From cc80a10088c8fd873fb683d978476e9280f4c408 Mon Sep 17 00:00:00 2001
From: Rainnny7 <braydonrainnny@gmail.com>
Date: Thu, 6 Jun 2024 22:09:04 -0400
Subject: [PATCH] Support elongated words

---
 .../processor/impl/VulgarityProcessor.java         |  8 +++++---
 .../profanity/service/FiltrationService.java       | 14 ++++++++++----
 2 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/API/src/main/java/me/braydon/profanity/processor/impl/VulgarityProcessor.java b/API/src/main/java/me/braydon/profanity/processor/impl/VulgarityProcessor.java
index 3f48636..3118b4b 100644
--- a/API/src/main/java/me/braydon/profanity/processor/impl/VulgarityProcessor.java
+++ b/API/src/main/java/me/braydon/profanity/processor/impl/VulgarityProcessor.java
@@ -28,6 +28,7 @@ public final class VulgarityProcessor extends TextProcessor {
      * Substitutions for characters in profane words.
      */
     private static final Map<Character, List<Character>> charSubstitutions = Collections.synchronizedMap(new HashMap<>());
+
     static { // Populate char substitutions
         charSubstitutions.put('e', Collections.singletonList('3'));
         charSubstitutions.put('i', List.of('1', '!'));
@@ -75,8 +76,9 @@ public final class VulgarityProcessor extends TextProcessor {
                 matched.add(word);
                 int start = offset + matcher.start();
                 int end = offset + matcher.end();
-                replacement.replace(start, end, Character.toString(replaceChar).repeat(word.length()));
-                offset += word.length() - (end - start);
+                String matchedWord = matcher.group();
+                replacement.replace(start, end, Character.toString(replaceChar).repeat(matchedWord.length()));
+                offset += matchedWord.length() - (end - start);
             }
         }
 
@@ -122,7 +124,7 @@ public final class VulgarityProcessor extends TextProcessor {
                 for (Character substitution : charSubstitutions.get(lowerChar)) {
                     chars.append(substitution);
                 }
-                obfuscatedWordRegex.append('[').append(chars).append(']');
+                obfuscatedWordRegex.append('[').append(chars).append("]+");
             } else {
                 obfuscatedWordRegex.append(lowerChar);
             }
diff --git a/API/src/main/java/me/braydon/profanity/service/FiltrationService.java b/API/src/main/java/me/braydon/profanity/service/FiltrationService.java
index 6e55826..f77be0a 100644
--- a/API/src/main/java/me/braydon/profanity/service/FiltrationService.java
+++ b/API/src/main/java/me/braydon/profanity/service/FiltrationService.java
@@ -88,16 +88,22 @@ public final class FiltrationService {
         log.info("Downloaded lists in {}ms", System.currentTimeMillis() - before);
     }
 
+    /**
+     * Filter the content in the given input.
+     *
+     * @param input the input to filter
+     * @return the response from filtering the content
+     */
     @NonNull
     public ContentProcessResponse process(@NonNull ContentProcessInput input) {
+        String content = StringEscapeUtils.escapeJava(input.getContent()).toLowerCase().trim(); // The content to filter
+
         List<String> matched = new ArrayList<>(); // The content that was matched
         List<ContentTag> tags = new ArrayList<>(); // Tags obtained from the processed content
         StringBuilder replacement = new StringBuilder(input.getContent());
 
         // Handle filtering if a profanity list is present
         if (profanityList != null) {
-            String content = StringEscapeUtils.escapeJava(input.getContent()).toLowerCase().trim(); // The content to filter
-
             // Invoke each text processor on the content
             for (TextProcessor textProcessor : textProcessors) {
                 int before = matched.size();
@@ -109,8 +115,8 @@ public final class FiltrationService {
         }
 
         // Calculate the score based on
-        // the matched profane content
-        double score = 0D;
+        // the matched profane content, that cannot be bypassed by changing the content length
+        double score = Math.min(matched.stream().mapToDouble(String::length).sum() / content.length(), 1D);
 
         return new ContentProcessResponse(replacement.toString(), matched, tags, score);
     }