Reimplement is_image and is_video in coreutils

Signed-off-by: trivernis <trivernis@protonmail.com>

Cargo.lock

@@ -183,6 +183,7 @@ name = "bot-coreutils"
 version = "0.1.0"
 dependencies = [
  "log 0.4.14",
+ "mime_guess",
  "tokio",
  "url",
 ]

bot-coreutils/Cargo.lock

@@ -11,6 +11,7 @@ name = "bot-coreutils"
 version = "0.1.0"
 dependencies = [
  "log",
+ "mime_guess",
  "tokio",
  "url",
 ]
@@ -57,6 +58,22 @@ version = "0.1.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7ffc5c5338469d4d3ea17d269fa8ea3512ad247247c30bd2df69e68309ed0a08"
 
+[[package]]
+name = "mime"
+version = "0.3.16"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2a60c7ce501c71e03a9c9c0d35b861413ae925bd979cc7a4e30d060069aaac8d"
+
+[[package]]
+name = "mime_guess"
+version = "2.0.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2684d4c2e97d99848d30b324b00c8fcc7e5c897b7cbb5819b09e7c90e8baf212"
+dependencies = [
+ "mime",
+ "unicase",
+]
+
 [[package]]
 name = "percent-encoding"
 version = "2.1.0"
@@ -94,6 +111,15 @@ dependencies = [
  "pin-project-lite",
 ]
 
+[[package]]
+name = "unicase"
+version = "2.6.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "50f37be617794602aabbeee0be4f259dc1778fabe05e2d67ee8f79326d5cb4f6"
+dependencies = [
+ "version_check",
+]
+
 [[package]]
 name = "unicode-bidi"
 version = "0.3.5"
@@ -123,3 +149,9 @@ dependencies = [
  "matches",
  "percent-encoding",
 ]
+
+[[package]]
+name = "version_check"
+version = "0.9.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5fecdca9a5291cc2b8dcf7dc02453fee791a280f3743cb0905f8822ae463b3fe"

bot-coreutils/Cargo.toml

@@ -9,4 +9,5 @@ edition = "2018"
 [dependencies]
 tokio = "1.4.0"
 log = "0.4.14"
-url = "2.2.1"
+url = "2.2.1"
+mime_guess = "2.0.3"

bot-coreutils/src/tests/url_tests.rs

@@ -17,3 +17,21 @@ fn it_returns_the_domain_name() {
         Some("domain.com".to_string())
     );
 }
+
+#[test]
+fn it_checks_for_image() {
+    assert!(is_image("domain.com/image.png"));
+    assert!(is_image("https://domain.com/image.jpeg?yo=someparam"));
+    assert!(!is_image("https://domain.com"));
+    assert!(!is_image("https://domain.com/file.pdf"));
+    assert!(!is_image("not an url"));
+}
+
+#[test]
+fn it_checks_for_video() {
+    assert!(is_video("domain.com/video.mp4"));
+    assert!(is_video("https://domain.com/video.webm?yo=someparam"));
+    assert!(!is_video("https://domain.com"));
+    assert!(!is_video("https://domain.com/file.pdf"));
+    assert!(!is_video("not an url"));
+}

bot-coreutils/src/url.rs

@@ -1,3 +1,5 @@
+use mime_guess::{mime, Mime};
+
 pub use url::*;
 
 static PROTOCOL_HTTP: &str = "http://";
@@ -41,3 +43,29 @@ pub fn get_domain_for_url(url_str: &str) -> Option<String> {
 
     Some(domain.trim_start_matches("www.").to_string())
 }
+
+/// Guesses the mime for a given url string
+#[inline]
+fn guess_mime_for_url(url_str: &str) -> Option<Mime> {
+    parse_url(url_str)
+        .ok()
+        .and_then(|u| mime_guess::from_path(u.path()).first())
+}
+
+/// Returns if a given url could be an image
+pub fn is_image(url_str: &str) -> bool {
+    if let Some(guess) = guess_mime_for_url(url_str) {
+        guess.type_() == mime::IMAGE
+    } else {
+        false
+    }
+}
+
+/// Returns if a given url could be a video
+pub fn is_video(url_str: &str) -> bool {
+    if let Some(guess) = guess_mime_for_url(url_str) {
+        guess.type_() == mime::VIDEO
+    } else {
+        false
+    }
+}

src/commands/misc/sauce.rs

@@ -1,5 +1,6 @@
 use crate::messages::sauce::show_sauce_menu;
-use crate::utils::{get_previous_message_or_reply, is_image, is_video};
+use crate::utils::get_previous_message_or_reply;
+use bot_coreutils::url;
 
 use sauce_api::Sauce;
 
@@ -29,7 +30,7 @@ async fn sauce(ctx: &Context, msg: &Message) -> CommandResult {
         .attachments
         .into_iter()
         .map(|a| a.url)
-        .filter(|url| is_image(url) || is_video(url))
+        .filter(|url| url::is_image(url) || url::is_video(url))
         .collect();
 
     log::debug!("Getting embedded images...");
@@ -37,7 +38,7 @@ async fn sauce(ctx: &Context, msg: &Message) -> CommandResult {
         .embeds
         .into_iter()
         .filter_map(|e| e.thumbnail.map(|t| t.url).or(e.image.map(|i| i.url)))
-        .filter(|url| is_image(url) || is_video(url))
+        .filter(|url| url::is_image(url) || url::is_video(url))
         .collect::<Vec<String>>();
 
     attachment_urls.append(&mut embed_images);

src/utils/mod.rs

@@ -1,6 +1,5 @@
 use crate::utils::error::BotResult;
 use rand::Rng;
-use regex::Regex;
 use serenity::client::Context;
 use serenity::model::channel::Message;
 use std::collections::VecDeque;
@@ -38,38 +37,3 @@ pub async fn get_previous_message_or_reply(
 
     Ok(referenced)
 }
-
-/// Returns the file for a given domain
-pub fn get_file_name_for_domain(url: &str) -> Option<String> {
-    lazy_static::lazy_static! {
-        static ref FILE_REGEX: Regex = Regex::new(r"^(https?://)?(www\.)?(\w+\.)+\w+([^/]*/)*([^/]+)$").unwrap();
-    }
-
-    let captures = FILE_REGEX.captures(url)?;
-
-    captures.get(3).map(|c| c.as_str().to_string())
-}
-
-/// Returns if the given file is an image
-#[inline]
-pub fn is_image(url: &str) -> bool {
-    static IMAGE_EXTENSIONS: &[&str] = &["png", "webp", "jpg", "jpeg", "ico", "gif"];
-    for ext in IMAGE_EXTENSIONS {
-        if url.ends_with(ext) {
-            return true;
-        }
-    }
-    return false;
-}
-
-/// Returns if the given file is an image
-#[inline]
-pub fn is_video(url: &str) -> bool {
-    static IMAGE_EXTENSIONS: &[&str] = &["mp4", "flv", "mkv", "webm"];
-    for ext in IMAGE_EXTENSIONS {
-        if url.ends_with(ext) {
-            return true;
-        }
-    }
-    return false;
-}