From 78a87f207c9b4cddf21306ed096e00725b0c8bd5 Mon Sep 17 00:00:00 2001
From: Myzel394 <50424412+Myzel394@users.noreply.github.com>
Date: Sat, 3 Feb 2024 21:35:50 +0100
Subject: [PATCH] feat: Add first wip for http server

---
 Cargo.lock                 |  27 ++++++
 Cargo.toml                 |   4 +-
 src/engines/duckduckgo.rs  | 173 ++++++++++++++++++++++++------------
 src/engines/engine_base.rs |  15 +++-
 src/main.rs                | 177 ++++++++++++++++++++++++-------------
 5 files changed, 273 insertions(+), 123 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 36377d6..729a834 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -310,6 +310,7 @@ checksum = "645c6916888f6cb6350d2550b80fb63e734897a8498abe35cfb732b6487804b0"
 dependencies = [
  "futures-channel",
  "futures-core",
+ "futures-executor",
  "futures-io",
  "futures-sink",
  "futures-task",
@@ -332,12 +333,34 @@ version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "dfc6580bb841c5a68e9ef15c77ccc837b40a7504914d52e47b8b0e9bbda25a1d"
 
+[[package]]
+name = "futures-executor"
+version = "0.3.30"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a576fc72ae164fca6b9db127eaa9a9dda0d61316034f33a0a0d4eda41f02b01d"
+dependencies = [
+ "futures-core",
+ "futures-task",
+ "futures-util",
+]
+
 [[package]]
 name = "futures-io"
 version = "0.3.30"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a44623e20b9681a318efdd71c299b6b222ed6f231972bfe2f224ebad6311f0c1"
 
+[[package]]
+name = "futures-macro"
+version = "0.3.30"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "87750cf4b7a4c0625b1529e4c543c2182106e4dedc60a2a6455e00d212c489ac"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
 [[package]]
 name = "futures-sink"
 version = "0.3.30"
@@ -359,6 +382,7 @@ dependencies = [
  "futures-channel",
  "futures-core",
  "futures-io",
+ "futures-macro",
  "futures-sink",
  "futures-task",
  "memchr",
@@ -1416,6 +1440,8 @@ dependencies = [
 name = "tcp_test"
 version = "0.1.0"
 dependencies = [
+ "async-trait",
+ "futures",
  "lazy-regex",
  "lazy_static",
  "regex",
@@ -1505,6 +1531,7 @@ dependencies = [
  "libc",
  "mio",
  "num_cpus",
+ "parking_lot",
  "pin-project-lite",
  "signal-hook-registry",
  "socket2",
diff --git a/Cargo.toml b/Cargo.toml
index 34c315a..3ef482e 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -6,12 +6,14 @@ edition = "2021"
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+async-trait = "0.1.77"
+futures = "0.3.30"
 lazy-regex = "3.1.0"
 lazy_static = "1.4.0"
 regex = "1.10.3"
 reqwest = "0.11.23"
 rocket = "0.5.0"
 rustls = "0.22.2"
-tokio = "1.35.1"
+tokio = {version = "1.35.1", features = ["full"]}
 urlencoding = "2.1.3"
 webpki-roots = "0.26.0"
diff --git a/src/engines/duckduckgo.rs b/src/engines/duckduckgo.rs
index 1023895..99eb11d 100644
--- a/src/engines/duckduckgo.rs
+++ b/src/engines/duckduckgo.rs
@@ -1,46 +1,23 @@
 // Search engine parser for DuckDuckGo
 pub mod duckduckgo {
-    use lazy_regex::regex_replace_all;
-    // Results start at:
-    //     <div id="links" class="results">
-    // Example for a result:
-    //     <div class="result results_links results_links_deep web-result ">
-    //         <div class="links_main links_deep result__body">
-    //             <h2 class="result__title">
-    //                 <a
-    //                     rel="nofollow" class="result__a"
-    //                     href="https://www.speedtest.net/">
-    //                     Speedtest by Ookla - The Global Broadband Speed Test
-    //                 </a>
-    //             </h2>
-    //             <div class="result__extras">
-    //                 <div class="result__extras__url">
-    //                     <span class="result__icon">
-    //                       <a rel="nofollow" href="https://www.speedtest.net/">
-    //                         <img class="result__icon__img" width="16" height="16" alt=""
-    //                           src="//external-content.duckduckgo.com/ip3/www.speedtest.net.ico" name="i15" />
-    //                       </a>
-    //                   </span>
-    //                     <a class="result__url" href="https://www.speedtest.net/">
-    //                         www.speedtest.net
-    //                     </a>
-    //                 </div>
-    //             </div>
-    //             <a
-    //                 class="result__snippet"
-    //                 href="https://www.speedtest.net/">
-    //                     Use Speedtest on all your devices with our free desktop and mobile apps.
-    //             </a>
-    //             <div class="clear"></div>
-    //         </div>
-    //     </div>
+    use std::{
+        cmp::min,
+        collections::VecDeque,
+        pin::Pin,
+        str::Bytes,
+        task::{Context, Poll},
+    };
+
+    use async_trait::async_trait;
+    use futures::Stream;
     use lazy_static::lazy_static;
     use regex::Regex;
     use urlencoding::decode;
 
     use crate::{
-        engines::engine_base::engine_base::{EngineBase, SearchResult},
-        utils::utils::{decode_html_text, replace_html_unicode},
+        client::client::{Client, PACKET_SIZE},
+        engines::engine_base::engine_base::{EngineBase, SearchEngine, SearchResult},
+        utils::utils::decode_html_text,
     };
 
     lazy_static! {
@@ -50,10 +27,15 @@ pub mod duckduckgo {
         static ref STRIP_HTML_TAGS: Regex = Regex::new(r#"<(?:"[^"]*"['"]*|'[^']*'['"]*|[^'">])+>"#).unwrap();
     }
 
+    pub type CallbackType = Box<dyn FnMut(SearchResult) -> () + Send + Sync>;
+
     pub struct DuckDuckGo {
-        pub search_results: Vec<SearchResult>,
+        callback: CallbackType,
+        pub completed: bool,
         results_started: bool,
         previous_block: String,
+        // Holds all results until consumed by iterator
+        pub results: VecDeque<SearchResult>,
     }
 
     impl DuckDuckGo {
@@ -65,14 +47,86 @@ pub mod duckduckgo {
             self.previous_block.clear();
             self.previous_block.push_str(&remaining_text);
         }
-    }
 
-    impl EngineBase for DuckDuckGo {
-        fn get_search_results(&self) -> &Vec<SearchResult> {
-            &self.search_results
+        pub fn new() -> Self {
+            Self {
+                callback: Box::new(|_: SearchResult| {}),
+                results_started: false,
+                previous_block: String::new(),
+                results: VecDeque::new(),
+                completed: false,
+            }
         }
 
-        fn parse_packet<'a>(&mut self, packet: impl Iterator<Item = &'a u8>) -> String {
+        pub fn set_callback(&mut self, callback: CallbackType) {
+            self.callback = callback;
+        }
+    }
+
+    // impl Stream for DuckDuckGo {
+    //     type Item = String;
+    //
+    //     fn poll_next(
+    //         self: Pin<&mut Self>,
+    //         cx: &mut Context<'_>,
+    //     ) -> std::task::Poll<Option<Self::Item>> {
+    //         if self.results.len() > 0 {
+    //             let result = &mut self.results.pop_front().unwrap();
+    //
+    //             let html = format!("<br><h2>{}</h2><p>{}</p>", result.title, result.description);
+    //
+    //             return Poll::Ready(Some(html));
+    //         }
+    //
+    //         if self.completed {
+    //             return Poll::Ready(None);
+    //         }
+    //
+    //         Poll::Pending
+    //     }
+    // }
+
+    // impl Iterator for DuckDuckGo {
+    //     type Item = SearchResult;
+    //
+    //     fn next(&mut self) -> Option<SearchResult> {
+    //         if self.results.len() > 0 {
+    //             let oldest = self.results.pop_front().unwrap();
+    //
+    //             Some(oldest)
+    //         } else {
+    //             None
+    //         }
+    //     }
+    // }
+
+    #[async_trait]
+    impl EngineBase for DuckDuckGo {
+        fn search(&mut self, query: &str) {
+            dbg!("searching duckduckgo");
+
+            let client = Client::new("https://html.duckduckgo.com/html/");
+
+            let packets = client.request(&"POST").unwrap();
+
+            for ii in (0..packets.len()).step_by(PACKET_SIZE) {
+                let end_range = min(packets.len(), ii + PACKET_SIZE);
+
+                let slice = &packets[ii..end_range];
+                self.parse_packet(slice.iter());
+
+                // Call callback, there is probably a better way to do this
+                // while self.results.len() > 0 {
+                //     let result = self.results.pop_front().unwrap();
+                //
+                //     (self.callback)(result);
+                // }
+            }
+
+            self.completed = true;
+        }
+
+        fn parse_packet<'a>(&mut self, packet: impl Iterator<Item = &'a u8>) {
             let bytes: Vec<u8> = packet.map(|bit| *bit).collect();
             let raw_text = String::from_utf8_lossy(&bytes);
             let text = STRIP.replace_all(&raw_text, " ");
@@ -82,33 +136,38 @@ pub mod duckduckgo {
 
                 match SINGLE_RESULT.captures(&self.previous_block.to_owned()) {
                     Some(captures) => {
-                        let title = decode(captures.name("title").unwrap().as_str()).unwrap();
+                        let title = decode(captures.name("title").unwrap().as_str())
+                            .unwrap()
+                            .into_owned();
                         let description_raw =
                             decode_html_text(captures.name("description").unwrap().as_str())
                                 .unwrap();
-                        let description = STRIP_HTML_TAGS.replace_all(&description_raw, "");
-                        let url = decode(captures.name("url").unwrap().as_str()).unwrap();
+                        let description = STRIP_HTML_TAGS
+                            .replace_all(&description_raw, "")
+                            .into_owned();
+                        let url = decode(captures.name("url").unwrap().as_str())
+                            .unwrap()
+                            .into_owned();
+
+                        let result = SearchResult {
+                            title,
+                            description,
+                            url,
+                            engine: SearchEngine::DuckDuckGo,
+                        };
 
                         let end_position = captures.name("end").unwrap().end();
                         self.slice_remaining_block(&end_position);
+
+                        // (self.callback)(result);
+
+                        self.results.push_back(result);
                     }
                     None => {}
                 }
             } else if RESULTS_START.is_match(&text) {
                 self.results_started = true;
             }
-
-            "".to_owned()
-        }
-    }
-
-    impl DuckDuckGo {
-        pub fn new() -> Self {
-            Self {
-                search_results: Vec::new(),
-                results_started: false,
-                previous_block: String::new(),
-            }
         }
     }
 }
diff --git a/src/engines/engine_base.rs b/src/engines/engine_base.rs
index 3523a68..470d7ac 100644
--- a/src/engines/engine_base.rs
+++ b/src/engines/engine_base.rs
@@ -1,13 +1,22 @@
 pub mod engine_base {
+    use async_trait::async_trait;
+
+    #[derive(Clone, Copy, Debug, Hash)]
+    pub enum SearchEngine {
+        DuckDuckGo,
+    }
+
+    #[derive(Clone, Debug, Hash)]
     pub struct SearchResult {
         pub title: String,
         pub url: String,
         pub description: String,
+        pub engine: SearchEngine,
     }
 
+    #[async_trait]
     pub trait EngineBase {
-        fn get_search_results(&self) -> &Vec<SearchResult>;
-
-        fn parse_packet<'a>(&mut self, packet: impl Iterator<Item = &'a u8>) -> String;
+        fn parse_packet<'a>(&mut self, packet: impl Iterator<Item = &'a u8>);
+        fn search(&mut self, query: &str);
     }
 }
diff --git a/src/main.rs b/src/main.rs
index 5cf2475..ddd0e29 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -1,69 +1,122 @@
-use std::cmp::min;
-
-use client::client::{Client, PACKET_SIZE};
-use engines::{duckduckgo::duckduckgo::DuckDuckGo, engine_base::engine_base::EngineBase};
-
-pub mod client;
-pub mod engines;
-pub mod utils;
-
-fn main() {
-    let mut ddg = DuckDuckGo::new();
-    let client = Client::new("https://html.duckduckgo.com/html/");
-
-    let packets = client.request(&"POST").unwrap();
-
-    for ii in (0..packets.len()).step_by(PACKET_SIZE) {
-        let end_range = min(packets.len(), ii + PACKET_SIZE);
-
-        let slice = &packets[ii..end_range];
-        &ddg.parse_packet(slice.iter());
-    }
-}
-
-// use std::cmp::min;
-//
-// use rocket::response::stream::TextStream;
-// use rocket::tokio::time::{self, Duration};
-//
-// use client::client::{Client, PACKET_SIZE};
-// use engines::{duckduckgo::duckduckgo::DuckDuckGo, engine_base::engine_base::EngineBase};
+// use engines::{
+//     duckduckgo::duckduckgo::DuckDuckGo,
+//     engine_base::engine_base::{EngineBase, SearchResult},
+// };
 //
 // pub mod client;
 // pub mod engines;
 // pub mod utils;
 //
-// #[macro_use]
-// extern crate rocket;
+// #[tokio::main]
+// async fn main() {
+//     let callback = Box::new(|result: SearchResult| {
+//         dbg!(&result);
+//     });
+//     let mut ddg = DuckDuckGo::new(callback);
+//     ddg.search(&"test").await;
 //
-// #[get("/")]
-// fn index() -> &'static str {
-//     "Hello, world!"
-// }
-//
-// #[get("/infinite-hellos")]
-// fn hello() -> TextStream![String] {
-//     let mut ddg = DuckDuckGo::new();
-//     let client = Client::new("https://html.duckduckgo.com/html/");
-//
-//     let packets = client.request(&"POST").unwrap();
-//
-//     TextStream! {
-//         let mut interval = time::interval(Duration::from_secs(1));
-//         interval.tick().await;
-//
-//         for ii in (0..packets.len()).step_by(PACKET_SIZE) {
-//             let end_range = min(packets.len(), ii + PACKET_SIZE);
-//
-//             let slice = &packets[ii..end_range];
-//             yield ddg.parse_packet(slice.iter()).to_string();
-//         }
-//     }
-// }
-//
-// #[launch]
-// fn rocket() -> _ {
-//     rocket::build()
-//         .mount("/", routes![index])
-//         .mount("/", routes![hello])
+//     println!("done");
 // }
+
+// Found no other way to make this work
+#![feature(async_closure)]
+
+use std::{
+    cmp::max,
+    sync::{Arc, RwLock},
+};
+
+use rocket::response::stream::{ReaderStream, TextStream};
+
+use engines::{
+    duckduckgo::duckduckgo::DuckDuckGo, engine_base::engine_base::EngineBase,
+    engine_base::engine_base::SearchResult,
+};
+use tokio::sync::Mutex;
+
+pub mod client;
+pub mod engines;
+pub mod utils;
+
+#[macro_use]
+extern crate rocket;
+
+struct SearchParams<'r> {
+    query: &'r str,
+}
+
+#[get("/")]
+fn index() -> &'static str {
+    "Hello, world!"
+}
+
+#[get("/search")]
+fn search_get() -> &'static str {
+    "<html>
+        <body>
+            <form method='get' action='/searchquery'>
+                <input name='query'>
+                <button type='submit'>Search</button>
+            </form>
+        </body>
+    </html>"
+}
+
+#[get("/searchquery?<query>")]
+async fn hello<'a>(query: &str) -> TextStream![String] {
+    let query_box = Box::new(query.to_string());
+
+    TextStream! {
+        let start = "<html><body>".to_string();
+        yield start;
+
+        let ddg_tv = Arc::new(
+            Mutex::new(
+                DuckDuckGo::new(),
+            ),
+        );
+        let ddg_tv_clone = ddg_tv.clone();
+
+        tokio::spawn(async move {
+            ddg_tv_clone.lock().await.search(&query_box);
+        });
+
+        let mut last_position: i32 = -1;
+
+        loop {
+            let ddg = ddg_tv.lock().await;
+            let len = ddg.results.len() as i32;
+
+            if ddg.completed && last_position == len {
+                break;
+            }
+
+            if last_position < (len - 1) {
+                for i in max(0, last_position)..=(len - 1) {
+                    match ddg.results.get(i as usize).clone() {
+                        Some(result) => {
+                            let html = format!("<br><h2>{}</h2><p>{}</p>", result.title, result.description);
+                            yield html;
+                        }
+                        None => {
+                            break;
+                        }
+                    }
+                }
+
+                last_position = len;
+            }
+        }
+
+        let end = "</body></html>".to_string();
+        yield end;
+    }
+}
+
+#[launch]
+fn rocket() -> _ {
+    rocket::build()
+        .mount("/", routes![index])
+        .mount("/", routes![hello])
+        .mount("/", routes![search_get])
+}