Update app.py

app.py

@@ -6,6 +6,17 @@ from io import BytesIO
 import base64
 import time
 
+# Suppress optional model warnings BEFORE importing inference
+os.environ["CORE_MODEL_SAM_ENABLED"] = "False"
+os.environ["CORE_MODEL_SAM2_ENABLED"] = "False"
+os.environ["CORE_MODEL_SAM3_ENABLED"] = "False"
+os.environ["CORE_MODEL_GAZE_ENABLED"] = "False"
+os.environ["CORE_MODEL_GROUNDINGDINO_ENABLED"] = "False"
+os.environ["CORE_MODEL_YOLO_WORLD_ENABLED"] = "False"
+
+# Set stable ONNX providers
+os.environ["ONNXRUNTIME_EXECUTION_PROVIDERS"] = "CPUExecutionProvider"
+
 import cv2
 import numpy as np
 from PIL import Image
@@ -27,8 +38,7 @@ from more_itertools import chunked
 from sklearn.cluster import KMeans
 import umap
 
-from inference import get_model
-from inference_sdk.http.errors import HTTPCallErrorError
+from inference_sdk import InferenceHTTPClient
 
 # ==============================================
 # ENVIRONMENT VARIABLES
@@ -59,10 +69,12 @@ print(f"📏 Pitch config units - Length: {CONFIG.length}, Width: {CONFIG.width}
 print(f"📏 Scale factors - X: {SCALE_X:.6f} m/unit, Y: {SCALE_Y:.6f} m/unit")
 
 # ==============================================
-# MODEL INITIALIZATION
+# ROBOFLOW INFERENCE CLIENT
 # ==============================================
-PLAYER_DETECTION_MODEL = None
-FIELD_DETECTION_MODEL = None
+CLIENT = InferenceHTTPClient(
+    api_url="https://detect.roboflow.com",
+    api_key=ROBOFLOW_API_KEY
+)
 
 PLAYER_DETECTION_MODEL_ID = "football-players-detection-3zvbc/11"
 FIELD_DETECTION_MODEL_ID = "football-field-detection-f07vi/14"
@@ -71,28 +83,8 @@ FIELD_DETECTION_MODEL_ID = "football-field-detection-f07vi/14"
 BALL_ID, GOALKEEPER_ID, PLAYER_ID, REFEREE_ID = 0, 1, 2, 3
 
 
-def initialize_models():
-    """Initialize detection models with local inference (more reliable than HTTP API)"""
-    global PLAYER_DETECTION_MODEL, FIELD_DETECTION_MODEL
-    
-    try:
-        print("📦 Loading detection models locally...")
-        PLAYER_DETECTION_MODEL = get_model(
-            model_id=PLAYER_DETECTION_MODEL_ID,
-            api_key=ROBOFLOW_API_KEY
-        )
-        FIELD_DETECTION_MODEL = get_model(
-            model_id=FIELD_DETECTION_MODEL_ID,
-            api_key=ROBOFLOW_API_KEY
-        )
-        print("✅ Models loaded successfully")
-    except Exception as e:
-        print(f"❌ Failed to load models: {e}")
-        raise
-
-
 def infer_with_confidence(
-    model,
+    model_id: str,
     frame: np.ndarray,
     confidence_threshold: float = 0.3,
     max_retries: int = 3
@@ -101,7 +93,7 @@ def infer_with_confidence(
     Run inference with retry logic for transient errors.
     
     Args:
-        model: The detection model to use
+        model_id: The model ID to use
         frame: Input frame
         confidence_threshold: Confidence threshold for detections
         max_retries: Maximum number of retry attempts
@@ -111,7 +103,7 @@ def infer_with_confidence(
     """
     for attempt in range(max_retries):
         try:
-            result = model.infer(frame, confidence=confidence_threshold)[0]
+            result = CLIENT.infer(frame, model_id=model_id)
             detections = sv.Detections.from_inference(result)
             if len(detections) > 0:
                 detections = detections[detections.confidence > confidence_threshold]
@@ -126,10 +118,6 @@ def infer_with_confidence(
                 # Return empty detections to continue processing
                 return None, sv.Detections.empty()
 
-
-# Initialize models at startup
-initialize_models()
-
 # ==============================================
 # SIGLIP MODEL (Embeddings)
 # ==============================================
@@ -725,7 +713,7 @@ def analyze_football_video(video_path: str, progress=gr.Progress()
             if not ret:
                 break
             if frame_idx % STRIDE == 0:
-                _, detections = infer_with_confidence(PLAYER_DETECTION_MODEL, frame, 0.3)
+                _, detections = infer_with_confidence(PLAYER_DETECTION_MODEL_ID, frame, 0.3)
                 detections = detections.with_nms(threshold=0.5, class_agnostic=True)
                 players_detections = detections[detections.class_id == PLAYER_ID]
                 if len(players_detections.xyxy) > 0:
@@ -808,7 +796,7 @@ def analyze_football_video(video_path: str, progress=gr.Progress()
                          desc=f"🎬 Processing frame {frame_idx}/{total_frames}")
 
             # --- detections ---
-            _, detections = infer_with_confidence(PLAYER_DETECTION_MODEL, frame, 0.3)
+            _, detections = infer_with_confidence(PLAYER_DETECTION_MODEL_ID, frame, 0.3)
             if len(detections.xyxy) == 0:
                 out.write(frame)
                 ball_path_raw.append(np.empty((0, 2)))
@@ -853,7 +841,7 @@ def analyze_football_video(video_path: str, progress=gr.Progress()
 
             # --- field homography ---
             try:
-                result_field, _ = infer_with_confidence(FIELD_DETECTION_MODEL, frame, 0.3)
+                result_field, _ = infer_with_confidence(FIELD_DETECTION_MODEL_ID, frame, 0.3)
                 key_points = sv.KeyPoints.from_inference(result_field)
 
                 filter_mask = key_points.confidence[0] > 0.5