Grouping : smaller groups, but be smarter

annotating : fix sorting for good

annotating.py

@@ -381,6 +381,8 @@ def concat_display_image(subdir):
 
     if not images:
         return
+    images.sort(key=lambda f: [int(n) for n in re.findall(r'\d+', str(f))])
+
 
     # Load images
     opened_imgs = [Image.open(img) for img in images]
@@ -402,41 +404,6 @@ def concat_display_image(subdir):
     print(f"Saved: {save_path}")
     # subprocess.call(('xdg-open', save_path))
 
-def concat_anot_images(directory):
-    root = Path(directory)
-
-    for subdir in root.iterdir():
-        if subdir.is_dir() and subdir.name.startswith("Anot"):
-            # Find valid images, excluding previous concatenations
-            images = sorted([
-                f for f in subdir.glob("*.jpg")
-                if f.name != "Concat.jpg"
-            ])
-
-            images.sort(key=lambda f: [int(n) for n in re.findall(r'\d+', f)])
-
-            if not images:
-                continue
-
-            # Load images
-            opened_imgs = [Image.open(img) for img in images]
-
-            # Calculate dimensions (max width, sum of heights)
-            max_w = max(i.width for i in opened_imgs)
-            total_h = sum(i.height for i in opened_imgs)
-
-            # Create canvas and paste vertically
-            canvas = Image.new('RGB', (max_w, total_h))
-            current_y = 0
-            for img in opened_imgs:
-                canvas.paste(img, (0, current_y))
-                current_y += img.height
-
-            # Save
-            save_path = subdir / "Concat.jpg"
-            canvas.save(save_path)
-            print(f"Saved: {save_path}")
-            subprocess.call(('xdg-open', save_path))
 
 
 if len(sys.argv) < 2:

grouping.py

@@ -11,7 +11,7 @@ from pdf2image import convert_from_path, pdfinfo_from_path
 DPI = 200  # Good balance for readability and size
 A4_HEIGHT_INCHES = 11.69
 FULL_PAGE_PX = int(A4_HEIGHT_INCHES * DPI)
-MAX_GROUP_HEIGHT = 2.5 * FULL_PAGE_PX
+MAX_GROUP_HEIGHT = 2. * FULL_PAGE_PX
 MAX_GROUP_COUNT = 15
 SEPARATOR_HEIGHT = 20
 LABEL_HEIGHT = 50
@@ -79,38 +79,79 @@ def collect_files(root_dir):
     return data
 
 def group_files(file_list):
-    """Groups files based on constraints."""
-    sorted_files = sorted(file_list, key=lambda x: x[0])
+    """
+    Groups files using First Fit Decreasing algorithm to minimize group count.
+    """
+    # 1. Sort by height DESCENDING. Large items are hardest to fit, handle them first.
+    #    (Remove this sort if you must strictly preserve input order logic)
+    sorted_files = sorted(file_list, key=lambda x: x[2], reverse=True)
 
+    # Each group is a dict: {'items': [], 'current_height': 0}
     groups = []
-    current_group = []
-    current_height = 0
 
     for item in sorted_files:
         dd, path, height = item
+        placed = False
 
-        # Calculate added height (image + separator + approx text space)
-        # We add separator height only if it's not the first image
-        added_overhead = SEPARATOR_HEIGHT + 30 if current_group else 0
+        # 2. Try to fit item into an existing group (First Fit)
+        for group in groups:
+            # Check Count Constraint
+            if len(group['items']) >= MAX_GROUP_COUNT:
+                continue
 
-        # Check conditions
-        if (len(current_group) >= MAX_GROUP_COUNT or
-            (current_height + height + added_overhead) > MAX_GROUP_HEIGHT):
+            # Calculate Overhead (only if group is not empty)
+            overhead = (SEPARATOR_HEIGHT + 30) if group['items'] else 0
 
-            # Push current group and start new
-            if current_group:
-                groups.append(current_group)
-            current_group = []
-            current_height = 0
-            added_overhead = 0 # Reset for first file of new group
+            # Check Height Constraint
+            if (group['current_height'] + height + overhead) <= MAX_GROUP_HEIGHT:
+                group['items'].append(item)
+                group['current_height'] += height + overhead
+                placed = True
+                break
 
-        current_group.append(item)
-        current_height += height + added_overhead
+        # 3. If it doesn't fit anywhere, create a new group
+        if not placed:
+            groups.append({
+                'items': [item],
+                'current_height': height
+            })
 
-    if current_group:
-        groups.append(current_group)
+    # Return list of lists (strip the metadata)
+    return [g['items'] for g in groups]
 
-    return groups
+# def group_files(file_list):
+#     """Groups files based on constraints."""
+#     sorted_files = sorted(file_list, key=lambda x: x[0])
+
+#     groups = []
+#     current_group = []
+#     current_height = 0
+
+#     for item in sorted_files:
+#         dd, path, height = item
+
+#         # Calculate added height (image + separator + approx text space)
+#         # We add separator height only if it's not the first image
+#         added_overhead = SEPARATOR_HEIGHT + 30 if current_group else 0
+
+#         # Check conditions
+#         if (len(current_group) >= MAX_GROUP_COUNT or
+#             (current_height + height + added_overhead) > MAX_GROUP_HEIGHT):
+
+#             # Push current group and start new
+#             if current_group:
+#                 groups.append(current_group)
+#             current_group = []
+#             current_height = 0
+#             added_overhead = 0 # Reset for first file of new group
+
+#         current_group.append(item)
+#         current_height += height + added_overhead
+
+#     if current_group:
+#         groups.append(current_group)
+
+#     return groups
 
 def stitch_pdf_pages(images_list):
     """Vertically concatenates a list of PIL images with no separator."""