Update detector.py

detector.py

@@ -94,9 +94,6 @@ def extract_platform(url: str) -> str:
     else:
         return "Direct Link / Other"
 
-# ─────────────────────────────────────────────
-# VIDEO DOWNLOAD 
-# ─────────────────────────────────────────────
 def download_video(url: str) -> tuple[bool, str]:
     tmp = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False)
     tmp.close()
@@ -130,6 +127,9 @@ def download_video(url: str) -> tuple[bool, str]:
 # ─────────────────────────────────────────────
 @spaces.GPU(duration=120)
 def encode_and_classify(windows: list) -> list:
+    gpu_start_time = time.time()
+    print(f"\n[TIMING] --- GPU Allocated & Inference Started ---")
+    
     if not windows:
         return []
 
@@ -176,14 +176,17 @@ def encode_and_classify(windows: list) -> list:
             "start_sec": window["start_sec"],
             "end_sec": window["end_sec"],
         })
-        
+    
+    gpu_end_time = time.time()
+    print(f"[TIMING] GPU execution completed in {gpu_end_time - gpu_start_time:.2f}s")
     return scores
 
 # ─────────────────────────────────────────────
 # BACKGROUND DATA UPLOAD & ALERTS (Asynchronous)
 # ─────────────────────────────────────────────
 def upload_and_notify_background(video_path: str, meta: dict):
-    """Executes network requests and file cleanups without blocking the user response."""
+    bg_start_time = time.time()
+    print("\n[TIMING] Background thread started for dataset upload...")
     try:
         if not DATASET_REPO_ID or not HF_TOKEN:
             return
@@ -232,22 +235,24 @@ def upload_and_notify_background(video_path: str, meta: dict):
     except Exception as e:
         print(f"Background task execution failed: {e}")
     finally:
-        # File deletion safe-guards executed inside the thread after upload attempts complete
         for f in glob.glob(video_path + "*") + [video_path]:
             if os.path.exists(f):
                 try:
                     os.remove(f)
                 except Exception:
                     pass
+        print(f"[TIMING] Background thread finished in {time.time() - bg_start_time:.2f}s")
 
 # ─────────────────────────────────────────────
 # GRADIO PREDICT
 # ─────────────────────────────────────────────
 def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -> tuple[str, str, str, dict]:
+    global_start_time = time.time()
+    print(f"\n{'='*30}\n[TIMING] Request Received.")
+    
     if not consent:
         return "❌ Required: You must check the data collection consent box to process videos.", "", "", current_state
     
-    start_time = time.time()
     url = (url or "").strip()
     
     if current_state and "video_path" in current_state:
@@ -259,6 +264,7 @@ def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -
                 except Exception:
                     pass
 
+    file_start_time = time.time()
     if uploaded_video:
         tmp = tempfile.NamedTemporaryFile(suffix=".mp4", delete=False)
         shutil.copy(uploaded_video, tmp.name)
@@ -275,6 +281,9 @@ def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -
     else:
         return "Please enter a URL or upload a video.", "", "", {}
 
+    file_ready_time = time.time()
+    print(f"[TIMING] File Acquisition (Download/Copy) took: {file_ready_time - file_start_time:.2f}s")
+
     try:
         windows, total_frames, fps, duration_sec = get_video_windows(video_path)
     except Exception as e:
@@ -282,7 +291,15 @@ def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -
             os.remove(video_path)
         return f"Video decode failed: {e}", "", "", {}
         
+    cpu_decode_time = time.time()
+    print(f"[TIMING] CPU Video Decoding took: {cpu_decode_time - file_ready_time:.2f}s")
+    
+    print(f"[TIMING] Requesting GPU Hardware from ZeroGPU Queue...")
+    queue_start_time = time.time()
     scores = encode_and_classify(windows)
+    queue_and_gpu_time = time.time()
+    
+    print(f"[TIMING] Total time spent requesting queue + GPU execution: {queue_and_gpu_time - queue_start_time:.2f}s")
     
     if not scores:
         if os.path.exists(video_path):
@@ -291,7 +308,7 @@ def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -
         
     final_prob = max(s["prob"] for s in scores)
     label = "⚠️  AI-GENERATED (FAKE)" if final_prob > 0.5 else "✅  REAL"
-    elapsed_time = time.time() - start_time
+    elapsed_time = time.time() - global_start_time
     confidence = f"{final_prob * 100:.1f}%"
     
     lines = [
@@ -312,14 +329,13 @@ def predict(url: str, uploaded_video: str, consent: bool, current_state: dict) -
         "total_frames": total_frames
     }
     
-    # SOLUTION: Dispatches upload tasks asynchronously to an detached thread.
-    # The application instantly proceeds to return the results to the user interface.
     threading.Thread(
         target=upload_and_notify_background,
         args=(video_path, meta),
         daemon=True
     ).start()
-                
+    
+    print(f"[TIMING] Total Blocking Time Before UI Return: {elapsed_time:.2f}s")
     return label, confidence, "\n".join(lines), {}
 
 # ─────────────────────────────────────────────