Update other categorization

app/Services/Categorization/Categorizers/MiscCategorizer.php

@@ -100,7 +100,10 @@ class MiscCategorizer extends AbstractCategorizer
 
         $signalScore = count($markers);
         $isCoreToken = preg_match('/^[A-Za-z0-9+\/_=-]+$/', $coreName) === 1;
-        $lowSignal = $signalScore === 0 && $isCoreToken && strlen($coreName) >= 12;
+        $lowSignal = $signalScore === 0
+            && $isCoreToken
+            && strlen($coreName) >= 12
+            && ! $this->hasStrongWordStructure($name, $coreName);
 
         return [
             'coreName' => $coreName,
@@ -226,6 +229,10 @@ class MiscCategorizer extends AbstractCategorizer
     {
         $analysis = $this->inspectSignals($name);
 
+        if ($this->hasStrongWordStructure($name, $analysis['coreName'])) {
+            return null;
+        }
+
         if ($this->isZeroVowelLongToken($analysis['coreName'])) {
             return null;
         }
@@ -246,6 +253,23 @@ class MiscCategorizer extends AbstractCategorizer
         return null;
     }
 
+    protected function hasStrongWordStructure(string $name, string $coreName): bool
+    {
+        return $this->getMaxConsecutiveLetters($coreName) >= 5
+            && $this->hasNormalVowelRatio($coreName)
+            && $this->countAlphabeticWordTokens($name) >= 2;
+    }
+
+    protected function countAlphabeticWordTokens(string $name): int
+    {
+        $tokens = preg_split('/[.\s_-]+/', $this->stripExtensionsForAnalysis($name)) ?: [];
+
+        return count(array_filter(
+            $tokens,
+            static fn (string $token): bool => preg_match('/[a-z]{3,}/i', $token) === 1
+        ));
+    }
+
     protected function checkArchive(string $name): ?CategorizationResult
     {
         if (preg_match('/\.(zip|rar|7z|tar|gz|bz2|xz|tgz|tbz2|cab|iso|img|dmg|pkg|archive)$/i', $name)) {

app/Services/Categorization/Categorizers/PcCategorizer.php

@@ -148,6 +148,10 @@ class PcCategorizer extends AbstractCategorizer
 
     protected function check0day(string $name): ?CategorizationResult
     {
+        if (preg_match('/\.(msix|msixbundle|appx|appxbundle|msi)$/i', $name)) {
+            return $this->matched(Category::PC_0DAY, 0.9, '0day_msix_installer');
+        }
+
         // Explicit 0day indicators
         if (preg_match('/[._ -]exe$|[._ -](utorrent|Virtualbox)[._ -]|\b0DAY\b|incl.+crack| DRM$|>DRM</i', $name)) {
             return $this->matched(Category::PC_0DAY, 0.9, '0day_explicit');

app/Traits/DetectsHashedNames.php

@@ -264,6 +264,23 @@ trait DetectsHashedNames
         return preg_match('/[aeiou]/i', $letterOnly) !== 1;
     }
 
+    /**
+     * Check whether the letter-only portion of a token has a normal vowel ratio.
+     */
+    protected function hasNormalVowelRatio(string $str, float $min = 0.2): bool
+    {
+        $letters = preg_replace('/[^a-z]/i', '', $str);
+
+        if ($letters === '' || strlen($letters) < 5) {
+            return false;
+        }
+
+        preg_match_all('/[aeiou]/i', $letters, $matches);
+        $vowelCount = count($matches[0]);
+
+        return ($vowelCount / strlen($letters)) >= $min;
+    }
+
     /**
      * Check if a string looks like a random/obfuscated string rather than a real title.
      *
@@ -315,7 +332,7 @@ trait DetectsHashedNames
     protected function stripExtensionsForAnalysis(string $name): string
     {
         return preg_replace(
-            '/\.(mkv|avi|mp4|m4v|mpg|mpeg|wmv|flv|mov|ts|vob|iso|divx|par2?|nfo|sfv|nzb|rar|r\d{2,3}|zip|7z|gz|tar|001)$/i',
+            '/\.(mkv|avi|mp4|m4v|mpg|mpeg|wmv|flv|mov|ts|vob|iso|divx|par2?|nfo|sfv|nzb|rar|r\d{2,3}|zip|7z|gz|tar|001|msix|msixbundle|appx|appxbundle|apk|xap|ipa|deb|rpm|pkg|dmg|exe|msi)$/i',
             '',
             trim($name)
         );