feat(ocr): Add validation system and CLIENT CUI extraction

OCR Data Extraction Validation System:
- Add 7 validation rules (amount range, TVA ratio, payment sum, etc.)
- Add Medium preprocessing to replace Heavy (fixes digit concatenation)
- Add validation warnings to API responses
- Flag receipts needing manual review (needs_manual_review field)
- Add database migration for needs_manual_review column

CLIENT CUI Extraction Improvements:
- Support all format variations: CIF CLIENT:, CLIENT C.U.I/C.I.F., etc.
- Handle OCR errors (R0 vs RO, C1F vs CIF)
- Add client_name, client_cui, client_address to API response
- Add validation fields to API response (was missing)

QA Review: 12 issues found, 9 fixed (5 errors + 4 warnings)
- Fixed type safety in validation rules
- Fixed ZeroDivisionError risk
- Fixed schema mismatch (Optional[bool] for needs_manual_review)
- All 37 unit tests passing

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

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

backend/modules/data_entry/services/image_preprocessor.py

@@ -104,10 +104,80 @@ class ImagePreprocessor:
         # NO binarization, NO morphological ops - preserve original quality
         return enhanced
 
+    def preprocess_medium(self, image: np.ndarray) -> np.ndarray:
+        """
+        Medium preprocessing for MIXED-QUALITY images.
+        Balance between Light (too gentle) and Heavy (too aggressive).
+
+        Use cases:
+        - Moderately faded receipts
+        - Photos with uneven lighting
+        - Scans with slight blur
+
+        Preprocessing steps:
+        - Moderate contrast enhancement (CLAHE clipLimit=2.0)
+        - Light denoising (fastNlMeansDenoising h=6)
+        - Gentle sharpening
+        - NO binarization (preserves text boundaries)
+        - NO morphological operations (avoids digit concatenation)
+
+        This method was created to replace preprocess_heavy() which caused
+        digit concatenation errors on high-quality PDFs (85.99 → 859,762.16).
+        """
+        # 0. Add safety padding to protect edge content during deskew rotation
+        image = self._add_safety_padding(image)
+
+        # 1. Grayscale
+        if len(image.shape) == 3:
+            gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+        else:
+            gray = image.copy()
+
+        # 2a. Scale DOWN if any side exceeds 4000px (PaddleOCR limit)
+        height, width = gray.shape
+        max_side = max(height, width)
+        if max_side > 4000:
+            scale = 4000 / max_side
+            gray = cv2.resize(gray, None, fx=scale, fy=scale, interpolation=cv2.INTER_AREA)
+            height, width = gray.shape
+
+        # 2b. Scale UP if too small
+        if width < 1500:
+            scale = 1500 / width
+            # Ensure we don't exceed 4000px after upscaling
+            new_width = int(width * scale)
+            new_height = int(height * scale)
+            if max(new_width, new_height) > 4000:
+                scale = 4000 / max(new_width, new_height)
+            gray = cv2.resize(gray, None, fx=scale, fy=scale, interpolation=cv2.INTER_CUBIC)
+
+        # 3. Deskew
+        gray = self._deskew(gray)
+
+        # 4. Moderate contrast enhancement (CLAHE clipLimit=2.0)
+        clahe = cv2.createCLAHE(clipLimit=2.0, tileGridSize=(8, 8))
+        enhanced = clahe.apply(gray)
+
+        # 5. Light denoising (less aggressive than Heavy)
+        denoised = cv2.fastNlMeansDenoising(enhanced, h=6, templateWindowSize=7, searchWindowSize=15)
+
+        # 6. Gentle sharpening
+        gaussian = cv2.GaussianBlur(denoised, (0, 0), 1.0)
+        sharpened = cv2.addWeighted(denoised, 1.3, gaussian, -0.3, 0)
+
+        # NO binarization, NO morphological operations
+        # This preserves text boundaries and avoids digit concatenation
+        return sharpened
+
     def preprocess_heavy(self, image: np.ndarray) -> np.ndarray:
         """
         Heavy preprocessing for FADED thermal receipts.
         Aggressive binarization to recover faded text.
+
+        ⚠️ DEPRECATED: Use preprocess_medium() instead.
+        Heavy preprocessing causes digit concatenation on clear PDFs
+        (e.g., 85.99 → 859,762.16 due to binarization + morphological operations).
+        Kept for backward compatibility only.
         """
         # 0. Add safety padding to protect edge content during deskew rotation
         image = self._add_safety_padding(image)