detect added

comic_panel_extractor/annorator_server.py

@@ -55,11 +55,22 @@ def get_label_path(image_name: str) -> str:
     return os.path.join(LABEL_ROOT, os.path.splitext(image_name)[0] + ".txt")
 
 # === Core Functions ===
-def load_yolo_boxes(image_path: str, label_path: str):
+def load_yolo_boxes(image_path: str, label_path: str, detect: bool = False):
     try:
         img = Image.open(image_path)
         w, h = img.size
         boxes = []
+        if detect and not os.path.exists(label_path):
+            from .yolo_manager import YOLOManager
+            from .utils import Config
+            yolo_manager = YOLOManager()
+            weights_path = f'{current_path}/{Config.YOLO_MODEL_NAME}.pt'
+
+            yolo_manager.load_model(weights_path)
+
+            # Run inference
+            _, label_path = yolo_manager.annotate_images(image_paths=[image_path], output_dir=IMAGE_LABEL_ROOT, save_image=False, label_path=label_path)
+
         if os.path.exists(label_path):
             with open(label_path, "r") as f:
                 for line in f:
@@ -160,6 +171,21 @@ async def get_annotations(image_name: str):
         "original_height": height
     }
 
+@app.get("/api/annotate/detect_annotations/{image_name:path}")
+async def get_annotations(image_name: str):
+    image_path = get_image_path(image_name)
+    label_path = get_label_path(image_name)
+
+    if not os.path.exists(image_path):
+        raise HTTPException(status_code=404, detail="Image not found")
+
+    boxes, (width, height) = load_yolo_boxes(image_path, label_path, True)
+    return {
+        "boxes": boxes,
+        "original_width": width,
+        "original_height": height
+    }
+
 @app.post("/api/annotate/annotations")
 async def save_annotations(request: SaveAnnotationsRequest):
     label_path = get_label_path(request.image_name)

comic_panel_extractor/static/annotator.html

@@ -547,6 +547,9 @@
                 <button class="btn btn-primary btn-block" id="reloadBtn">
                     🔄 Reload Annotations
                 </button>
+                <button class="btn btn-primary btn-block" id="detectBtn">
+                    🔄 Detect Annotations
+                </button>
                 <button class="btn btn-secondary btn-block" id="downloadBtn" style="display: none; margin-top: 8px;">
                     📥 Download
                 </button>
@@ -661,6 +664,7 @@
                 document.getElementById('undoBtn').addEventListener('click', () => this.undoLastBox());
                 document.getElementById('clearBtn').addEventListener('click', () => this.clearAllBoxes());
                 document.getElementById('reloadBtn').addEventListener('click', () => this.reloadAnnotations());
+                document.getElementById('detectBtn').addEventListener('click', () => this.detectAnnotations());
                 document.getElementById('downloadBtn').addEventListener('click', () => this.downloadAnnotations());
 
                 // Canvas events
@@ -1369,6 +1373,27 @@
                 }
             }
 
+            async detectAnnotations() {
+                if (!this.currentImage) return;
+
+                try {
+                    const response = await fetch(`/api/annotate/detect_annotations/${encodeURIComponent(this.currentImage)}`);
+                    const data = await response.json();
+
+                    this.boxes = (data.boxes || []).map(box => ({
+                        ...box,
+                        saved: true
+                    }));
+                    this.selectedBoxIndex = -1;
+                    document.getElementById('boxCount').textContent = this.boxes.length;
+                    this.updateSelectedBoxInfo();
+                    this.drawCanvas();
+                    this.showAlert('Annotations reloaded from file', 'success');
+                } catch (error) {
+                    this.showAlert('Error reloading annotations: ' + error.message, 'error');
+                }
+            }
+
             downloadAnnotations() {
                 if (!this.currentImage) return;
 

comic_panel_extractor/yolo_manager.py

@@ -66,7 +66,7 @@ import os
 import cv2
 from ultralytics import YOLO
 from typing import List, Optional, Dict, Any
-from utils import Config, get_abs_path, clean_directory
+from .utils import Config, get_abs_path, clean_directory
 
 class YOLOManager:
     """Manages YOLO model training and inference operations."""
@@ -153,13 +153,13 @@ class YOLOManager:
         
         return weights_path
     
-    def annotate_images(self, image_paths: List[str], output_dir: str = 'temp_dir', image_size: int = None) -> None:
+    def annotate_images(self, image_paths: List[str], output_dir: str = 'temp_dir', image_size: int = None, save_image: bool = True, label_path: str = None) -> None:
         """
-        Annotate images with model predictions.
+        Annotate images with model predictions and save YOLO-format label files.
         
         Args:
             image_paths: List of image file paths
-            output_dir: Directory to save annotated images
+            output_dir: Directory to save annotated images and labels
             image_size: Size for inference
         """
         if not self.model:
@@ -169,30 +169,63 @@ class YOLOManager:
             raise ValueError("❌ No images provided for annotation.")
         
         image_size = image_size or Config.DEFAULT_IMAGE_SIZE
-        clean_directory(output_dir)
-        
-        print(f"🎨 Annotating {len(image_paths)} images...")
+        # clean_directory(output_dir)
+        totla_images = len(image_paths)
+        print(f"🎨 Annotating {totla_images} images and saving labels...")
         
         for idx, image_path in enumerate(image_paths):
             if not os.path.isfile(image_path):
                 print(f"⚠️ Warning: Skipping non-existent file {image_path}")
                 continue
-                
+            
             print(f'🔍 Processing ({idx+1}/{len(image_paths)}): {os.path.basename(image_path)}')
             
             try:
+                # Load image for size info
+                img = cv2.imread(image_path)
+                h, w = img.shape[:2]
+                
+                # Run inference
                 results = self.model(image_path, imgsz=image_size)
                 annotated_frame = results[0].plot()
                 
-                # Use original filename with prefix
+                # Prepare save paths
                 original_name = os.path.basename(image_path)
                 name, ext = os.path.splitext(original_name)
-                save_path = os.path.join(output_dir, f'annotated_{name}{ext}')
-                
-                cv2.imwrite(save_path, annotated_frame)
-                print(f'✅ Saved: {save_path}')
+
+                save_img_path = None
+                save_txt_path = os.path.join(output_dir, f'{name}.txt')  # YOLO label txt
+                if save_image:
+                    save_img_path = os.path.join(output_dir, f'annotated_{name}{ext}')
+                    # Save annotated image
+                    cv2.imwrite(save_img_path, annotated_frame)
+
+                # Write YOLO label file
+                with open(save_txt_path, 'w') as f:
+                    for box in results[0].boxes:
+                        # box.xyxy format: (xmin, ymin, xmax, ymax)
+                        xyxy = box.xyxy[0].tolist()
+                        cls_id = int(box.cls[0].item())  # class id
+
+                        xmin, ymin, xmax, ymax = xyxy
+                        # Convert to YOLO format (normalized)
+                        x_center = ((xmin + xmax) / 2) / w
+                        y_center = ((ymin + ymax) / 2) / h
+                        width = (xmax - xmin) / w
+                        height = (ymax - ymin) / h
+
+                        # Write one line per object
+                        f.write(f"{cls_id} {x_center:.6f} {y_center:.6f} {width:.6f} {height:.6f}\n")
+
+                if label_path:
+                    shutil.copyfile(save_txt_path, label_path)
+                print(f'✅ Saved annotated image: {save_img_path}')
+                print(f'✅ Saved label file: {save_txt_path}')
+                print(f"🎉 Annotation and label saving complete! Results saved to: {output_dir}")
+
+                if totla_images == 1:
+                    return save_img_path, save_txt_path
                 
             except Exception as e:
                 print(f"❌ Error processing {image_path}: {str(e)}")
-        
-        print(f"🎉 Annotation complete! Results saved to: {output_dir}")
+                return None, None