feat(medsam): follow original behavior; remove SAM auto from inference path; MedSAM boxes only (max 8)

app.py

@@ -692,52 +692,18 @@ def analyze_with_medsam(base_result, image):
 
         # AUTO segmentation path
         try:
-            from segment_anything import sam_model_registry, SamAutomaticMaskGenerator
-            import cv2 as _cv2
-            # If ViT-H checkpoint present, use SAM automatic mask generator (download if missing)
-            vit_h_ckpt = os.path.join(HF_CACHE_DIR, "sam_vit_h_4b8939.pth")
-            if not os.path.exists(vit_h_ckpt):
-                try:
-                    from huggingface_hub import hf_hub_download
-                    vit_h_ckpt = hf_hub_download(
-                        repo_id="Aniketg6/SAM",
-                        filename="sam_vit_h_4b8939.pth",
-                        cache_dir=HF_CACHE_DIR
-                    )
-                    print(f"✅ Downloaded SAM ViT-H checkpoint to: {vit_h_ckpt}")
-                except Exception as dlh:
-                    print(f"⚠ Failed to download SAM ViT-H checkpoint: {dlh}")
-            if os.path.exists(vit_h_ckpt):
-                img_bgr = _cv2.imread(img_path)
-                sam = sam_model_registry["vit_h"](checkpoint=vit_h_ckpt)
-                mask_generator = SamAutomaticMaskGenerator(sam)
-                masks = mask_generator.generate(img_bgr)
-                # Keep top-12 masks by stability_score
-                masks = sorted(masks, key=lambda m: m.get('stability_score', 0), reverse=True)[:12]
-                for m in masks:
-                    seg = m.get('segmentation', None)
-                    if seg is None:
-                        continue
-                    seg_u8 = seg.astype(np.uint8)
-                    segmentations.append({
-                        "mask": seg_u8.tolist(),
-                        "confidence": float(m.get('stability_score', 1.0)),
-                        "method": "sam_auto"
-                    })
-                    masks_for_overlay.append({"mask": seg_u8})
-            else:
-                # Fallback: derive candidate boxes and run MedSAM per box
-                cand_bboxes = _find_topk_foreground_bboxes(pil_img, max_regions=8, min_area=200)
-                for bbox in cand_bboxes:
-                    m = _medsam.segment_with_box(bbox)
-                    if m is None or not isinstance(m.get('mask'), np.ndarray):
-                        continue
-                    segmentations.append({
-                        "mask": m['mask'].astype(np.uint8).tolist(),
-                        "confidence": float(m.get('confidence', 1.0)),
-                        "method": m.get("method", "medsam_box_auto")
-                    })
-                    masks_for_overlay.append(m)
+            # Follow original behavior: use MedSAM with box prompts; no SAM auto in main path
+            cand_bboxes = _find_topk_foreground_bboxes(pil_img, max_regions=8, min_area=200)
+            for bbox in cand_bboxes:
+                m = _medsam.segment_with_box(bbox)
+                if m is None or not isinstance(m.get('mask'), np.ndarray):
+                    continue
+                segmentations.append({
+                    "mask": m['mask'].astype(np.uint8).tolist(),
+                    "confidence": float(m.get('confidence', 1.0)),
+                    "method": m.get("method", "medsam_box_auto")
+                })
+                masks_for_overlay.append(m)
         except Exception as auto_e:
             print(f"Automatic MedSAM segmentation failed: {auto_e}")