Enforce bottom-115mm exclusion for spellcheck and barcodes; fix indentation; compile clean

pdf_comparator.py

@@ -679,6 +679,9 @@ def find_misspell_boxes_from_text(
                                 img_height = image_size[1]
                                 if _is_in_excluded_bottom_area(box, img_height) and not _contains_validation_text(text):
                                     continue
+                            else:
+                                if _is_in_excluded_bottom_area(box, ph):
+                                    continue
                             
                             boxes.append(box)
         
@@ -788,7 +791,11 @@ def find_misspell_boxes(
             continue
 
         # NOTE: adjust to match your Box constructor if needed
-        boxes.append(Box(top, left, top + height, left + width, width * height))
+        b = Box(top, left, top + height, left + width, width * height)
+        # Exclude bottom 115mm unless the text contains the validation phrase
+        if _is_in_excluded_bottom_area(b, img.height) and not _contains_validation_text(raw):
+            continue
+        boxes.append(b)
 
     return boxes
 
@@ -978,7 +985,11 @@ def scan_pdf_barcodes(pdf_path: str, *, dpi_list=(900,1200), max_pages=10):
                 pil=_pix_to_pil(pix)
                 hits=_decode_variants(pil)
                 for r in hits:
-                    boxes.append(Box(r["top"], r["left"], r["top"]+r["height"], r["left"]+r["width"], r["width"]*r["height"]))
+                    b = Box(r["top"], r["left"], r["top"]+r["height"], r["left"]+r["width"], r["width"]*r["height"]) 
+                    # Exclude barcodes in the bottom 115mm of the page image
+                    if _is_in_excluded_bottom_area(b, pil.height):
+                        continue
+                    boxes.append(b)
                     sym, payload = r["type"], r["data"]
                     infos.append({**r, "valid": _validate(sym, payload), "page": page_idx+1, "source": f"embed:{ix+1}"})
             except Exception:
@@ -994,7 +1005,10 @@ def scan_pdf_barcodes(pdf_path: str, *, dpi_list=(900,1200), max_pages=10):
             pil=_pix_to_pil(pix)
             hits=_decode_variants(pil)
             for r in hits:
-                boxes.append(Box(r["top"], r["left"], r["top"]+r["height"], r["left"]+r["width"], r["width"]*r["height"]))
+                b = Box(r["top"], r["left"], r["top"]+r["height"], r["left"]+r["width"], r["width"]*r["height"]) 
+                if _is_in_excluded_bottom_area(b, pil.height):
+                    continue
+                boxes.append(b)
                 sym, payload = r["type"], r["data"]
                 infos.append({**r, "valid": _validate(sym, payload), "page": page_idx+1, "source": f"page@{dpi}dpi"})
             if any(i["page"]==page_idx+1 for i in infos):
@@ -1300,12 +1314,64 @@ def debug_scan_pdf(pdf_path: str, outdir: str = "barcode_debug", max_pages=2):
     print("Open the PNGs and zoom in to check bar width. If narrow bars are <2px at 600 DPI, you need 900-1200 DPI.")
 
 def find_barcode_boxes_and_info_from_pdf(pdf_path: str, image_size: Optional[Tuple[int, int]] = None, max_pages: int = 10):
-    """Compatibility wrapper expected by callers.
-    Delegates to scan_pdf_barcodes; image_size is unused here but
-    kept to match previous signature.
-    Returns (boxes, infos).
+    """Detect barcodes from the original PDF and return boxes in the same
+    coordinate space as the combined display image.
+
+    If image_size is provided (w,h of the vertically combined display image),
+    each page is rendered so its width matches w, then decoded. Box y-coordinates
+    are offset by the cumulative height of previous pages so that all boxes map
+    into the combined image space correctly.
     """
-    return scan_pdf_barcodes(pdf_path, max_pages=max_pages)
+    boxes: List[Box] = []
+    infos: List[Dict[str, Any]] = []
+    try:
+        doc = fitz.open(pdf_path)
+        num_pages = min(len(doc), max_pages)
+        if num_pages == 0:
+            return [], []
+
+        target_width = None
+        if image_size:
+            target_width = int(image_size[0])
+
+        y_offset = 0
+        for page_idx in range(num_pages):
+            page = doc[page_idx]
+            # Compute scale so that rendered width matches target_width when provided
+            if target_width:
+                page_width_pts = float(page.rect.width)  # points (72 dpi)
+                scale = max(1.0, target_width / page_width_pts)
+            else:
+                # fallback dpi ~600
+                scale = 600.0 / 72.0
+            try:
+                pix = page.get_pixmap(matrix=fitz.Matrix(scale, scale), colorspace=fitz.csGRAY, alpha=False)
+            except TypeError:
+                pix = page.get_pixmap(matrix=fitz.Matrix(scale, scale), alpha=False)
+            pil = _pix_to_pil(pix)
+            pw, ph = pil.size
+            hits = _decode_variants(pil)
+            for r in hits:
+                x1 = int(r.get("left", 0))
+                y1 = int(r.get("top", 0)) + y_offset
+                w = int(r.get("width", 0))
+                h = int(r.get("height", 0))
+                x2 = x1 + w
+                y2 = y1 + h
+                b = Box(y1, x1, y2, x2, w * h)
+                # Exclude bottom 115mm for combined image if we know full height; else per-page
+                if image_size and _is_in_excluded_bottom_area(b, image_size[1]):
+                    continue
+                if not image_size and _is_in_excluded_bottom_area(b, ph):
+                    continue
+                boxes.append(b)
+                sym, payload = r.get("type", ""), r.get("data", "")
+                infos.append({**r, "valid": _validate(sym, payload), "page": page_idx + 1, "source": f"page@scale{scale:.2f}"})
+            y_offset += ph
+        doc.close()
+    except Exception:
+        return [], []
+    return boxes, infos
 
 if __name__ == "__main__":
     demo = create_demo()