feat(classifier): count_placeholder_fields() helper for drop-logic

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

includes/class-classifier.php

@@ -46,6 +46,14 @@ final class Classifier
         'rambler.ru', 'gmail.com',
     ];
 
+    /** Service field keys excluded from placeholder count. */
+    private const SERVICE_FIELD_KEYS = [
+        'acceptance-data',
+        '_wpcf7', '_wpcf7_version', '_wpcf7_locale', '_wpcf7_unit_tag',
+        '_wpcf7_container_post', '_wpcf7_posted_data_hash',
+        'g-recaptcha-response',
+    ];
+
     /**
      * @param array<string,mixed> $fields CF7 posted_data (field => value or array of values)
      * @param array{reason?:string} $meta  Context: reason = Akismet / Honeypot / CF7 rules / Неизвестно
@@ -266,4 +274,32 @@ final class Classifier
         if ($name === '') return false;
         return (bool)preg_match('/^[A-Za-z0-9]{1,3}$/', $name);
     }
+
+    /**
+     * Counts non-empty fields whose value is exactly 1-2 ASCII alphanumeric chars
+     * (the typical "AB", "70" placeholder pattern bots fill custom-keyed fields with).
+     *
+     * Skips service keys (acceptance-data, _wpcf7*, g-recaptcha-response) and any key
+     * starting with `_wpcf7`. Array values are joined without separators before length
+     * check (so `['A','B']` collapses to "AB" — still counted).
+     *
+     * @param array<string,string|string[]|null> $fields CF7 posted_data
+     * @return int Number of fields matching the placeholder pattern
+     */
+    public static function count_placeholder_fields(array $fields): int
+    {
+        $count = 0;
+        foreach ($fields as $k => $v) {
+            if (is_string($k)) {
+                if (in_array($k, self::SERVICE_FIELD_KEYS, true)) continue;
+                if (strpos($k, '_wpcf7') === 0) continue;
+            }
+            if (is_array($v)) $v = implode('', array_map('strval', $v));
+            $v = trim((string)$v);
+            if ($v === '') continue;
+            if (!preg_match('/^[A-Za-z0-9]{1,2}$/', $v)) continue;
+            $count++;
+        }
+        return $count;
+    }
 }

tests/ClassifierDropTest.php

@@ -102,4 +102,43 @@ final class ClassifierDropTest extends Cf7stgTestCase
         $fields = ['your-name' => ['OB', 'MU']];
         self::assertFalse(Classifier::has_short_latin_name($fields));
     }
+
+    public function test_count_placeholder_fields_screenshot_1(): void
+    {
+        $fields = require __DIR__ . '/fixtures/spam-screenshot-1.php';
+        // text-785: OB(2), text-787: 70(2), text-786: BP(2), text-788: ZZ(2), text-789: QV(2)
+        // text-790: 10 цифр — не placeholder. acceptance-data: служебное → исключено.
+        self::assertGreaterThanOrEqual(3, Classifier::count_placeholder_fields($fields));
+    }
+
+    public function test_count_placeholder_fields_ignores_empty(): void
+    {
+        $fields = ['a' => 'AB', 'b' => '', 'c' => 'XY', 'd' => null, 'e' => 'CD'];
+        self::assertSame(3, Classifier::count_placeholder_fields($fields));
+    }
+
+    public function test_count_placeholder_fields_ignores_long_values(): void
+    {
+        $fields = ['a' => 'AB', 'b' => 'Hello world', 'c' => 'CD'];
+        self::assertSame(2, Classifier::count_placeholder_fields($fields));
+    }
+
+    public function test_count_placeholder_fields_ignores_service_keys(): void
+    {
+        $fields = [
+            'acceptance-data' => '1',
+            '_wpcf7' => 'AB',
+            '_wpcf7_unit_tag' => 'XY',
+            'g-recaptcha-response' => 'CD',
+            'real' => 'EF',
+        ];
+        self::assertSame(1, Classifier::count_placeholder_fields($fields));
+    }
+
+    public function test_count_placeholder_fields_ignores_non_alnum(): void
+    {
+        // Кириллица или с пробелами не считается placeholder
+        $fields = ['a' => 'Ия', 'b' => 'a b', 'c' => 'XY'];
+        self::assertSame(1, Classifier::count_placeholder_fields($fields));
+    }
 }