update

app.py

@@ -340,7 +340,7 @@ def run(img_str):
     from concurrent.futures import ProcessPoolExecutor
     from concurrent.futures import as_completed
     gif_results = []
-    with ProcessPoolExecutor() as executor:
+    with ProcessPoolExecutor(max_workers=8) as executor:
         futures = [executor.submit(run_code, new_folder, i, code) for i, code in enumerate(codes)]
         for future in as_completed(futures):
             try:

myturtle_cv.py

@@ -3,7 +3,7 @@ import cv2
 
 
 def crop_and_scaled_imgs(imgs):
-    PAD = 10
+    PAD = 5
     # use the last image to find the bounding box of the non-white area and the transformation parameters
     # and then apply the transformation to all images
 
@@ -18,34 +18,41 @@ def crop_and_scaled_imgs(imgs):
 
     # Get the bounding box of the non-zero pixels
     x, y, w, h = cv2.boundingRect(coords)
-    x -= PAD
-    y -= PAD
-    w += 2 * PAD
-    h += 2 * PAD
-
-    # Calculate the position to center the ROI in the 256x256 image
-    start_x = max(0, (256 - w) // 2)
-    start_y = max(0, (256 - h) // 2)
-
-    # Create a new 256x256 rgb images
-    new_imgs = [np.ones((256, 256, 3), dtype=np.uint8) * 255 for _ in range(len(imgs))]
+    x = max(0, x-PAD)
+    y = max(0, y-PAD)
+    x_end = min(img.shape[1], x+w+2*PAD)
+    y_end = min(img.shape[0], y+h+2*PAD)
+    w = x_end - x
+    h = y_end - y
+
+    SIZE = 400
+    # Calculate the position to center the ROI in the SIZExSIZE image
+    start_x = max(0, (SIZE - w) // 2)
+    start_y = max(0, (SIZE - h) // 2)
+
+    # Create a new SIZExSIZE rgb images
+    new_imgs = [np.ones((SIZE, SIZE, 3), dtype=np.uint8) * 255 for _ in range(len(imgs))]
     for i in range(len(imgs)):
         # Extract the ROI (region of interest) of the non-white area
         roi = imgs[i][y:y+h, x:x+w]
         # If the ROI is larger than 256x256, resize it
 
-        if w > 256 or h > 256:
-            scale = min(256 / w, 256 / h)
+        if w > SIZE or h > SIZE:
+            scale = min(SIZE / w, SIZE / h)
             new_w = int(w * scale)
             new_h = int(h * scale)
             roi = cv2.resize(roi, (new_w, new_h), interpolation=cv2.INTER_AREA)
-            w, h = new_w, new_h
+        else:
+            new_w = w
+            new_h = h
+
+
 
         # new_imgs[i] = np.ones((256, 256), dtype=np.uint8) * 255
         # centered_img = np.ones((256, 256), dtype=np.uint8) * 255
 
         # Place the ROI in the centered position
-        new_imgs[i][start_y:start_y+h, start_x:start_x+w] = roi
+        new_imgs[i][start_y:start_y+new_h, start_x:start_x+new_w] = roi
     
     return new_imgs
 
@@ -115,6 +122,7 @@ class Turtle:
         self.time_since_last_frame += abs(dist) / MOVE_SPEED
         # self.frames.append(self.canvas.copy())
         # self.save_frame_with_turtle()
+        # print(self.x, self.y)
 
     def save_frame_with_turtle(self):
         # save the current frame to frames buffer
@@ -365,7 +373,7 @@ if __name__ == "__main__":
         for i in range(7):
             with fork_state():
                 for j in range(4):
-                    forward(2*i)
+                    forward(3*i)
                     left(90.0)
         return turtle.frames