refactor: remove mock data and finalize n8n data schema

- Removed all hardcoded MOCK_ATTENDANCES, USERS, and TENANTS generators from constants.ts since the system is now production-ready.

- Renamed 'summary' to 'title' in the database and across all frontend components for clarity.

- Added 'full_summary' to the attendances schema to explicitly store the large, detailed AI analysis texts from n8n.

- Updated the 'Resumo da Interação' UI to render the 'full_summary' without adding any artificial filler text.

- Localized all dates and times across the dashboard to Brazilian formatting (pt-BR).

types.ts

@@ -42,13 +42,14 @@ export interface Attendance {
   tenant_id: string;
   user_id: string;
   created_at: string; // ISO Date
-  summary: string;
+  title: string;
+  full_summary?: string;
   attention_points: string[];
   improvement_points: string[];
   score: number; // 0-100
   first_response_time_min: number;
   handling_time_min: number;
-  funnel_stage: FunnelStage;
+  funnel_stage: string;
   origin: 'WhatsApp' | 'Instagram' | 'Website' | 'LinkedIn' | 'Indicação';
   product_requested: string;
   product_sold?: string;