✨ Option to use a proxy for outgoing `upstream search engine` requests (#573)

* Add option to use a proxy for outgoing (search engine) requests.

* Enable socks feature in reqwest

* Fix formatting

* add proxy feature

* Update src/config/parser.rs

Co-authored-by: neon_arch <[email protected]>

* Update websurfx/config.lua

Co-authored-by: neon_arch <[email protected]>

* Update Cargo.toml

Co-authored-by: neon_arch <[email protected]>

* fix

* Update Cargo.toml

Co-authored-by: neon_arch <[email protected]>

---------

Co-authored-by: neon_arch <[email protected]>

Cargo.toml

@@ -17,7 +17,8 @@ reqwest = { version = "0.12.5", default-features = false, features = [
     "rustls-tls",
     "brotli",
     "gzip",
-    "http2"
+    "http2",
+    "socks",
 ] }
 tokio = { version = "1.32.0", features = [
     "rt-multi-thread",

src/config/parser.rs

@@ -6,6 +6,7 @@ use crate::handler::{file_path, FileType};
 use crate::models::parser_models::{AggregatorConfig, RateLimiter, Style};
 use log::LevelFilter;
 use mlua::Lua;
+use reqwest::Proxy;
 use std::{collections::HashMap, fs, thread::available_parallelism};
 
 /// A named struct which stores the parsed config file options.
@@ -48,6 +49,8 @@ pub struct Config {
     pub tcp_connection_keep_alive: u8,
     /// It stores the pool idle connection timeout in seconds.
     pub pool_idle_connection_timeout: u8,
+    /// Url of the proxy to use for outgoing requests.
+    pub proxy: Option<Proxy>,
     /// It stores the number of https connections to keep in the pool.
     pub number_of_https_connections: u8,
 }
@@ -120,6 +123,14 @@ impl Config {
             _ => parsed_cet,
         };
 
+        let proxy_opt = globals.get::<_, Option<String>>("proxy")?;
+        let proxy = proxy_opt.and_then(|proxy_str| {
+            Proxy::all(proxy_str).ok().and_then(|_| {
+                log::error!("Invalid proxy url, defaulting to no proxy.");
+                None
+            })
+        });
+
         Ok(Config {
             port: globals.get::<_, u16>("port")?,
             binding_ip: globals.get::<_, String>("binding_ip")?,
@@ -151,6 +162,7 @@ impl Config {
             safe_search,
             #[cfg(any(feature = "redis-cache", feature = "memory-cache"))]
             cache_expiry_time,
+            proxy,
         })
     }
 }

src/results/aggregator.rs

@@ -75,7 +75,7 @@ pub async fn aggregate(
     safe_search: u8,
 ) -> Result<SearchResults, Box<dyn std::error::Error>> {
     let client = CLIENT.get_or_init(|| {
-        ClientBuilder::new()
+        let mut cb = ClientBuilder::new()
             .timeout(Duration::from_secs(config.request_timeout as u64)) // Add timeout to request to avoid DDOSing the server
             .pool_idle_timeout(Duration::from_secs(
                 config.pool_idle_connection_timeout as u64,
@@ -86,9 +86,13 @@ pub async fn aggregate(
             .https_only(true)
             .gzip(true)
             .brotli(true)
-            .http2_adaptive_window(config.adaptive_window)
-            .build()
-            .unwrap()
+            .http2_adaptive_window(config.adaptive_window);
+
+        if config.proxy.is_some() {
+            cb = cb.proxy(config.proxy.clone().unwrap());
+        }
+
+        cb.build().unwrap()
     });
 
     let user_agent: &str = random_user_agent();
@@ -242,6 +246,7 @@ pub async fn filter_with_lists(
 
     Ok(())
 }
+
 /// Sorts  SearchResults by relevance score.
 /// <br> sort_unstable is used as its faster,stability is not an issue on our side.
 /// For reasons why, check out [`this`](https://rust-lang.github.io/rfcs/1884-unstable-sort.html)
@@ -257,6 +262,7 @@ fn sort_search_results(results: &mut [SearchResult]) {
             .unwrap_or(Ordering::Less)
     })
 }
+
 #[cfg(test)]
 mod tests {
     use super::*;

websurfx/config.lua

@@ -75,3 +75,5 @@ upstream_search_engines = {
     Mojeek = false,
     Bing = false,
 } -- select the upstream search engines from which the results should be fetched.
+
+proxy = nil -- Proxy to send outgoing requests through. Set to nil to disable.