more fixes

This commit is contained in:
Simon
2026-01-02 14:58:29 +00:00
parent 89eecbe790
commit 262b908692

View File

@@ -7,43 +7,45 @@ 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::{ICodedDataTrait, decode};
use htmlentity::entity::{decode, ICodedDataTrait};
use wreq::Version;
use std::{vec};
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 {
let provider = NoodlemagazineProvider {
Self {
url: "https://noodlemagazine.com".to_string(),
};
provider
}
}
fn build_channel(&self, clientversion: ClientVersion) -> Channel {
let _ = clientversion;
fn build_channel(&self, _clientversion: ClientVersion) -> Channel {
Channel {
id: "noodlemagazine".to_string(),
name: "Noodlemagazine".to_string(),
description: "The Best Search Engine of HD Videos".to_string(),
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"
.to_string(),
status: "active".to_string(),
favicon: "https://www.google.com/s2/favicons?sz=64&domain=noodlemagazine.com".into(),
status: "active".into(),
categories: vec![],
options: vec![],
nsfw: true,
@@ -55,40 +57,40 @@ impl NoodlemagazineProvider {
&self,
cache: VideoCache,
page: u8,
sort: &str,
_sort: &str,
options: ServerOptions,
) -> Result<Vec<VideoItem>> {
let _ = sort;
let video_url = format!(
"{}/popular/recent?sort_by=views&sort_order=desc&p={}",
self.url,
page - 1
page.saturating_sub(1)
);
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![]
}
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 = options.requester.clone().unwrap();
let text = requester.get(&video_url, Some(Version::HTTP_2)).await.unwrap();
let video_items: Vec<VideoItem> = self
.get_video_items_from_html(text.clone(), requester.clone())
.await;
if !video_items.is_empty() {
cache.remove(&video_url);
cache.insert(video_url.clone(), video_items.clone());
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 {
return Ok(old_items);
cache.remove(&video_url);
cache.insert(video_url, items.clone());
Ok(items)
}
Ok(video_items)
}
async fn query(
@@ -98,37 +100,34 @@ impl NoodlemagazineProvider {
query: &str,
options: ServerOptions,
) -> Result<Vec<VideoItem>> {
let search_string = query.trim().replace(" ", "%20").to_string();
let q = query.trim().replace(' ', "%20");
let video_url = format!("{}/video/{}?p={}", self.url, q, page.saturating_sub(1));
let video_url = format!("{}/video/{}?p={}", self.url, search_string, page - 1);
// 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![]
}
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 = options.requester.clone().unwrap();
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();
let video_items: Vec<VideoItem> = self
.get_video_items_from_html(text.clone(), requester.clone())
.await;
if !video_items.is_empty() {
cache.remove(&video_url);
cache.insert(video_url.clone(), video_items.clone());
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 {
return Ok(old_items);
cache.remove(&video_url);
cache.insert(video_url, items.clone());
Ok(items)
}
Ok(video_items)
}
async fn get_video_items_from_html(
@@ -139,19 +138,29 @@ impl NoodlemagazineProvider {
if html.is_empty() || html.contains("404 Not Found") {
return vec![];
}
let raw_videos = html.split(">Show more</div>").collect::<Vec<&str>>()[0]
.split("<div class=\"list_videos\" id=\"list_videos\">")
.collect::<Vec<&str>>()[1]
.split("<div class=\"item\">")
.collect::<Vec<&str>>()[1..]
.to_vec();
let futures = raw_videos
.into_iter()
.map(|el| self.get_video_item(el.to_string(), requester.clone()));
let results: Vec<Result<VideoItem>> = join_all(futures).await;
let video_items: Vec<VideoItem> = results.into_iter().filter_map(Result::ok).collect();
return video_items;
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(
@@ -159,97 +168,109 @@ impl NoodlemagazineProvider {
video_segment: String,
requester: Requester,
) -> Result<VideoItem> {
let video_url: String = format!(
"{}{}",
self.url,
video_segment.split("<a href=\"").collect::<Vec<&str>>()[1]
.split("\"")
.collect::<Vec<&str>>()[0]
.to_string()
);
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\">")
.collect::<Vec<&str>>()[1]
.split("<")
.collect::<Vec<&str>>()[0]
.nth(1)
.and_then(|s| s.split('<').next())
.unwrap_or("")
.trim()
.to_string();
// html decode
title = decode(title.as_bytes())
.to_string()
.unwrap_or(title)
.titlecase();
let id = video_url.split("/").collect::<Vec<&str>>()[4]
.split(".")
.collect::<Vec<&str>>()[0]
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("<img ").collect::<Vec<&str>>()[1]
let thumb = video_segment
.split("data-src=\"")
.collect::<Vec<&str>>()[1]
.split("\"")
.collect::<Vec<&str>>()[0]
.nth(1)
.and_then(|s| s.split('"').next())
.unwrap_or("")
.to_string();
let raw_duration = video_segment
.split("#clock-o\"></use></svg>")
.collect::<Vec<&str>>()[1]
.split("<")
.collect::<Vec<&str>>()[0]
.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("<use xlink:href=\"#eye\"></use></svg>")
.collect::<Vec<&str>>()[1]
.split("<")
.collect::<Vec<&str>>()[0]
.trim()).unwrap_or(0);
.nth(1)
.and_then(|s| s.split('<').next())
.unwrap_or("0:00");
let formats = match self.extract_media(&video_url, requester.clone()).await{
Some(f) => f,
None => return Err(Error::from("Failed to extract media formats")),
};
let video_item = VideoItem::new(
id,
title,
video_url,
"noodlemagazine".to_string(),
thumb,
duration,
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),
)
.views(views)
.formats(formats)
;
return Ok(video_item);
}
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("".to_string());
if text.is_empty() {
return None;
}
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![];
let json_str = text.split("window.playlist = ")
.collect::<Vec<&str>>()[1]
.split(";")
.collect::<Vec<&str>>()[0];
let json: serde_json::Value = serde_json::from_str(json_str).unwrap();
let sources = json["sources"].as_array().unwrap();
for source in sources {
let file = source["file"].as_str().unwrap().to_string();
let label = source["label"].as_str().unwrap_or("unknown").to_string();
let format = VideoFormat::new(
file,
label.clone(),
"video/mp4".to_string(),
)
.format_id(label.clone())
.format_note(label.clone())
.http_header("Referer".to_string(), video_url.clone())
;
formats.push(format);
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()),
);
}
return Some(formats.into_iter().rev().collect());
Some(formats.into_iter().rev().collect())
}
}
@@ -265,27 +286,23 @@ impl Provider for NoodlemagazineProvider {
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
}
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,
};
match videos {
Ok(v) => v,
Err(e) => {
println!("Error fetching videos: {}", e);
vec![]
}
}
res.unwrap_or_else(|e| {
eprintln!("Noodlemagazine error: {e}");
vec![]
})
}
fn get_channel(&self, clientversion: ClientVersion) -> crate::status::Channel {
fn get_channel(&self, clientversion: ClientVersion) -> Channel {
self.build_channel(clientversion)
}
}