Add no links rule

src/groups/calls/embeds.rs

@@ -34,7 +34,7 @@ pub fn first_warning_embed(
                 )
             )
         .title("Warning")
-        .description("You broke [Payphone's code of conduct](https://payphone.lat/code-of-conduct) during a call. Since it's your first offense, we assumed you didn't read the rules and made this a warning instead. Read the code of conduct to prevent getting banned.")
+        .description("You broke [Payphone's code of conduct](https://payphone.lat/code-of-conduct) during a call. Read the code of conduct to prevent getting banned.")
         .field(EmbedFieldBuilder::new("What message broke the rules?", ellipse(message.content.clone(), 500)))
         .field(EmbedFieldBuilder::new(
             "What rules did you break?",

src/groups/calls/moderation.rs

@@ -85,6 +85,7 @@
 
 use std::{collections::HashMap, sync::Arc};
 
+use linkify::LinkFinder;
 use once_cell::sync::Lazy;
 use regex::Regex;
 use twilight_model::gateway::payload::incoming::MessageCreate;
@@ -260,6 +261,10 @@ fn regex_from_phrases_list(phrases: &[&str]) -> Regex {
     Regex::new(&pattern).expect("Failed to compile regex")
 }
 
+/// The results of the analysis of a rule, including all detections (on the original message first,
+/// on the normalized message later) and the associated rule.
+type RuleAnalysis = ((Vec<String>, Vec<String>), Arc<dyn Rule>);
+
 /// Scans all rules on the message and returns the results for each rule which gave detections.
 ///
 /// Arguments:
@@ -267,15 +272,18 @@ fn regex_from_phrases_list(phrases: &[&str]) -> Regex {
 /// * `normalized` - The normalized version of the message.
 ///
 /// Returns:
-/// [`Vec<(Vec<String>, Arc<dyn Rule>)>`] - The results of the analysis, including all detections
-/// and their associated rules.
-pub fn analyze_message(original: &str, normalized: &str) -> Vec<(Vec<String>, Arc<dyn Rule>)> {
+/// [`Vec<RuleAnalysis>`] - The results of the analysis, including all detections (on the original
+/// message first, on the normalized message later) and their associated rules.
+pub fn analyze_message(
+    original: &str,
+    normalized: &str,
+) -> Vec<RuleAnalysis> {
     let mut results = Vec::new();
 
     for rule in RULES.iter() {
-        let detections = rule.detect(original, normalized);
-        if !detections.is_empty() {
-            results.push((detections, rule.clone()));
+        let (original_detections, normalized_detections) = rule.detect(original, normalized);
+        if !normalized_detections.is_empty() || !original_detections.is_empty() {
+            results.push(((original_detections, normalized_detections), rule.clone()));
         }
     }
 
@@ -295,15 +303,15 @@ pub fn analyze_message(original: &str, normalized: &str) -> Vec<(Vec<String>, Ar
 pub fn censor_message(
     original: &str,
     normalized: NormalizedMessage,
-    analysis: &[(Vec<String>, Arc<dyn Rule>)],
+    analysis: &[RuleAnalysis],
 ) -> String {
     let mut censored = original.to_string();
 
     // Collect all ranges to censor
     let mut ranges = Vec::new();
 
-    for (findings, _) in analysis {
-        for finding in findings {
+    for ((original_findings, normalized_findings), _) in analysis {
+        for finding in normalized_findings {
             for (index, _) in normalized.message.match_indices(finding) {
                 let start_index = normalized.mappings.get(index).cloned().unwrap_or(0);
                 let end_index = normalized
@@ -315,6 +323,15 @@ pub fn censor_message(
                 ranges.push((start_index, end_index));
             }
         }
+
+        for finding in original_findings {
+            for (index, _) in original.match_indices(finding) {
+                let start_index = index;
+                let end_index = index + finding.len();
+
+                ranges.push((start_index, end_index));
+            }
+        }
     }
 
     // Sort ranges in reverse order to avoid shifting indices
@@ -380,10 +397,9 @@ pub trait Rule: Send + Sync {
     /// * `normalized` - The normalized version of the message.
     ///
     /// Returns:
-    /// [`Vec<String>`] - The list of detected rule violations in the normalized message.
-    ///
-    /// TODO: Support detecting violations in the original message.
-    fn detect(&self, original: &str, normalized: &str) -> Vec<String>;
+    /// [`(Vec<String>, Vec<String>)`] - The list of detected rule violations in the original
+    /// (first vector) and normalized (second vector) message.
+    fn detect(&self, original: &str, normalized: &str) -> (Vec<String>, Vec<String>);
 }
 
 /// Parses a file containing phrases to be detected by rules.
@@ -429,11 +445,14 @@ impl Rule for HateSpeechAndSlursRule {
         &Action::BlockAndPenalty
     }
 
-    fn detect(&self, _original: &str, normalized: &str) -> Vec<String> {
-        self.0
-            .find_iter(normalized)
-            .map(|m| normalized[m.start()..m.end()].to_string())
-            .collect()
+    fn detect(&self, _original: &str, normalized: &str) -> (Vec<String>, Vec<String>) {
+        (
+            vec![],
+            self.0
+                .find_iter(normalized)
+                .map(|m| normalized[m.start()..m.end()].to_string())
+                .collect(),
+        )
     }
 }
 
@@ -459,11 +478,14 @@ impl Rule for ThreatsAndSelfHarmRule {
         &Action::BlockAndPenalty
     }
 
-    fn detect(&self, _original: &str, normalized: &str) -> Vec<String> {
-        self.0
-            .find_iter(normalized)
-            .map(|m| normalized[m.start()..m.end()].to_string())
-            .collect()
+    fn detect(&self, _original: &str, normalized: &str) -> (Vec<String>, Vec<String>) {
+        (
+            vec![],
+            self.0
+                .find_iter(normalized)
+                .map(|m| normalized[m.start()..m.end()].to_string())
+                .collect(),
+        )
     }
 }
 
@@ -489,11 +511,14 @@ impl Rule for ExplicitSexualLanguageRule {
         "Explicit Sexual Language"
     }
 
-    fn detect(&self, _original: &str, normalized: &str) -> Vec<String> {
-        self.0
-            .find_iter(normalized)
-            .map(|m| normalized[m.start()..m.end()].to_string())
-            .collect()
+    fn detect(&self, _original: &str, normalized: &str) -> (Vec<String>, Vec<String>) {
+        (
+            vec![],
+            self.0
+                .find_iter(normalized)
+                .map(|m| normalized[m.start()..m.end()].to_string())
+                .collect(),
+        )
     }
 }
 
@@ -518,11 +543,36 @@ impl Rule for SuggestiveLanguageRule {
         "Suggestive Language"
     }
 
-    fn detect(&self, _original: &str, normalized: &str) -> Vec<String> {
-        self.0
-            .find_iter(normalized)
-            .map(|m| normalized[m.start()..m.end()].to_string())
-            .collect()
+    fn detect(&self, _original: &str, normalized: &str) -> (Vec<String>, Vec<String>) {
+        (
+            vec![],
+            self.0
+                .find_iter(normalized)
+                .map(|m| normalized[m.start()..m.end()].to_string())
+                .collect(),
+        )
+    }
+}
+
+struct LinksRule;
+
+impl Rule for LinksRule {
+    fn name(&self) -> &'static str {
+        "Links"
+    }
+
+    fn action(&self) -> &'static Action {
+        &Action::SoftCensorAndWarn
+    }
+
+    fn detect(&self, original: &str, _normalized: &str) -> (Vec<String>, Vec<String>) {
+        let mut detections = Vec::new();
+
+        for link in LinkFinder::new().url_must_have_scheme(false).links(original) {
+            detections.push(link.as_str().to_string());
+        }
+
+        (detections, vec![])
     }
 }
 
@@ -533,5 +583,6 @@ pub static RULES: Lazy<Vec<Arc<dyn Rule>>> = Lazy::new(|| {
         Arc::new(ThreatsAndSelfHarmRule::new()),
         Arc::new(SuggestiveLanguageRule::new()),
         Arc::new(ExplicitSexualLanguageRule::new()),
+        Arc::new(LinksRule),
     ]
 });