Merge pull request #190 from neon-mmd/feat-content-blocking-using-lists

✨ Filter user specified content from search results using a `blacklist` and `Allowlist`.

Cargo.lock

@@ -77,7 +77,7 @@ dependencies = [
  "encoding_rs",
  "flate2",
  "futures-core",
- "h2 0.3.20",
+ "h2 0.3.21",
  "http 0.2.9",
  "httparse",
  "httpdate",
@@ -475,9 +475,9 @@ checksum = "37b2a672a2cb129a2e41c10b1224bb368f9f37a2b16b612598138befd7b37eb5"
 
 [[package]]
 name = "cc"
-version = "1.0.82"
+version = "1.0.83"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "305fe645edc1442a0fa8b6726ba61d422798d37a52e12eaecf4b022ebbb88f01"
+checksum = "f1174fb0b6ec23863f8b971027804a42614e347eafb0a95bf0b12cdae21fc4d0"
 dependencies = [
  "jobserver",
  "libc",
@@ -816,9 +816,9 @@ dependencies = [
 
 [[package]]
 name = "deranged"
-version = "0.3.7"
+version = "0.3.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7684a49fb1af197853ef7b2ee694bc1f5b4179556f1e5710e1760c5db6f5e929"
+checksum = "f2696e8a945f658fd14dc3b87242e6b80cd0f36ff04ea560fa39082368847946"
 
 [[package]]
 name = "derive_more"
@@ -1176,9 +1176,9 @@ dependencies = [
 
 [[package]]
 name = "h2"
-version = "0.3.20"
+version = "0.3.21"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "97ec8491ebaf99c8eaa73058b045fe58073cd6be7f596ac993ced0b0a0c01049"
+checksum = "91fc23aa11be92976ef4729127f1a74adf36d8436f7816b185d18df956790833"
 dependencies = [
  "bytes 1.4.0",
  "fnv",
@@ -1363,7 +1363,7 @@ dependencies = [
  "futures-channel",
  "futures-core",
  "futures-util",
- "h2 0.3.20",
+ "h2 0.3.21",
  "http 0.2.9",
  "http-body 0.4.5",
  "httparse",
@@ -2454,16 +2454,16 @@ dependencies = [
 
 [[package]]
 name = "reqwest"
-version = "0.11.18"
+version = "0.11.19"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cde824a14b7c14f85caff81225f411faacc04a2013f41670f41443742b1c1c55"
+checksum = "20b9b67e2ca7dd9e9f9285b759de30ff538aab981abaaf7bc9bd90b84a0126c3"
 dependencies = [
  "base64 0.21.2",
  "bytes 1.4.0",
  "encoding_rs",
  "futures-core",
  "futures-util",
- "h2 0.3.20",
+ "h2 0.3.21",
  "http 0.2.9",
  "http-body 0.4.5",
  "hyper 0.14.27",
@@ -2486,7 +2486,7 @@ dependencies = [
  "wasm-bindgen",
  "wasm-bindgen-futures",
  "web-sys",
- "winreg 0.10.1",
+ "winreg 0.50.0",
 ]
 
 [[package]]
@@ -2684,18 +2684,18 @@ checksum = "388a1df253eca08550bef6c72392cfe7c30914bf41df5269b68cbd6ff8f570a3"
 
 [[package]]
 name = "serde"
-version = "1.0.183"
+version = "1.0.185"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "32ac8da02677876d532745a130fc9d8e6edfa81a269b107c5b00829b91d8eb3c"
+checksum = "be9b6f69f1dfd54c3b568ffa45c310d6973a5e5148fd40cf515acaf38cf5bc31"
 dependencies = [
  "serde_derive",
 ]
 
 [[package]]
 name = "serde_derive"
-version = "1.0.183"
+version = "1.0.185"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "aafe972d60b0b9bee71a91b92fee2d4fb3c9d7e8f6b179aa99f27203d99a4816"
+checksum = "dc59dfdcbad1437773485e0367fea4b090a2e0a16d9ffc46af47764536a298ec"
 dependencies = [
  "proc-macro2 1.0.66",
  "quote 1.0.33",
@@ -2797,9 +2797,9 @@ checksum = "7bd3e3206899af3f8b12af284fafc038cc1dc2b41d1b89dd17297221c5d225de"
 
 [[package]]
 name = "slab"
-version = "0.4.8"
+version = "0.4.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6528351c9bc8ab22353f9d776db39a20288e8d6c37ef8cfe3317cf875eecfc2d"
+checksum = "8f92a496fb766b417c996b9c5e57daf2f7ad3b0bebe1ccfca4856390e3d3bb67"
 dependencies = [
  "autocfg 1.1.0",
 ]
@@ -3328,9 +3328,9 @@ checksum = "ed646292ffc8188ef8ea4d1e0e0150fb15a5c2e12ad9b8fc191ae7a8a7f3c4b9"
 
 [[package]]
 name = "unicase"
-version = "2.6.0"
+version = "2.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "50f37be617794602aabbeee0be4f259dc1778fabe05e2d67ee8f79326d5cb4f6"
+checksum = "f7d2d4dafb69621809a81864c9c1b864479e1235c0dd4e199924b9742439ed89"
 dependencies = [
  "version_check",
 ]
@@ -3543,7 +3543,7 @@ dependencies = [
 
 [[package]]
 name = "websurfx"
-version = "0.17.0"
+version = "0.18.0"
 dependencies = [
  "actix-cors",
  "actix-files",
@@ -3559,7 +3559,8 @@ dependencies = [
  "once_cell",
  "rand 0.8.5",
  "redis",
- "reqwest 0.11.18",
+ "regex",
+ "reqwest 0.11.19",
  "rlua",
  "rusty-hook",
  "scraper",
@@ -3688,11 +3689,12 @@ dependencies = [
 
 [[package]]
 name = "winreg"
-version = "0.10.1"
+version = "0.50.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "80d0f4e272c85def139476380b12f9ac60926689dd2e01d4923222f40580869d"
+checksum = "524e57b2c537c0f9b1e69f1965311ec12182b4122e45035b1508cd24d2adadb1"
 dependencies = [
- "winapi 0.3.9",
+ "cfg-if 1.0.0",
+ "windows-sys",
 ]
 
 [[package]]

Cargo.toml

@@ -1,15 +1,15 @@
 [package]
 name = "websurfx"
-version = "0.17.0"
+version = "0.18.0"
 edition = "2021"
 description = "An open-source alternative to Searx that provides clean, ad-free, and organic results with incredible speed while keeping privacy and security in mind."
 repository = "https://github.com/neon-mmd/websurfx"
 license = "AGPL-3.0"
 
 [dependencies]
-reqwest = {version="0.11.18",features=["json"]}
+reqwest = {version="0.11.19",features=["json"]}
 tokio = {version="1.32.0",features=["full"]}
-serde = {version="1.0.183",features=["derive"]}
+serde = {version="1.0.185",features=["derive"]}
 handlebars = { version = "4.3.7", features = ["dir_source"] }
 scraper = {version="0.17.1"}
 actix-web = {version="4.3.1", features = ["cookies"]}
@@ -26,6 +26,7 @@ rand={version="0.8.5"}
 once_cell = {version="1.18.0"}
 error-stack = {version="0.3.1"}
 async-trait = {version="0.1.73"}
+regex = {version="1.9.3", features=["perf"]}
 
 [dev-dependencies]
 rusty-hook = "^0.11.2"
@@ -50,7 +51,7 @@ split-debuginfo = '...'
 debug-assertions = false
 overflow-checks = false
 lto = 'thin'
-panic = 'unwind'
+panic = 'abort'
 incremental = false
 codegen-units = 16
 rpath = false

src/config/parser.rs

@@ -1,14 +1,12 @@
 //! This module provides the functionality to parse the lua config and convert the config options
 //! into rust readable form.
 
+use crate::handler::paths::{file_path, FileType};
+
 use super::parser_models::Style;
 use log::LevelFilter;
 use rlua::Lua;
-use std::{collections::HashMap, format, fs, path::Path, thread::available_parallelism};
-
-// ------- Constants --------
-static COMMON_DIRECTORY_NAME: &str = "websurfx";
-static CONFIG_FILE_NAME: &str = "config.lua";
+use std::{collections::HashMap, fs, thread::available_parallelism};
 
 /// A named struct which stores the parsed config file options.
 ///
@@ -69,7 +67,7 @@ impl Config {
             let globals = context.globals();
 
             context
-                .load(&fs::read_to_string(Config::config_path()?)?)
+                .load(&fs::read_to_string(file_path(FileType::Config)?)?)
                 .exec()?;
 
             let parsed_threads: u8 = globals.get::<_, u8>("threads")?;
@@ -114,52 +112,6 @@ impl Config {
             })
         })
     }
-
-    /// A helper function which returns an appropriate config file path checking if the config
-    /// file exists on that path.
-    ///
-    /// # Error
-    ///
-    /// Returns a `config file not found!!` error if the config file is not present under following
-    /// paths which are:
-    /// 1. `~/.config/websurfx/` if it not present here then it fallbacks to the next one (2)
-    /// 2. `/etc/xdg/websurfx/config.lua` if it is not present here then it fallbacks to the next
-    ///    one (3).
-    /// 3. `websurfx/` (under project folder ( or codebase in other words)) if it is not present
-    ///    here then it returns an error as mentioned above.
-    fn config_path() -> Result<String, Box<dyn std::error::Error>> {
-        // check user config
-
-        let path = format!(
-            "{}/.config/{}/config.lua",
-            std::env::var("HOME").unwrap(),
-            COMMON_DIRECTORY_NAME
-        );
-        if Path::new(path.as_str()).exists() {
-            return Ok(format!(
-                "{}/.config/{}/{}",
-                std::env::var("HOME").unwrap(),
-                COMMON_DIRECTORY_NAME,
-                CONFIG_FILE_NAME
-            ));
-        }
-
-        // look for config in /etc/xdg
-        if Path::new(format!("/etc/xdg/{}/{}", COMMON_DIRECTORY_NAME, CONFIG_FILE_NAME).as_str())
-            .exists()
-        {
-            return Ok("/etc/xdg/websurfx/config.lua".to_string());
-        }
-
-        // use dev config
-        if Path::new(format!("./{}/{}", COMMON_DIRECTORY_NAME, CONFIG_FILE_NAME).as_str()).exists()
-        {
-            return Ok("./websurfx/config.lua".to_string());
-        }
-
-        // if no of the configs above exist, return error
-        Err("Config file not found!!".to_string().into())
-    }
 }
 
 /// a helper function that sets the proper logging level

src/handler/mod.rs

@@ -1 +1 @@
-pub mod public_paths;
+pub mod paths;

src/handler/paths.rs

@@ -0,0 +1,111 @@
+//! This module provides the functionality to handle theme folder present on different paths and
+//! provide one appropriate path on which it is present and can be used.
+
+use std::collections::HashMap;
+use std::io::Error;
+use std::path::Path;
+
+// ------- Constants --------
+static PUBLIC_DIRECTORY_NAME: &str = "public";
+static COMMON_DIRECTORY_NAME: &str = "websurfx";
+static CONFIG_FILE_NAME: &str = "config.lua";
+static ALLOWLIST_FILE_NAME: &str = "allowlist.txt";
+static BLOCKLIST_FILE_NAME: &str = "blocklist.txt";
+
+#[derive(Hash, PartialEq, Eq, Debug)]
+pub enum FileType {
+    Config,
+    AllowList,
+    BlockList,
+    Theme,
+}
+
+static FILE_PATHS_FOR_DIFF_FILE_TYPES: once_cell::sync::Lazy<HashMap<FileType, Vec<String>>> =
+    once_cell::sync::Lazy::new(|| {
+        HashMap::from([
+            (
+                FileType::Config,
+                vec![
+                    format!(
+                        "{}/.config/{}/{}",
+                        std::env::var("HOME").unwrap(),
+                        COMMON_DIRECTORY_NAME,
+                        CONFIG_FILE_NAME
+                    ),
+                    format!("/etc/xdg/{}/{}", COMMON_DIRECTORY_NAME, CONFIG_FILE_NAME),
+                    format!("./{}/{}", COMMON_DIRECTORY_NAME, CONFIG_FILE_NAME),
+                ],
+            ),
+            (
+                FileType::Theme,
+                vec![
+                    format!("/opt/websurfx/{}/", PUBLIC_DIRECTORY_NAME),
+                    format!("./{}/", PUBLIC_DIRECTORY_NAME),
+                ],
+            ),
+            (
+                FileType::AllowList,
+                vec![
+                    format!(
+                        "{}/.config/{}/{}",
+                        std::env::var("HOME").unwrap(),
+                        COMMON_DIRECTORY_NAME,
+                        ALLOWLIST_FILE_NAME
+                    ),
+                    format!("/etc/xdg/{}/{}", COMMON_DIRECTORY_NAME, ALLOWLIST_FILE_NAME),
+                    format!("./{}/{}", COMMON_DIRECTORY_NAME, ALLOWLIST_FILE_NAME),
+                ],
+            ),
+            (
+                FileType::BlockList,
+                vec![
+                    format!(
+                        "{}/.config/{}/{}",
+                        std::env::var("HOME").unwrap(),
+                        COMMON_DIRECTORY_NAME,
+                        BLOCKLIST_FILE_NAME
+                    ),
+                    format!("/etc/xdg/{}/{}", COMMON_DIRECTORY_NAME, BLOCKLIST_FILE_NAME),
+                    format!("./{}/{}", COMMON_DIRECTORY_NAME, BLOCKLIST_FILE_NAME),
+                ],
+            ),
+        ])
+    });
+
+/// A helper function which returns an appropriate config file path checking if the config
+/// file exists on that path.
+///
+/// # Error
+///
+/// Returns a `config file not found!!` error if the config file is not present under following
+/// paths which are:
+/// 1. `~/.config/websurfx/` if it not present here then it fallbacks to the next one (2)
+/// 2. `/etc/xdg/websurfx/config.lua` if it is not present here then it fallbacks to the next
+///    one (3).
+/// 3. `websurfx/` (under project folder ( or codebase in other words)) if it is not present
+///    here then it returns an error as mentioned above.
+
+/// A function which returns an appropriate theme directory path checking if the theme
+/// directory exists on that path.
+///
+/// # Error
+///
+/// Returns a `Theme (public) folder not found!!` error if the theme folder is not present under following
+/// paths which are:
+/// 1. `/opt/websurfx` if it not present here then it fallbacks to the next one (2)
+/// 2. Under project folder ( or codebase in other words) if it is not present
+///    here then it returns an error as mentioned above.
+pub fn file_path(file_type: FileType) -> Result<String, Error> {
+    let file_path = FILE_PATHS_FOR_DIFF_FILE_TYPES.get(&file_type).unwrap();
+    for (idx, _) in file_path.iter().enumerate() {
+        if Path::new(file_path[idx].as_str()).exists() {
+            return Ok(file_path[idx].clone());
+        }
+    }
+
+    // if no of the configs above exist, return error
+    Err(Error::new(
+        std::io::ErrorKind::NotFound,
+        format!("{:?} file not found!!", file_type),
+    ))
+}

src/handler/public_paths.rs

@@ -1,33 +0,0 @@
-//! This module provides the functionality to handle theme folder present on different paths and
-//! provide one appropriate path on which it is present and can be used.
-
-use std::io::Error;
-use std::path::Path;
-
-// ------- Constants --------
-static PUBLIC_DIRECTORY_NAME: &str = "public";
-
-/// A function which returns an appropriate theme directory path checking if the theme
-/// directory exists on that path.
-///
-/// # Error
-///
-/// Returns a `Theme (public) folder not found!!` error if the theme folder is not present under following
-/// paths which are:
-/// 1. `/opt/websurfx` if it not present here then it fallbacks to the next one (2)
-/// 2. Under project folder ( or codebase in other words) if it is not present
-///    here then it returns an error as mentioned above.
-pub fn public_path() -> Result<String, Error> {
-    if Path::new(format!("/opt/websurfx/{}/", PUBLIC_DIRECTORY_NAME).as_str()).exists() {
-        return Ok(format!("/opt/websurfx/{}", PUBLIC_DIRECTORY_NAME));
-    }
-
-    if Path::new(format!("./{}/", PUBLIC_DIRECTORY_NAME).as_str()).exists() {
-        return Ok(format!("./{}", PUBLIC_DIRECTORY_NAME));
-    }
-
-    Err(Error::new(
-        std::io::ErrorKind::NotFound,
-        "Themes (public) folder not found!!",
-    ))
-}

src/lib.rs

@@ -17,7 +17,7 @@ use actix_files as fs;
 use actix_web::{dev::Server, http::header, middleware::Logger, web, App, HttpServer};
 use config::parser::Config;
 use handlebars::Handlebars;
-use handler::public_paths::public_path;
+use handler::paths::{file_path, FileType};
 
 /// Runs the web server on the provided TCP listener and returns a `Server` instance.
 ///
@@ -42,7 +42,7 @@ use handler::public_paths::public_path;
 pub fn run(listener: TcpListener, config: Config) -> std::io::Result<Server> {
     let mut handlebars: Handlebars = Handlebars::new();
 
-    let public_folder_path: String = public_path()?;
+    let public_folder_path: String = file_path(FileType::Theme)?;
 
     handlebars
         .register_templates_directory(".html", format!("{}/templates", public_folder_path))

src/results/aggregator.rs

@@ -1,18 +1,26 @@
 //! This module provides the functionality to scrape and gathers all the results from the upstream
 //! search engines and then removes duplicate results.
 
-use std::{collections::HashMap, time::Duration};
-
-use error_stack::Report;
-use rand::Rng;
-use tokio::task::JoinHandle;
+use std::{
+    collections::HashMap,
+    io::{BufReader, Read},
+    time::Duration,
+};
 
 use super::{
     aggregation_models::{EngineErrorInfo, SearchResult, SearchResults},
     user_agent::random_user_agent,
 };
+use error_stack::Report;
+use rand::Rng;
+use regex::Regex;
+use std::{fs::File, io::BufRead};
+use tokio::task::JoinHandle;
 
-use crate::engines::engine_models::{EngineError, EngineHandler};
+use crate::{
+    engines::engine_models::{EngineError, EngineHandler},
+    handler::paths::{file_path, FileType},
+};
 
 /// Aliases for long type annotations
 type FutureVec = Vec<JoinHandle<Result<HashMap<String, SearchResult>, Report<EngineError>>>>;
@@ -106,7 +114,7 @@ pub async fn aggregate(
         log::error!("Engine Error: {:?}", error);
         engine_errors_info.push(EngineErrorInfo::new(
             error.downcast_ref::<EngineError>().unwrap(),
-            engine_name.to_string(),
+            engine_name,
         ));
     };
 
@@ -143,7 +151,22 @@ pub async fn aggregate(
         }
     }
 
-    let results = result_map.into_values().collect();
+    let mut blacklist_map: HashMap<String, SearchResult> = HashMap::new();
+    filter_with_lists(
+        &mut result_map,
+        &mut blacklist_map,
+        &file_path(FileType::BlockList)?,
+    )?;
+
+    filter_with_lists(
+        &mut blacklist_map,
+        &mut result_map,
+        &file_path(FileType::AllowList)?,
+    )?;
+
+    drop(blacklist_map);
+
+    let results: Vec<SearchResult> = result_map.into_values().collect();
 
     Ok(SearchResults::new(
         results,
@@ -151,3 +174,23 @@ pub async fn aggregate(
         engine_errors_info,
     ))
 }
+
+fn filter_with_lists(
+    map_to_be_filtered: &mut HashMap<String, SearchResult>,
+    resultant_map: &mut HashMap<String, SearchResult>,
+    file_path: &str,
+) -> Result<(), Box<dyn std::error::Error>> {
+    let mut reader = BufReader::new(File::open(file_path)?);
+    for line in reader.by_ref().lines() {
+        let re = Regex::new(&line?)?;
+        for (url, search_result) in map_to_be_filtered.clone().into_iter() {
+            if re.is_match(&url.to_lowercase())
+                || re.is_match(&search_result.title.to_lowercase())
+                || re.is_match(&search_result.description.to_lowercase())
+            {
+                resultant_map.insert(url.clone(), map_to_be_filtered.remove(&url).unwrap());
+            }
+        }
+    }
+    Ok(())
+}

src/server/routes.rs

@@ -8,7 +8,7 @@ use crate::{
     cache::cacher::RedisCache,
     config::parser::Config,
     engines::engine_models::EngineHandler,
-    handler::public_paths::public_path,
+    handler::paths::{file_path, FileType},
     results::{aggregation_models::SearchResults, aggregator::aggregate},
 };
 use actix_web::{get, web, HttpRequest, HttpResponse};
@@ -215,7 +215,8 @@ async fn results(
 /// Handles the route of robots.txt page of the `websurfx` meta search engine website.
 #[get("/robots.txt")]
 pub async fn robots_data(_req: HttpRequest) -> Result<HttpResponse, Box<dyn std::error::Error>> {
-    let page_content: String = read_to_string(format!("{}/robots.txt", public_path()?))?;
+    let page_content: String =
+        read_to_string(format!("{}/robots.txt", file_path(FileType::Theme)?))?;
     Ok(HttpResponse::Ok()
         .content_type("text/plain; charset=ascii")
         .body(page_content))