Video Uploading Default

app.py

@@ -128,6 +128,30 @@ CONFIG = {
     "DEVICE"           : "cuda:0" if torch.cuda.is_available() else "cpu",
 }
 
+# ── Demo videos (hosted in HF Hub model repo alongside checkpoints) ──────────
+# Upload your two XCA demo videos to MuhammadAdil63/angio-ai-checkpoints as:
+#   demo_video_1.mp4
+#   demo_video_2.mp4
+# The buttons below download and load them automatically.
+DEMO_VIDEO_1_NAME = "demo_video_1.mp4"
+DEMO_VIDEO_2_NAME = "demo_video_2.mp4"
+
+def _get_demo_video(filename: str) -> str:
+    """Download demo video from HF Hub and return local path."""
+    try:
+        from huggingface_hub import hf_hub_download
+        path = hf_hub_download(
+            repo_id  = "MuhammadAdil63/angio-ai-checkpoints",
+            filename = filename,
+            repo_type= "model",
+            token    = os.environ.get("HF_TOKEN"),
+        )
+        return path
+    except Exception as e:
+        print(f"[DEMO] Could not load {filename}: {e}")
+        return None
+
+
 # ─────────────────────────────────────────────────────────────────────────────
 #  ITU LOGO — encode to base64 for embedding in HTML
 #  Place ITULog.png in the same directory as this script.
@@ -748,15 +772,15 @@ def _load_yolo():
             f"YOLOv8 checkpoint not found:\n{ckpt}\n"
             f"Check HF_TOKEN secret and repo MuhammadAdil63/angio-ai-checkpoints"
         )
-    print(f"[YOLO] Loading checkpoint: {ckpt}")
+    print(f"[SEG] Loading checkpoint: {ckpt}")
     try:
         import torch.serialization
         from ultralytics.nn.tasks import SegmentationModel
         torch.serialization.add_safe_globals([SegmentationModel])
     except Exception as _e:
-        print(f"[YOLO] safe_globals warning: {_e}")
+        print(f"[SEG] safe_globals warning: {_e}")
     _yolo_model = YOLO(ckpt)
-    print(f"[YOLO] Model loaded OK")
+    print(f"[SEG] Model loaded OK")
 
 
 def run_yolo_seg(frame_rgb: np.ndarray, seg_conf: float) -> np.ndarray:
@@ -793,7 +817,7 @@ def run_yolo_seg(frame_rgb: np.ndarray, seg_conf: float) -> np.ndarray:
             verbose      = False,
         )
     except Exception as e:
-        print(f"[YOLO] predict() failed: {e}")
+        print(f"[SEG] predict() failed: {e}")
         return np.zeros((512, 512), dtype=np.uint8)
 
     result  = results[0]
@@ -801,7 +825,7 @@ def run_yolo_seg(frame_rgb: np.ndarray, seg_conf: float) -> np.ndarray:
     seg_map = np.zeros((h, w), dtype=np.uint8)
 
     if result.masks is None or result.boxes is None:
-        print("[YOLO] No masks returned -- check conf threshold or checkpoint.")
+        print("[SEG] No masks returned -- check conf threshold or checkpoint.")
         return seg_map
 
     masks   = result.masks.data.cpu().numpy()       # (N, Hm, Wm) float32
@@ -823,9 +847,9 @@ def run_yolo_seg(frame_rgb: np.ndarray, seg_conf: float) -> np.ndarray:
         n_written += int(binary.sum())
 
     detected_cls = [_YOLO_CLASS_NAMES[min(int(c) + 1, 25)] for c in cls_ids]
-    print(f"[YOLO] Detections: {len(cls_ids)}  |  classes: {detected_cls}  |  "
+    print(f"[SEG] Detections: {len(cls_ids)}  |  classes: {detected_cls}  |  "
           f"vessel px: {n_written} ({100 * n_written / (h * w):.1f}%)")
-    print(f"[YOLO] seg_map unique labels: {np.unique(seg_map)}")
+    print(f"[SEG] seg_map unique labels: {np.unique(seg_map)}")
     return seg_map
 
 # ─────────────────────────────────────────────────────────────────────────────
@@ -1510,8 +1534,10 @@ with gr.Blocks(css=CSS, title="Angio AI") as demo:
                     video_input = gr.Video(
                         label="Upload XCA video (mp4 / avi / dicom-wrapped)",
                         elem_id="upload-zone",
-                        height=240,
+                        height=220,
                     )
+                    btn_demo1 = gr.Button("▶  XCA Video Run", variant="secondary", size="sm")
+                    gr.HTML('<div style="font-size:11px;color:#7a9ab6;margin:4px 0 8px 2px">Click to load the demo XCA video, or upload your own above.</div>')
                     gr.HTML('</div>')
 
                     gr.HTML('<div class="card"><div class="card-title">Model controls</div>')
@@ -1540,7 +1566,7 @@ with gr.Blocks(css=CSS, title="Angio AI") as demo:
   <ol style="font-size:13px;color:#1a2533;line-height:1.9;padding-left:16px">
     <li>Best frame extracted from video</li>
     <li>Stenosis detection — Mask2Former</li>
-    <li>Coronary segmentation — YOLOv8m-seg (26-class)</li>
+    <li>Coronary segmentation — 26-class anatomy labelling</li>
     <li>Binary vessel mask — ResUNet</li>
     <li>FFR estimation — QFR v4</li>
     <li>SYNTAX score computation</li>
@@ -1592,8 +1618,8 @@ with gr.Blocks(css=CSS, title="Angio AI") as demo:
 
             with gr.Row():
                 with gr.Column():
-                    gr.HTML('<div class="card-title" style="margin-bottom:6px">Coronary segmentation (YOLOv8m-seg)</div>')
-                    seg_out   = gr.Image(label="26-class overlay", height=300)
+                    gr.HTML('<div class="card-title" style="margin-bottom:6px">Coronary segmentation (26-class)</div>')
+                    seg_out   = gr.Image(label="26-class coronary overlay", height=300)
                 with gr.Column():
                     gr.HTML('<div class="card-title" style="margin-bottom:6px">FFR analysis</div>')
                     ffr_out = gr.Image(label="FFR pipeline output", height=300)
@@ -1616,6 +1642,11 @@ with gr.Blocks(css=CSS, title="Angio AI") as demo:
     def run_and_switch(video, sc, sg, px):
         return analyse(video, sc, sg, px)
 
+    def load_demo1():
+        return _get_demo_video(DEMO_VIDEO_1_NAME)
+
+    btn_demo1.click(fn=load_demo1, inputs=[], outputs=[video_input])
+
     btn_analyse.click(
         fn=run_and_switch,
         inputs=[video_input, sten_conf, seg_conf, px_per_mm],