feat: Add PWA support and consolidate CSS design system

- Add PWA manifest, icons (192x192, 512x512), and service worker
- Register service worker in index.html with Apple mobile web app support
- Consolidate CSS variables and design tokens documentation
- Update PrimeVue overrides for consistent theming
- Refactor data-entry components to use shared CSS patterns
- Add frontend-style-auditor agent for style consistency checks
- Minor OCR validation and job worker improvements
- Update start-prod.sh configuration

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/modules/data_entry/services/ocr/job_worker.py

@@ -353,9 +353,9 @@ async def start_job_worker() -> bool:
             logger.info("[JobWorker] Pre-warming OCR worker pool (background)...")
             warmup_success = await ocr_worker_pool.prewarm(timeout=90.0)
             if warmup_success:
-                logger.info("[JobWorker] ✅ OCR worker pool pre-warmed successfully")
+                logger.info("[JobWorker] OCR worker pool pre-warmed successfully")
             else:
-                logger.warning("[JobWorker] ⚠️ Worker pool pre-warm failed, first request will be slower")
+                logger.warning("[JobWorker] Worker pool pre-warm failed, first request will be slower")
 
         asyncio.create_task(_background_prewarm())
 

backend/modules/data_entry/services/ocr/validation.py

@@ -238,7 +238,7 @@ class PaymentSumRule(ValidationRule):
             return ValidationResult(
                 is_valid=False,
                 confidence_penalty=0.4,
-                message=f"Payment sum {payment_sum:.2f} RON ≠ Total {total:.2f} RON (diff: {diff:.2f} RON). Consider auto-correction.",
+                message=f"Payment sum {payment_sum:.2f} RON != Total {total:.2f} RON (diff: {diff:.2f} RON). Consider auto-correction.",
                 severity="error"
             )
 
@@ -299,7 +299,7 @@ class TVAEntriesSumRule(ValidationRule):
             return ValidationResult(
                 is_valid=False,
                 confidence_penalty=0.2,
-                message=f"TVA entries sum {entries_sum:.2f} RON ≠ TVA total {tva_total:.2f} RON (diff: {diff:.2f} RON)",
+                message=f"TVA entries sum {entries_sum:.2f} RON != TVA total {tva_total:.2f} RON (diff: {diff:.2f} RON)",
                 severity="warning"
             )
 
@@ -783,7 +783,7 @@ class OCRValidationEngine:
                 else:
                     warnings.append(msg)
 
-                print(f"  ❌ {msg}", flush=True)
+                print(f"  [X] {msg}", flush=True)
 
                 # Track confidence penalty for the relevant field based on rule
                 if result.confidence_penalty > 0:
@@ -797,7 +797,7 @@ class OCRValidationEngine:
                         if f in extraction_result:
                             confidence_adjustments[f] = result.confidence_penalty
             else:
-                print(f"  ✅ {rule.rule_name}: {result.message}", flush=True)
+                print(f"  [OK] {rule.rule_name}: {result.message}", flush=True)
 
         # Step 2: Cross-field validation
         print("\n[Validation] Step 2: Cross-field validation...", flush=True)
@@ -812,7 +812,7 @@ class OCRValidationEngine:
                 else:
                     warnings.append(msg)
 
-                print(f"  ❌ {msg}", flush=True)
+                print(f"  [X] {msg}", flush=True)
 
                 # Track confidence penalty for the relevant field based on rule
                 if result.confidence_penalty > 0:
@@ -826,7 +826,7 @@ class OCRValidationEngine:
                         if f in extraction_result:
                             confidence_adjustments[f] = result.confidence_penalty
             else:
-                print(f"  ✅ {rule.rule_name}: {result.message}", flush=True)
+                print(f"  [OK] {rule.rule_name}: {result.message}", flush=True)
 
         # Step 3: Inter-OCR consistency checks
         if light_result and medium_result:
@@ -845,7 +845,7 @@ class OCRValidationEngine:
                 if not result.is_valid:
                     msg = f"[Inter-OCR] {result.message}"
                     warnings.append(msg)
-                    print(f"  ❌ {msg}", flush=True)
+                    print(f"  [X] {msg}", flush=True)
 
                     # Store ratio for metadata
                     ratio = max(
@@ -854,7 +854,7 @@ class OCRValidationEngine:
                     ) / min(light_result["amount"], medium_result["amount"])
                     inter_ocr_ratios["amount"] = ratio
                 else:
-                    print(f"  ✅ {result.message}", flush=True)
+                    print(f"  [OK] {result.message}", flush=True)
 
         # Determine if manual review is needed
         # Only flag for review if there are errors OR high-severity warnings

backend/modules/data_entry/services/ocr_service.py

@@ -124,7 +124,7 @@ class OCRService:
 
         # Check if memory is high - force GC before processing
         if mem_before > self._memory_threshold_mb:
-            print(f"[OCR Service] ⚠️ Memory high ({mem_before:.0f}MB > {self._memory_threshold_mb}MB), forcing GC...", flush=True)
+            print(f"[OCR Service] WARNING: Memory high ({mem_before:.0f}MB > {self._memory_threshold_mb}MB), forcing GC...", flush=True)
             gc.collect()
             mem_after_gc = get_memory_usage_mb()
             print(f"[OCR Service] Memory after pre-GC: {mem_after_gc:.0f}MB", flush=True)
@@ -188,7 +188,7 @@ class OCRService:
                     extraction.raw_text = "\n\n".join(raw_texts)
                     elapsed_ms = int((time.time() - start_time) * 1000)
                     extraction.processing_time_ms = elapsed_ms
-                    print(f"[OCR] ✓✓✓ EARLY EXIT at Step 1 - All fields found! ({elapsed_ms}ms) ✓✓✓", flush=True)
+                    print(f"[OCR] *** EARLY EXIT at Step 1 - All fields found! ({elapsed_ms}ms) ***", flush=True)
                     # Cleanup before return
                     del image
                     if images:
@@ -244,7 +244,7 @@ class OCRService:
                     extraction.ocr_engine = "paddle-adaptive"
                     elapsed_ms = int((time.time() - start_time) * 1000)
                     extraction.processing_time_ms = elapsed_ms
-                    print(f"[OCR] ✓✓✓ EARLY EXIT at Step 2 - All fields found after merge! ({elapsed_ms}ms) ✓✓✓", flush=True)
+                    print(f"[OCR] *** EARLY EXIT at Step 2 - All fields found after merge! ({elapsed_ms}ms) ***", flush=True)
                     # Cleanup before return
                     del image
                     if images:
@@ -399,7 +399,7 @@ class OCRService:
         print(f"  - Needs Manual Review: {extraction.needs_manual_review}", flush=True)
         if extraction.validation_warnings:
             for warning in extraction.validation_warnings:
-                print(f"    ⚠️  {warning}", flush=True)
+                print(f"    [!] {warning}", flush=True)
 
         return True, message, extraction
 
@@ -611,7 +611,7 @@ class OCRService:
             print(f"[OCR] Missing: {', '.join(missing)} - continuing", flush=True)
             return False
 
-        print(f"[OCR] ✓ All 5 fields found with {ext.overall_confidence:.0%} confidence", flush=True)
+        print(f"[OCR] OK: All 5 fields found with {ext.overall_confidence:.0%} confidence", flush=True)
         return True
 
     def _complement_extraction(

backend/modules/reports/cache/event_monitor.py

@@ -157,7 +157,7 @@ class EventMonitor:
                 # Changes detected!
                 logger.info(
                     f"Schema {schema} (company {company_id}): "
-                    f"id_act changed {cached_watermark} → {current_max}"
+                    f"id_act changed {cached_watermark} -> {current_max}"
                 )
 
                 # Update watermark