simon/hottub
0
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

noodlemagazine thumb proxy

Browse Source
This commit is contained in:
Simon
2026-03-20 13:52:06 +00:00
parent 99e4a77507
commit dd7c4ec6a1
3 changed files with 323 additions and 33 deletions
Download Patch File Download Diff File Expand all files Collapse all files

191
src/proxies/noodlemagazine.rs
View File

@@ -1,11 +1,21 @@
use ntex::http::header::CONTENT_TYPE;
use ntex::web::{self, HttpRequest, error};
use ntex::http::header::{CONTENT_LENGTH, CONTENT_TYPE};
use ntex::{
http::Response,
web::{self, HttpRequest, error},
};
use serde_json::Value;
use url::Url;
use wreq::Version;
use crate::util::requester::Requester;
const FIREFOX_USER_AGENT: &str =
"Mozilla/5.0 (X11; Linux x86_64; rv:147.0) Gecko/20100101 Firefox/147.0";
const HTML_ACCEPT: &str =
"text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,*/*;q=0.8";
const IMAGE_ACCEPT: &str =
"image/avif,image/webp,image/png,image/svg+xml,image/*;q=0.8,*/*;q=0.5";
#[derive(Debug, Clone)]
pub struct NoodlemagazineProxy {}
@@ -57,6 +67,75 @@ impl NoodlemagazineProxy {
}
}
fn normalize_image_url(url: &str) -> String {
if url.starts_with("http://") || url.starts_with("https://") {
url.to_string()
} else {
format!("https://{}", url.trim_start_matches('/'))
}
}
fn root_referer() -> &'static str {
"https://noodlemagazine.com/"
}
fn root_html_headers() -> Vec<(String, String)> {
vec![
("Referer".to_string(), Self::root_referer().to_string()),
("User-Agent".to_string(), FIREFOX_USER_AGENT.to_string()),
("Accept".to_string(), HTML_ACCEPT.to_string()),
("Accept-Language".to_string(), "en-US,en;q=0.9".to_string()),
]
}
fn image_headers(requester: &Requester, image_url: &str) -> Vec<(String, String)> {
let mut headers = vec![
("Referer".to_string(), Self::root_referer().to_string()),
("User-Agent".to_string(), FIREFOX_USER_AGENT.to_string()),
("Accept".to_string(), IMAGE_ACCEPT.to_string()),
("Accept-Language".to_string(), "en-US,en;q=0.9".to_string()),
];
if let Some(cookie) = requester.cookie_header_for_url(image_url) {
headers.push(("Cookie".to_string(), cookie));
}
headers
}
fn has_allowed_image_extension(path: &str) -> bool {
let path = path.to_ascii_lowercase();
[".jpg", ".jpeg", ".png", ".webp", ".avif", ".gif"]
.iter()
.any(|ext| path.ends_with(ext))
}
fn is_allowed_thumb_url(url: &str) -> bool {
let Some(url) = Url::parse(url).ok() else {
return false;
};
if url.scheme() != "https" {
return false;
}
let Some(host) = url.host_str() else {
return false;
};
(host == "noodlemagazine.com" || host.ends_with(".noodlemagazine.com"))
&& Self::has_allowed_image_extension(url.path())
}
fn is_binary_image_content_type(content_type: &str) -> bool {
let media_type = content_type
.split(';')
.next()
.unwrap_or_default()
.trim()
.to_ascii_lowercase();
media_type.starts_with("image/")
}
fn is_hls_url(url: &str) -> bool {
Url::parse(url)
.ok()
@@ -184,6 +263,82 @@ pub async fn serve_media(
.body(rewritten_manifest))
}
pub async fn get_image(
req: HttpRequest,
requester: web::types::State<Requester>,
) -> Result<impl web::Responder, web::Error> {
let endpoint = req.match_info().query("endpoint").to_string();
let image_url = NoodlemagazineProxy::normalize_image_url(&endpoint);
if !NoodlemagazineProxy::is_allowed_thumb_url(&image_url) {
return Ok(web::HttpResponse::BadRequest().finish());
}
let mut requester = requester.get_ref().clone();
let _ = requester
.get_with_headers(
NoodlemagazineProxy::root_referer(),
NoodlemagazineProxy::root_html_headers(),
Some(Version::HTTP_11),
)
.await;
let mut headers = NoodlemagazineProxy::image_headers(&requester, image_url.as_str());
let mut upstream = requester
.get_raw_with_headers(image_url.as_str(), headers.clone())
.await
.ok();
let needs_warmup = upstream
.as_ref()
.map(|response| !response.status().is_success())
.unwrap_or(true);
if needs_warmup {
let _ = requester
.get_with_headers(
image_url.as_str(),
headers.clone(),
Some(Version::HTTP_11),
)
.await;
headers = NoodlemagazineProxy::image_headers(&requester, image_url.as_str());
upstream = requester
.get_raw_with_headers(image_url.as_str(), headers)
.await
.ok();
}
let Some(upstream) = upstream.filter(|response| response.status().is_success()) else {
return Ok(web::HttpResponse::NotFound().finish());
};
let status = upstream.status();
let headers = upstream.headers().clone();
let content_type = headers
.get(CONTENT_TYPE)
.and_then(|value| value.to_str().ok())
.map(str::to_string)
.unwrap_or_default();
if !NoodlemagazineProxy::is_binary_image_content_type(&content_type) {
return Ok(web::HttpResponse::BadGateway().finish());
}
let bytes = upstream.bytes().await.map_err(error::ErrorBadGateway)?;
let mut resp = Response::build(status);
if !content_type.is_empty() {
resp.set_header(CONTENT_TYPE, content_type);
}
if let Some(cl) = headers.get(CONTENT_LENGTH) {
if let Ok(cl_str) = cl.to_str() {
resp.set_header(CONTENT_LENGTH, cl_str);
}
}
Ok(resp.body(bytes.to_vec()))
}
#[cfg(test)]
mod tests {
use super::NoodlemagazineProxy;
@@ -231,4 +386,36 @@ mod tests {
"#EXTM3U\n#EXT-X-STREAM-INF:BANDWIDTH=1\nhttps://cdn.example/hls/low/index.m3u8\n#EXT-X-KEY:METHOD=AES-128,URI=\"https://cdn.example/hls/keys/key.bin\"\nhttps://cdn.example/hls/segment0.ts"
);
}
#[test]
fn allows_only_noodlemagazine_image_thumbs() {
assert!(NoodlemagazineProxy::is_allowed_thumb_url(
"https://noodlemagazine.com/thumbs/example.webp"
));
assert!(NoodlemagazineProxy::is_allowed_thumb_url(
"https://img.noodlemagazine.com/previews/example.jpg"
));
assert!(!NoodlemagazineProxy::is_allowed_thumb_url(
"https://noodlemagazine.com/watch/-123_456"
));
assert!(!NoodlemagazineProxy::is_allowed_thumb_url(
"https://cdn.example/thumb.jpg"
));
}
#[test]
fn recognizes_binary_image_content_types() {
assert!(NoodlemagazineProxy::is_binary_image_content_type(
"image/webp"
));
assert!(NoodlemagazineProxy::is_binary_image_content_type(
"image/jpeg; charset=binary"
));
assert!(!NoodlemagazineProxy::is_binary_image_content_type(
"text/html; charset=utf-8"
));
assert!(!NoodlemagazineProxy::is_binary_image_content_type(
"application/json"
));
}
}