show the exact image that the model analyzed1

app.py

@@ -89,8 +89,6 @@ def run_leukocoria_prediction(iris_crop):
 # --- 3. FastAPI Application ---
 app = FastAPI()
 
-# In app.py - an updated full_detection_pipeline function
-
 @app.post("/detect/")
 async def full_detection_pipeline(image: UploadFile = File(...)):
     with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as tmp:
@@ -106,28 +104,31 @@ async def full_detection_pipeline(image: UploadFile = File(...)):
         if not detect_faces_roboflow(temp_image_path):
             return JSONResponse(status_code=400, content={"error": "No face detected."})
 
-        # --- NEW: Logic to track which image is used for analysis ---
-        image_for_analysis = raw_image
-
-        # First, try to detect eyes on the original image
+        # --- This is the corrected, robust logic ---
+        
+        # By default, assume we will use the original image
+        image_to_process = raw_image
+        
+        # First, try detection on the original image
         print("--- Attempting detection on original image... ---")
-        eye_crops, error_msg = detect_eyes_roboflow(temp_image_path, raw_image)
+        eye_crops, error_msg = detect_eyes_roboflow(temp_image_path, image_to_process)
 
-        # If that fails, create a mirrored version and try again
+        # If the first attempt fails, try the mirrored version
         if len(eye_crops) != 2:
             print("--- Original failed. Attempting detection on mirrored image... ---")
             mirrored_image = cv2.flip(raw_image, 1)
-            image_for_analysis = mirrored_image # Set the mirrored image as the one to use
+            image_to_process = mirrored_image # Set the mirrored image as the one to use now
 
             with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as tmp_mirrored:
                 cv2.imwrite(tmp_mirrored.name, mirrored_image)
                 temp_mirrored_image_path = tmp_mirrored.name
             try:
-                eye_crops, error_msg = detect_eyes_roboflow(temp_mirrored_image_path, mirrored_image)
+                # Ensure we pass the mirrored image data for cropping
+                eye_crops, error_msg = detect_eyes_roboflow(temp_mirrored_image_path, image_to_process)
             finally:
                 os.remove(temp_mirrored_image_path)
 
-        # After trying both, check if we have a valid result
+        # Final check after both attempts
         if error_msg or len(eye_crops) != 2:
             return JSONResponse(
                 status_code=400,
@@ -154,19 +155,18 @@ async def full_detection_pipeline(image: UploadFile = File(...)):
                 flags[side] = has_leuko
             else:
                 flags[side] = None
-
-        # --- NEW: Encode the successfully analyzed image to send back ---
-        is_success_main, buffer_main = cv2.imencode(".jpg", image_for_analysis)
+        
+        # Encode the image that was successfully analyzed (original or mirrored)
+        is_success_main, buffer_main = cv2.imencode(".jpg", image_to_process)
         analyzed_image_b64 = ""
         if is_success_main:
             analyzed_image_b64 = "data:image/jpeg;base64," + base64.b64encode(buffer_main).decode("utf-8")
 
-        # --- NEW: Add the analyzed_image to the final response ---
         return JSONResponse(content={
             "leukocoria": flags,
             "warnings": [],
             "two_eyes": eye_images_b64,
-            "analyzed_image": analyzed_image_b64 # This is the new field
+            "analyzed_image": analyzed_image_b64
         })
 
     finally: