feat: add AI recommendations flow for tenders

Implement company AI onboarding/recommendation models, services, repository, and tender UI integration with supporting tests.

Co-authored-by: Cursor <cursoragent@cursor.com>

lib/core/constants/pref_keys.dart

@@ -9,4 +9,13 @@ class PrefKeys {
   static const String bearer = 'bearer';
   static const String refreshToken = 'refresh_token';
   static const String customerData = 'customer_data';
+
+  /// Fingerprint (hash of the company's document ids + website) of the last
+  /// profile version we triggered AI onboarding for. Lets us skip redundant
+  /// `POST /api/v1/onboarding` calls until the company's data actually changes.
+  static const String onboardingFingerprint = 'ai_onboarding_fingerprint';
+
+  /// Epoch milliseconds when AI onboarding was last started. Used by the
+  /// Recommended tab to show a "still learning" hint for a short grace period.
+  static const String onboardingStartedAt = 'ai_onboarding_started_at';
 }