Use lookbehind for spelling

Cargo.lock

@@ -8,6 +8,15 @@ version = "1.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f26201604c87b1e01bd3d98f8d5d9a8fcbb815e8cedb41ffccbeb4bf593a35fe"
 
+[[package]]
+name = "aho-corasick"
+version = "0.7.18"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1e37cfd5e7657ada45f742d6e99ca5788580b5c529dc78faf11ece6dc702656f"
+dependencies = [
+ "memchr",
+]
+
 [[package]]
 name = "async-trait"
 version = "0.1.53"
@@ -58,12 +67,27 @@ name = "basedbot"
 version = "0.1.0"
 dependencies = [
  "async-trait",
+ "fancy-regex",
  "lazy_static",
- "regex",
  "serenity",
  "tokio",
 ]
 
+[[package]]
+name = "bit-set"
+version = "0.5.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6e11e16035ea35e4e5997b393eacbf6f63983188f7a2ad25bfb13465f5ad59de"
+dependencies = [
+ "bit-vec",
+]
+
+[[package]]
+name = "bit-vec"
+version = "0.6.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "349f9b6a179ed607305526ca489b34ad0a41aed5f7980fa90eb03160b69598fb"
+
 [[package]]
 name = "bitflags"
 version = "1.3.2"
@@ -169,6 +193,16 @@ dependencies = [
  "cfg-if 1.0.0",
 ]
 
+[[package]]
+name = "fancy-regex"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0678ab2d46fa5195aaf59ad034c083d351377d4af57f3e073c074d0da3e3c766"
+dependencies = [
+ "bit-set",
+ "regex",
+]
+
 [[package]]
 name = "flate2"
 version = "1.0.24"
@@ -657,6 +691,8 @@ version = "1.5.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d83f127d94bdbcda4c8cc2e50f6f84f4b611f69c902699ca385a39c3a75f9ff1"
 dependencies = [
+ "aho-corasick",
+ "memchr",
  "regex-syntax",
 ]
 

Cargo.toml

@@ -7,6 +7,6 @@ edition = "2021"
 [dependencies]
 serenity = { version = "0.10", default_features = false, features = ["client", "rustls_backend", "model", "builder", "gateway", "utils", "cache"] }
 lazy_static = "1.4.0"
-regex = { version = "1.5.6", default_features = false, features = ["std", "unicode"] }
+fancy-regex = "0.10"
 async-trait = "0.1.53"
 tokio = { version = "1.18.2", features = ["rt-multi-thread"] }

src/main.rs

@@ -1,6 +1,7 @@
+#![feature(option_result_contains)]
 use async_trait::async_trait;
+use fancy_regex::Regex;
 use lazy_static::lazy_static;
-use regex::{Captures, Regex};
 use serenity::{
     model::{
         channel::ReactionType,
@@ -17,7 +18,8 @@ lazy_static! {
     static ref SERVER_ID: GuildId = GuildId(std::env::args().nth(1).unwrap().parse().unwrap());
     static ref MEME_CHANNEL: ChannelId =
         ChannelId(std::env::args().nth(2).unwrap().parse().unwrap());
-    static ref RETARD_REGEX: Regex = Regex::new("([^djDJh ])a( |$)").unwrap();
+    static ref RETARD_REGEX: Regex =
+        Regex::new("(?<!(. | j| d|op|in|us|ng|si|tw|dd))a( |$)").unwrap();
 }
 
 #[async_trait]
@@ -29,32 +31,24 @@ impl EventHandler for Handler {
     }
 }
 
+fn fix_spelling(msg: &str) -> String {
+    RETARD_REGEX.replace_all(&msg, "**er** ").trim().to_owned()
+}
+
 async fn handle_message(ctx: Context, message: Message) -> Result<(), serenity::Error> {
     if message.guild_id != Some(*SERVER_ID) {
         return Ok(());
     }
     // That other idiot who ends words with “a” instead of “er”
-    if message.author.id == 261246789942902794 && RETARD_REGEX.is_match(&message.content) {
+    if message.author.id == 261246789942902794
+        && RETARD_REGEX.is_match(&message.content).contains(&true)
+        && !message.content.starts_with("a ")
+    {
         message.delete(&ctx).await?;
+        let fixed = fix_spelling(&message.content);
         message
             .channel_id
-            .say(&ctx, &format!("{} wollte sagen:", message.author.mention()))
-            .await?;
-        message
-            .channel_id
-            .say(
-                &ctx,
-                RETARD_REGEX
-                    .replace_all(&message.content_safe(&ctx).await, |caps: &Captures| {
-                        format!("{}**er**{}", &caps[1], &caps[2])
-                    })
-                    // some common false positives
-                    .replace("etw**er**", "etwa")
-                    .replace("europ**er**", "europa")
-                    .replace("amerik**er**", "amerika")
-                    .replace("chin**er**", "china")
-                    .replace("mang**er**", "manga"),
-            )
+            .say(&ctx, &format!("{}: {}", message.author.mention(), fixed))
             .await?;
     }
     // that one idiot who always posts 5 links per message
@@ -114,3 +108,23 @@ async fn main() {
             .expect("Could not create client");
     client.start().await.expect("could not start");
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn fix_spelling_test() {
+        let msg = "aba ihr meint es ja bessa zu wissen ohne euch damit auseinanda gesetzt zu haben oda zu wollen";
+        let expected = "ab**er** ihr meint es ja bess**er** zu wissen ohne euch damit auseinand**er** gesetzt zu haben od**er** zu wollen";
+        assert_eq!(fix_spelling(msg), expected);
+        let msg = "your a bad person";
+        assert_eq!(fix_spelling(msg), msg);
+        let msg = "china usa europa da ja manga asia etwa unsa";
+        let expected = "china usa europa da ja manga asia etwa uns**er**";
+        assert_eq!(fix_spelling(msg), expected);
+        let msg = "guta tip";
+        let expected = "gut**er** tip";
+        assert_eq!(fix_spelling(msg), expected);
+    }
+}