archived tube8 until its ready
This commit is contained in:
Simon
434
archive/tube8.rs
Normal file
434
archive/tube8.rs
Normal file
@@ -0,0 +1,434 @@
|
||||
use crate::DbPool;
|
||||
use crate::api::ClientVersion;
|
||||
use crate::providers::Provider;
|
||||
use crate::status::*;
|
||||
use crate::util::cache::VideoCache;
|
||||
use crate::util::parse_abbreviated_number;
|
||||
use crate::util::time::parse_time_to_seconds;
|
||||
use crate::videos::{ServerOptions, VideoItem};
|
||||
use async_trait::async_trait;
|
||||
use error_chain::error_chain;
|
||||
use htmlentity::entity::{ICodedDataTrait, decode};
|
||||
use std::sync::{Arc, RwLock};
|
||||
use std::vec;
|
||||
|
||||
error_chain! {
|
||||
foreign_links {
|
||||
Io(std::io::Error);
|
||||
HttpRequest(wreq::Error);
|
||||
}
|
||||
}
|
||||
|
||||
#[derive(Debug, Clone)]
|
||||
pub struct Tube8Provider {
|
||||
url: String,
|
||||
sites: Arc<RwLock<Vec<FilterOption>>>,
|
||||
stars: Arc<RwLock<Vec<FilterOption>>>,
|
||||
}
|
||||
impl Tube8Provider {
|
||||
pub fn new() -> Self {
|
||||
let provider = Tube8Provider {
|
||||
url: "https://www.tube8.com".to_string(),
|
||||
sites: Arc::new(RwLock::new(vec![FilterOption {
|
||||
id: "all".to_string(),
|
||||
title: "All".to_string(),
|
||||
}])),
|
||||
stars: Arc::new(RwLock::new(vec![FilterOption {
|
||||
id: "all".to_string(),
|
||||
title: "All".to_string(),
|
||||
}])),
|
||||
};
|
||||
|
||||
// Kick off the background load but return immediately
|
||||
provider
|
||||
}
|
||||
|
||||
// Push one item with minimal lock time and dedup by id
|
||||
fn push_unique(target: &Arc<RwLock<Vec<FilterOption>>>, item: FilterOption) {
|
||||
if let Ok(mut vec) = target.write() {
|
||||
if !vec.iter().any(|x| x.id == item.id) {
|
||||
vec.push(item);
|
||||
// Optional: keep it sorted for nicer UX
|
||||
// vec.sort_by(|a,b| a.title.cmp(&b.title));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fn build_channel(&self, clientversion: ClientVersion) -> Channel {
|
||||
let _ = clientversion;
|
||||
let sites: Vec<FilterOption> = self
|
||||
.sites
|
||||
.read()
|
||||
.map(|g| g.clone()) // or: .map(|g| g.to_vec())
|
||||
.unwrap_or_default(); // or: .unwrap_or_else(|_| Vec::new())
|
||||
|
||||
let stars: Vec<FilterOption> = self
|
||||
.stars
|
||||
.read()
|
||||
.map(|g| g.clone()) // or: .map(|g| g.to_vec())
|
||||
.unwrap_or_default(); // or: .unwrap_or_else(|_| Vec::new())
|
||||
|
||||
Channel {
|
||||
id: "tube8".to_string(),
|
||||
name: "Tube8".to_string(),
|
||||
description: "Tube8 Videos".to_string(),
|
||||
premium: false,
|
||||
favicon: "https://www.google.com/s2/favicons?sz=64&domain=www.tube8.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(),
|
||||
systemImage: "list.number".to_string(),
|
||||
colorName: "blue".to_string(),
|
||||
options: vec![
|
||||
FilterOption {
|
||||
id: "rating".into(),
|
||||
title: "Rating".into(),
|
||||
},
|
||||
FilterOption {
|
||||
id: "mostviewed".into(),
|
||||
title: "Most Viewed".into(),
|
||||
},
|
||||
FilterOption {
|
||||
id: "longest".into(),
|
||||
title: "Duration".into(),
|
||||
},
|
||||
FilterOption {
|
||||
id: "newest".into(),
|
||||
title: "Newest".into(),
|
||||
},
|
||||
],
|
||||
multiSelect: false,
|
||||
},
|
||||
ChannelOption {
|
||||
id: "sites".to_string(),
|
||||
title: "Sites".to_string(),
|
||||
description: "Filter for different Sites".to_string(),
|
||||
systemImage: "rectangle.stack".to_string(),
|
||||
colorName: "green".to_string(),
|
||||
options: sites,
|
||||
multiSelect: false,
|
||||
},
|
||||
ChannelOption {
|
||||
id: "stars".to_string(),
|
||||
title: "Stars".to_string(),
|
||||
description: "Filter for different Pornstars".to_string(),
|
||||
systemImage: "star.fill".to_string(),
|
||||
colorName: "yellow".to_string(),
|
||||
options: stars,
|
||||
multiSelect: false,
|
||||
},
|
||||
],
|
||||
nsfw: true,
|
||||
cacheDuration: None,
|
||||
}
|
||||
}
|
||||
|
||||
async fn get(
|
||||
&self,
|
||||
cache: VideoCache,
|
||||
page: u8,
|
||||
sort: &str,
|
||||
options: ServerOptions,
|
||||
) -> Result<Vec<VideoItem>> {
|
||||
let mut sort_string: String = match sort {
|
||||
"mostviewed" => "most-viewed/page/".to_string(),
|
||||
"longest" => "longest/page/".to_string(),
|
||||
"newest" => "newest/page/".to_string(),
|
||||
_ => "top/page/".to_string(),
|
||||
};
|
||||
if options.sites.is_some()
|
||||
&& !options.sites.as_ref().unwrap().is_empty()
|
||||
&& options.sites.as_ref().unwrap() != "all"
|
||||
{
|
||||
sort_string = match sort {
|
||||
"mostviewed.html" => "?orderBy=mv&page=".to_string(),
|
||||
"longest.html" => "?orderBy=ln&page=".to_string(),
|
||||
"newest.html" => "?page=".to_string(),
|
||||
_ => "?orderBy=tr&page=".to_string(),
|
||||
};
|
||||
}
|
||||
if options.stars.is_some()
|
||||
&& !options.stars.as_ref().unwrap().is_empty()
|
||||
&& options.stars.as_ref().unwrap() != "all"
|
||||
{
|
||||
sort_string = match sort {
|
||||
"mostviewed.html" => "views/?page=".to_string(),
|
||||
"longest.html" => "duration/?page=".to_string(),
|
||||
"newest.html" => "?page=".to_string(),
|
||||
_ => "rating/?page=".to_string(),
|
||||
};
|
||||
}
|
||||
let video_url = format!("{}/{}{}", self.url, sort_string, page);
|
||||
let old_items = match cache.get(&video_url) {
|
||||
Some((time, items)) => {
|
||||
if time.elapsed().unwrap_or_default().as_secs() < 60 * 5 {
|
||||
return Ok(items.clone());
|
||||
} else {
|
||||
items.clone()
|
||||
}
|
||||
}
|
||||
None => {
|
||||
vec![]
|
||||
}
|
||||
};
|
||||
println!("Video URL {:?}", video_url);
|
||||
let mut requester = options.requester.clone().unwrap();
|
||||
let text = requester.get(&video_url, None).await.unwrap();
|
||||
let video_items: Vec<VideoItem> = self.get_video_items_from_html(text.clone());
|
||||
if !video_items.is_empty() {
|
||||
cache.remove(&video_url);
|
||||
cache.insert(video_url.clone(), video_items.clone());
|
||||
} else {
|
||||
return Ok(old_items);
|
||||
}
|
||||
Ok(video_items)
|
||||
}
|
||||
|
||||
async fn query(
|
||||
&self,
|
||||
cache: VideoCache,
|
||||
page: u8,
|
||||
query: &str,
|
||||
options: ServerOptions,
|
||||
) -> Result<Vec<VideoItem>> {
|
||||
let mut sort_string: String = match options.sort.as_ref().unwrap().as_str() {
|
||||
"mostviewed.html" => "&orderby=views&page=".to_string(),
|
||||
"longest.html" => "&orderby=longest&page=".to_string(),
|
||||
"newest.html" => "&orderby=newest&page=".to_string(),
|
||||
_ => "&orderby=rating&page=".to_string(),
|
||||
};
|
||||
let mut search_string = query.to_string().to_ascii_lowercase().trim().to_string();
|
||||
let mut video_url = format!(
|
||||
"{}/searches.html/?q={}{}{}",
|
||||
self.url, query, sort_string, page
|
||||
);
|
||||
video_url = video_url.replace(" ", "+");
|
||||
match self
|
||||
.stars
|
||||
.read()
|
||||
.unwrap()
|
||||
.iter()
|
||||
.find(|s| s.title.to_ascii_lowercase() == search_string)
|
||||
{
|
||||
Some(star) => {
|
||||
sort_string = match options.sort.as_ref().unwrap().as_str() {
|
||||
"mostviewed.html" => "views/?page=".to_string(),
|
||||
"longest.html" => "duration/?page=".to_string(),
|
||||
"newest.html" => "?page=".to_string(),
|
||||
_ => "rating/?page=".to_string(),
|
||||
};
|
||||
video_url = format!("{}/{}{}{}", self.url, star.id, sort_string, page);
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
match self
|
||||
.sites
|
||||
.read()
|
||||
.unwrap()
|
||||
.iter()
|
||||
.find(|s| s.title.to_ascii_lowercase() == search_string)
|
||||
{
|
||||
Some(site) => {
|
||||
sort_string = match options.sort.as_ref().unwrap().as_str() {
|
||||
"mostviewed.html" => "?orderBy=mv&page=".to_string(),
|
||||
"longest.html" => "?orderBy=ln&page=".to_string(),
|
||||
"newest.html" => "?page=".to_string(),
|
||||
_ => "?orderBy=tr&page=".to_string(),
|
||||
};
|
||||
video_url = format!("{}/{}{}{}", self.url, site.id, sort_string, page);
|
||||
}
|
||||
_ => {}
|
||||
}
|
||||
// Check our Video Cache. If the result is younger than 1 hour, we return it.
|
||||
let old_items = match cache.get(&video_url) {
|
||||
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 mut requester = options.requester.clone().unwrap();
|
||||
|
||||
let text = requester.get(&video_url, None).await.unwrap();
|
||||
let video_items: Vec<VideoItem> = self.get_video_items_from_html(text.clone());
|
||||
if !video_items.is_empty() {
|
||||
cache.remove(&video_url);
|
||||
cache.insert(video_url.clone(), video_items.clone());
|
||||
} else {
|
||||
return Ok(old_items);
|
||||
}
|
||||
Ok(video_items)
|
||||
}
|
||||
|
||||
fn get_video_items_from_html(&self, html: String) -> Vec<VideoItem> {
|
||||
if html.is_empty() {
|
||||
println!("HTML is empty");
|
||||
return vec![];
|
||||
}
|
||||
let mut items: Vec<VideoItem> = Vec::new();
|
||||
if !html.contains("video-box ") {
|
||||
return items;
|
||||
}
|
||||
let raw_videos = html.split("id=\"pagination\"").collect::<Vec<&str>>()[0]
|
||||
.split("-thumbs")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("\"video-box ")
|
||||
.collect::<Vec<&str>>()[1..]
|
||||
.to_vec();
|
||||
for video_segment in &raw_videos {
|
||||
// let vid = video_segment.split("\n").collect::<Vec<&str>>();
|
||||
// for (index, line) in vid.iter().enumerate() {
|
||||
// println!("Line {}: {}", index, line);
|
||||
// }
|
||||
if video_segment.contains("adsbytrafficjunky"){
|
||||
continue;
|
||||
}
|
||||
let video_url: String = format!("{}{}", self.url, video_segment.split("<a href=\"").collect::<Vec<&str>>()[1]
|
||||
.split("\"")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.to_string());
|
||||
let mut title = video_segment.split("alt=\"").collect::<Vec<&str>>()[1]
|
||||
.split("\"")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.to_string();
|
||||
// html decode
|
||||
title = decode(title.as_bytes()).to_string().unwrap_or(title);
|
||||
let id = video_url.split("/").collect::<Vec<&str>>()[4].to_string();
|
||||
|
||||
let thumb = match video_segment.split("thumb-image ").collect::<Vec<&str>>()[1]
|
||||
.contains("data-src=\"")
|
||||
{
|
||||
true => video_segment.split("thumb-image ").collect::<Vec<&str>>()[1]
|
||||
.split("data-src=\"")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("\"")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.to_string(),
|
||||
false => video_segment.split("thumb-image ").collect::<Vec<&str>>()[1]
|
||||
.split("data-poster=\"")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("\"")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.to_string(),
|
||||
};
|
||||
let raw_duration = video_segment
|
||||
.split("video-duration ")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("</span>")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.split("<span>")
|
||||
.collect::<Vec<&str>>()
|
||||
.last()
|
||||
.unwrap_or(&"")
|
||||
.replace("\n", "")
|
||||
.trim()
|
||||
.to_string();
|
||||
let duration = parse_time_to_seconds(raw_duration.as_str()).unwrap_or(0) as u32;
|
||||
let views = parse_abbreviated_number(
|
||||
video_segment
|
||||
.split("<span class='info-views'>")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("<")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.to_string()
|
||||
.as_str(),
|
||||
)
|
||||
.unwrap_or(0) as u32;
|
||||
let mut tags = match video_segment.contains("info-views-container block") {
|
||||
true => video_segment
|
||||
.split("info-views-container block")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("view-rating-container")
|
||||
.collect::<Vec<&str>>()[0]
|
||||
.split("<a ")
|
||||
.collect::<Vec<&str>>()[1..]
|
||||
.into_iter()
|
||||
.map(|s| {
|
||||
let mut target = &self.stars;
|
||||
if s.contains("author-title-text "){
|
||||
target = &self.sites
|
||||
}
|
||||
let id = s.split("href=\"")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("\"")
|
||||
.collect::<Vec<&str>>()[0];
|
||||
let title = s.split(">")
|
||||
.collect::<Vec<&str>>()[1]
|
||||
.split("</a")
|
||||
.collect::<Vec<&str>>()[0];
|
||||
Self::push_unique(
|
||||
target,
|
||||
FilterOption {
|
||||
id: id.to_string(),
|
||||
title: title.to_string(),
|
||||
},
|
||||
);
|
||||
title.to_string()
|
||||
})
|
||||
.collect::<Vec<String>>()
|
||||
.to_vec(),
|
||||
false => vec![],
|
||||
};
|
||||
let video_item = VideoItem::new(
|
||||
id,
|
||||
title,
|
||||
video_url.to_string(),
|
||||
"tube8".to_string(),
|
||||
thumb,
|
||||
duration,
|
||||
)
|
||||
.views(views)
|
||||
.tags(tags);
|
||||
items.push(video_item);
|
||||
}
|
||||
return items;
|
||||
}
|
||||
}
|
||||
|
||||
#[async_trait]
|
||||
impl Provider for Tube8Provider {
|
||||
async fn get_videos(
|
||||
&self,
|
||||
cache: VideoCache,
|
||||
pool: DbPool,
|
||||
sort: String,
|
||||
query: Option<String>,
|
||||
page: String,
|
||||
per_page: String,
|
||||
options: ServerOptions,
|
||||
) -> Vec<VideoItem> {
|
||||
let _ = per_page;
|
||||
let _ = pool;
|
||||
let videos: std::result::Result<Vec<VideoItem>, Error> = match query {
|
||||
Some(q) => {
|
||||
self.query(cache, page.parse::<u8>().unwrap_or(1), &q, options)
|
||||
.await
|
||||
}
|
||||
None => {
|
||||
self.get(cache, page.parse::<u8>().unwrap_or(1), &sort, options)
|
||||
.await
|
||||
}
|
||||
};
|
||||
match videos {
|
||||
Ok(v) => v,
|
||||
Err(e) => {
|
||||
println!("Error fetching videos: {}", e);
|
||||
vec![]
|
||||
}
|
||||
}
|
||||
}
|
||||
fn get_channel(&self, clientversion: ClientVersion) -> Option<crate::status::Channel> {
|
||||
Some(self.build_channel(clientversion))
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user