Merge pull request #25 from ginkelsoft-development/develop

Develop into Main

Author: ginkelsoft-development

composer.json

@@ -20,7 +20,8 @@
   ],
   "require": {
     "php": "^8.1 || ^8.2 || ^8.3 || ^8.4",
-    "illuminate/support": "^9.0 || ^10.0 || ^11.0 || ^12.0"
+    "illuminate/support": "^9.0 || ^10.0 || ^11.0 || ^12.0",
+    "ext-intl": "*"
   },
   "require-dev": {
     "phpunit/phpunit": "^9.5.10 || ^10.0 || ^11.0",

config/encrypted-search.php

@@ -36,7 +36,7 @@
     |--------------------------------------------------------------------------
     |
     | The maximum number of prefix levels to generate for prefix-based search.
-    | For example, the term “wietse” would generate:
+    | For example, the term "wietse" would generate:
     |   ["w", "wi", "wie", "wiet", "wiets", "wietse"]
     |
     | Increasing this value improves search precision for short terms, but
@@ -45,6 +45,29 @@
     */
     'max_prefix_depth' => 6,
 
+    /*
+    |--------------------------------------------------------------------------
+    | Minimum Prefix Length
+    |--------------------------------------------------------------------------
+    |
+    | The minimum number of characters required for prefix-based searches.
+    | This prevents overly broad matches from very short search terms.
+    |
+    | For example, with min_prefix_length = 3:
+    | - Searching for "Wi" (2 chars) will return no results
+    | - Searching for "Wil" (3 chars) will work normally
+    |
+    | This helps prevent performance issues and reduces false positives
+    | when users search for very short terms like "a" or "de".
+    |
+    | Recommended values:
+    | - 2: Allow two-character searches (more flexible, more false positives)
+    | - 3: Require three characters (good balance)
+    | - 4: Require four characters (very precise, less flexible)
+    |
+    */
+    'min_prefix_length' => env('ENCRYPTED_SEARCH_MIN_PREFIX', 3),
+
     /*
     |--------------------------------------------------------------------------
     | Automatic Indexing of Encrypted Casts
@@ -82,4 +105,18 @@
         'host'    => env('ELASTICSEARCH_HOST', 'http://elasticsearch:9200'),
         'index'   => env('ELASTICSEARCH_INDEX', 'encrypted_search'),
     ],
+
+    /*
+    |--------------------------------------------------------------------------
+    | Debug Logging
+    |--------------------------------------------------------------------------
+    |
+    | Enable debug logging for encrypted search operations. When enabled,
+    | the package will log token generation, index updates, and deletions
+    | to help with debugging and monitoring.
+    |
+    | Warning: This can generate a lot of log entries in high-traffic applications.
+    |
+    */
+    'debug' => env('ENCRYPTED_SEARCH_DEBUG', false),
 ];

src/Contracts/SearchDriver.php

@@ -53,6 +53,9 @@ public function register(): void
      */
     public function boot(): void
     {
+        // Validate configuration
+        $this->validateConfiguration();
+
         // Publish configuration
         $this->publishes([
             __DIR__ . '/../config/encrypted-search.php' => config_path('encrypted-search.php'),
@@ -75,4 +78,35 @@ public function boot(): void
         // Listen for all Eloquent model events and route them through the observer
         Event::listen('eloquent.*: *', SearchIndexObserver::class);
     }
+
+    /**
+     * Validate package configuration at boot time.
+     *
+     * @return void
+     *
+     * @throws \InvalidArgumentException if configuration is invalid
+     */
+    protected function validateConfiguration(): void
+    {
+        // Validate Elasticsearch configuration if enabled
+        if (config('encrypted-search.elasticsearch.enabled', false)) {
+            $host = config('encrypted-search.elasticsearch.host');
+
+            if (empty($host)) {
+                throw new \InvalidArgumentException(
+                    'Elasticsearch is enabled but ELASTICSEARCH_HOST is not configured. ' .
+                    'Set it in your .env file or disable Elasticsearch mode.'
+                );
+            }
+
+            $index = config('encrypted-search.elasticsearch.index');
+
+            if (empty($index)) {
+                throw new \InvalidArgumentException(
+                    'Elasticsearch is enabled but ELASTICSEARCH_INDEX is not configured. ' .
+                    'Set it in your .env file.'
+                );
+            }
+        }
+    }
 }

src/EncryptedSearchServiceProvider.php

@@ -49,29 +49,41 @@ public function __construct(?string $host = null)
      * @param  string  $index  The Elasticsearch index name.
      * @param  string  $id  The unique document ID.
      * @param  array<string, mixed>  $body  The document body to be stored.
-     * @return bool  True if successful, false otherwise.
+     * @return void
+     *
+     * @throws \RuntimeException if the request fails
      */
-    public function indexDocument(string $index, string $id, array $body): bool
+    public function indexDocument(string $index, string $id, array $body): void
     {
         $url = "{$this->host}/{$index}/_doc/{$id}";
         $response = Http::put($url, $body);
 
-        return $response->successful();
+        if (!$response->successful()) {
+            throw new \RuntimeException(
+                "Failed to index document to Elasticsearch [{$url}]: " . $response->body()
+            );
+        }
     }
 
     /**
      * Delete a document from Elasticsearch by its ID.
      *
      * @param  string  $index  The Elasticsearch index name.
      * @param  string  $id  The document ID to delete.
-     * @return bool  True if successful, false otherwise.
+     * @return void
+     *
+     * @throws \RuntimeException if the request fails
      */
-    public function deleteDocument(string $index, string $id): bool
+    public function deleteDocument(string $index, string $id): void
     {
         $url = "{$this->host}/{$index}/_doc/{$id}";
         $response = Http::delete($url);
 
-        return $response->successful();
+        if (!$response->successful()) {
+            throw new \RuntimeException(
+                "Failed to delete document from Elasticsearch [{$url}]: " . $response->body()
+            );
+        }
     }
 
     /**
@@ -80,12 +92,35 @@ public function deleteDocument(string $index, string $id): bool
      * @param  string  $index  The Elasticsearch index name.
      * @param  array<string, mixed>  $query  The Elasticsearch query body.
      * @return array<int, mixed>  The array of matching documents (hits).
+     *
+     * @throws \RuntimeException if the request fails
      */
     public function search(string $index, array $query): array
     {
         $url = "{$this->host}/{$index}/_search";
         $response = Http::post($url, $query);
 
+        if (!$response->successful()) {
+            throw new \RuntimeException(
+                "Failed to search Elasticsearch [{$url}]: " . $response->body()
+            );
+        }
+
         return $response->json('hits.hits', []);
     }
+
+    /**
+     * Delete documents matching a query from an Elasticsearch index.
+     *
+     * @param  string  $index  The Elasticsearch index name.
+     * @param  array<string, mixed>  $query  The Elasticsearch query body.
+     * @return bool  True if successful, false otherwise.
+     */
+    public function deleteByQuery(string $index, array $query): bool
+    {
+        $url = "{$this->host}/{$index}/_delete_by_query";
+        $response = Http::post($url, $query);
+
+        return $response->successful();
+    }
 }

src/Services/ElasticsearchService.php

@@ -20,8 +20,11 @@
  *
  * Features:
  * - Lowercases all text (UTF-8 safe)
- * - Optionally removes diacritics using PHP’s Normalizer (if available)
+ * - Removes diacritics using PHP's intl extension (required)
  * - Strips all non-alphanumeric characters
+ *
+ * Requirements:
+ * - The intl PHP extension must be installed for consistent normalization
  */
 class Normalizer
 {
@@ -44,11 +47,9 @@ public static function normalize(?string $v): ?string
         // Convert to lowercase (UTF-8 safe)
         $s = mb_strtolower($v, 'UTF-8');
 
-        // Optionally remove diacritics if intl extension is available
-        if (class_exists(\Normalizer::class)) {
-            $s = \Normalizer::normalize($s, \Normalizer::FORM_D);
-            $s = preg_replace('/\p{M}/u', '', $s); // strip diacritics
-        }
+        // Remove diacritics using intl extension
+        $s = \Normalizer::normalize($s, \Normalizer::FORM_D);
+        $s = preg_replace('/\p{M}/u', '', $s); // strip diacritics
 
         // Retain only letters and digits
         $s = preg_replace('/[^a-z0-9]/u', '', $s);

src/Support/Normalizer.php

@@ -46,9 +46,18 @@ class Tokens
      *
      * @return string
      *     Hex-encoded SHA-256 hash (64 characters).
+     *
+     * @throws \RuntimeException if pepper is empty
      */
     public static function exact(string $normalized, string $pepper): string
     {
+        if (empty($pepper)) {
+            throw new \RuntimeException(
+                'SEARCH_PEPPER is not configured. Set it in your .env file for security. ' .
+                'Generate a random string: openssl rand -base64 32'
+            );
+        }
+
         return hash('sha256', $normalized . $pepper);
     }
 
@@ -60,25 +69,43 @@ public static function exact(string $normalized, string $pepper): string
      * These prefix hashes can be used to implement fast "starts-with"
      * queries while maintaining cryptographic privacy.
      *
-     * Example: "alex" with maxDepth=3 yields tokens for "a", "al", "ale".
+     * Only prefixes at or above the minimum length (from config) are generated.
+     * This prevents overly broad matches from very short search terms.
+     *
+     * Example: "alex" with maxDepth=4, minLength=2 yields tokens for "al", "ale", "alex".
+     * (skips "a" because it's below minimum length)
      *
      * @param string $normalized
      *     The normalized (lowercase, diacritic-free) string.
      * @param int $maxDepth
      *     The maximum number of prefix characters to hash.
      * @param string $pepper
      *     A secret application-level random string from configuration.
+     * @param int $minLength
+     *     The minimum prefix length to generate (default: 1 for backwards compatibility).
      *
      * @return string[]
      *     An array of hex-encoded SHA-256 prefix tokens.
+     *
+     * @throws \RuntimeException if pepper is empty
      */
-    public static function prefixes(string $normalized, int $maxDepth, string $pepper): array
+    public static function prefixes(string $normalized, int $maxDepth, string $pepper, int $minLength = 1): array
     {
+        if (empty($pepper)) {
+            throw new \RuntimeException(
+                'SEARCH_PEPPER is not configured. Set it in your .env file for security. ' .
+                'Generate a random string: openssl rand -base64 32'
+            );
+        }
+
         $out = [];
         $len = mb_strlen($normalized, 'UTF-8');
         $depth = min($maxDepth, $len);
 
-        for ($i = 1; $i <= $depth; $i++) {
+        // Start from minimum length instead of 1
+        $start = max(1, $minLength);
+
+        for ($i = $start; $i <= $depth; $i++) {
             $prefix = mb_substr($normalized, 0, $i, 'UTF-8');
             $out[] = hash('sha256', $prefix . $pepper);
         }