duration = video_segment.split("
").collect::>()[1] + .split("<") + .collect::>()[0] + .to_string(); + let duration = parse_time_to_seconds(&raw_duration).unwrap_or(0) as u32; + let views = parse_abbreviated_number(&video_segment + .split("
").collect::>()[1].split("").collect::>()[1] + .split("<") + .collect::>()[0]).unwrap_or(0); +//https://rule34video.com/get_file/47/5e71602b7642f9b997f90c979a368c99b8aad90d89/3942000/3942353/3942353_preview.mp4/ +//https://rule34video.com/get_file/47/5e71602b7642f9b997f90c979a368c99b8aad90d89/3942000/3942353/3942353_preview.mp4/ + let thumb = video_segment.split(">()[1].split("data-original=\"").collect::>()[1] + .split("\"") + .collect::>()[0] + .to_string(); + let url = video_segment.split(">()[1] + .split("\"") + .collect::>()[0] + .to_string(); + let preview = video_segment.split("
>()[1] + .split("\"") + .collect::>()[0] + .to_string(); + + + let mut video_item = VideoItem::new( + id, + title, + url.to_string(), + "Rule34video".to_string(), + thumb, + duration, + ) + .views(views) + // .preview(preview) + ; + + + items.push(video_item); + } + return items; + } + + +} + +impl Provider for Rule34videoProvider { + async fn get_videos( + &self, + cache: VideoCache, + pool: DbPool, + _channel: String, + sort: String, + query: Option, + page: String, + per_page: String, + featured: String, + category: String, + ) -> Vec { + let _ = category; + let _ = per_page; + let _ = featured; // Ignored in this implementation + let _ = pool; // Ignored in this implementation + let videos: std::result::Result, Error> = match query { + Some(q) => { + self.query(cache, page.parse::().unwrap_or(1), &q, &sort) + .await + } + None => { + self.get(cache, page.parse::().unwrap_or(1), &sort) + .await + } + }; + match videos { + Ok(v) => v, + Err(e) => { + println!("Error fetching videos: {}", e); + vec![] + } + } + } +}