Files
hottub/src/providers/noodlemagazine.rs
2026-01-13 18:13:51 +00:00

309 lines
8.4 KiB
Rust

use crate::DbPool;
use crate::api::ClientVersion;
use crate::providers::Provider;
use crate::status::*;
use crate::util::cache::VideoCache;
use crate::util::requester::Requester;
use crate::util::parse_abbreviated_number;
use crate::util::time::parse_time_to_seconds;
use crate::videos::{ServerOptions, VideoFormat, VideoItem};
use async_trait::async_trait;
use error_chain::error_chain;
use futures::future::join_all;
use htmlentity::entity::{decode, ICodedDataTrait};
use wreq::Version;
use titlecase::Titlecase;
use std::vec;
error_chain! {
foreign_links {
Io(std::io::Error);
HttpRequest(wreq::Error);
}
errors {
Parse(msg: String)
}
}
#[derive(Debug, Clone)]
pub struct NoodlemagazineProvider {
url: String,
}
impl NoodlemagazineProvider {
pub fn new() -> Self {
Self {
url: "https://noodlemagazine.com".to_string(),
}
}
fn build_channel(&self, _clientversion: ClientVersion) -> Channel {
Channel {
id: "noodlemagazine".into(),
name: "Noodlemagazine".into(),
description: "The Best Search Engine of HD Videos".into(),
premium: false,
favicon: "https://www.google.com/s2/favicons?sz=64&domain=noodlemagazine.com".into(),
status: "active".into(),
categories: vec![],
options: vec![],
nsfw: true,
cacheDuration: Some(1800),
}
}
async fn get(
&self,
cache: VideoCache,
page: u8,
_sort: &str,
options: ServerOptions,
) -> Result<Vec<VideoItem>> {
let video_url = format!(
"{}/popular/recent?sort_by=views&sort_order=desc&p={}",
self.url,
page.saturating_sub(1)
);
let old_items = match cache.get(&video_url) {
Some((t, i)) if t.elapsed().unwrap_or_default().as_secs() < 300 => return Ok(i.clone()),
Some((_, i)) => i.clone(),
None => vec![],
};
let mut requester = match options.requester.clone() {
Some(r) => r,
None => return Ok(old_items),
};
let text = requester
.get(&video_url, Some(Version::HTTP_2))
.await
.unwrap_or_default();
let items = self.get_video_items_from_html(text, requester).await;
if items.is_empty() {
Ok(old_items)
} else {
cache.remove(&video_url);
cache.insert(video_url, items.clone());
Ok(items)
}
}
async fn query(
&self,
cache: VideoCache,
page: u8,
query: &str,
options: ServerOptions,
) -> Result<Vec<VideoItem>> {
let q = query.trim().replace(' ', "%20");
let video_url = format!("{}/video/{}?p={}", self.url, q, page.saturating_sub(1));
let old_items = match cache.get(&video_url) {
Some((t, i)) if t.elapsed().unwrap_or_default().as_secs() < 300 => return Ok(i.clone()),
Some((_, i)) => i.clone(),
None => vec![],
};
let mut requester = match options.requester.clone() {
Some(r) => r,
None => return Ok(old_items),
};
let text = requester
.get(&video_url, Some(Version::HTTP_2))
.await
.unwrap_or_default();
let items = self.get_video_items_from_html(text, requester).await;
if items.is_empty() {
Ok(old_items)
} else {
cache.remove(&video_url);
cache.insert(video_url, items.clone());
Ok(items)
}
}
async fn get_video_items_from_html(
&self,
html: String,
requester: Requester,
) -> Vec<VideoItem> {
if html.is_empty() || html.contains("404 Not Found") {
return vec![];
}
let section = match html.split(">Show more</div>").next() {
Some(s) => s,
None => return vec![],
};
let list = match section
.split("<div class=\"list_videos\" id=\"list_videos\">")
.nth(1)
{
Some(l) => l,
None => return vec![],
};
let raw_videos = list
.split("<div class=\"item\">")
.skip(1)
.map(|s| s.to_string());
let futures = raw_videos.map(|v| self.get_video_item(v, requester.clone()));
let results = join_all(futures).await;
results.into_iter().filter_map(Result::ok).collect()
}
async fn get_video_item(
&self,
video_segment: String,
requester: Requester,
) -> Result<VideoItem> {
let href = video_segment
.split("<a href=\"")
.nth(1)
.and_then(|s| s.split('"').next())
.ok_or_else(|| Error::from("missing href"))?;
let video_url = format!("{}{}", self.url, href);
let mut title = video_segment
.split("<div class=\"title\">")
.nth(1)
.and_then(|s| s.split('<').next())
.unwrap_or("")
.trim()
.to_string();
title = decode(title.as_bytes())
.to_string()
.unwrap_or(title)
.titlecase();
let id = video_url
.split('/')
.nth(4)
.and_then(|s| s.split('.').next())
.ok_or_else(|| Error::from("missing id"))?
.to_string();
let thumb = video_segment
.split("data-src=\"")
.nth(1)
.and_then(|s| s.split('"').next())
.unwrap_or("")
.to_string();
let raw_duration = video_segment
.split("#clock-o\"></use></svg>")
.nth(1)
.and_then(|s| s.split('<').next())
.unwrap_or("0:00");
let duration = parse_time_to_seconds(raw_duration).unwrap_or(0) as u32;
let views = video_segment
.split("#eye\"></use></svg>")
.nth(1)
.and_then(|s| s.split('<').next())
.and_then(|v| parse_abbreviated_number(v.trim()))
.unwrap_or(0);
let formats = self
.extract_media(&video_url, requester)
.await
.ok_or_else(|| Error::from("media extraction failed"))?;
Ok(
VideoItem::new(
id,
title,
video_url,
"noodlemagazine".into(),
thumb,
duration,
)
.views(views)
.formats(formats),
)
}
async fn extract_media(
&self,
video_url: &String,
mut requester: Requester,
) -> Option<Vec<VideoFormat>> {
let text = requester
.get(video_url, Some(Version::HTTP_2))
.await
.unwrap_or_default();
let json_str = text
.split("window.playlist = ")
.nth(1)?
.split(';')
.next()?;
let json: serde_json::Value = serde_json::from_str(json_str).ok()?;
let sources = json["sources"].as_array()?;
let mut formats = vec![];
for s in sources {
let file = s["file"].as_str()?.to_string();
let label = s["label"].as_str().unwrap_or("unknown").to_string();
formats.push(
VideoFormat::new(file, label.clone(), "video/mp4".into())
.format_id(label.clone())
.format_note(label.clone())
.http_header("Referer".into(), video_url.clone()),
);
}
Some(formats.into_iter().rev().collect())
}
}
#[async_trait]
impl Provider for NoodlemagazineProvider {
async fn get_videos(
&self,
cache: VideoCache,
pool: DbPool,
sort: String,
query: Option<String>,
page: String,
per_page: String,
options: ServerOptions,
) -> Vec<VideoItem> {
let _ = pool;
let _ = per_page;
let page = page.parse::<u8>().unwrap_or(1);
let res = match query {
Some(q) => self.query(cache, page, &q, options).await,
None => self.get(cache, page, &sort, options).await,
};
res.unwrap_or_else(|e| {
eprintln!("Noodlemagazine error: {e}");
vec![]
})
}
fn get_channel(&self, clientversion: ClientVersion) -> Option<Channel> {
Some(self.build_channel(clientversion))
}
}