I have two functions with opencv and python that look like this. Individually they work like a charm but when put in a way like this(see down below), the code ignores the findclosestspot() function. Does anyone know how this can be? Any ideas are appreciated because chatgpt does a terrible job explaining this. Could it be that it is going too fast to process the image?
example:
if not confirmclosestspot(): findclosestspot()The functions:
def confirmclosestspot(): top_left = 246, 231 bottom_right = 279, 262 screenshot = aut.screenshot(region=(top_left[0], top_left[1], bottom_right[0] - top_left[0], bottom_right[1] - top_left[1])) screenshot_cv = cv.cvtColor(np.array(screenshot),cv.COLOR_RGB2BGR) grayscale1 = cv.cvtColor(screenshot_cv, cv.COLOR_BGR2GRAY) template1 = cv.imread(str("picture1.png"), cv.IMREAD_GRAYSCALE) result = cv.matchTemplate(template1, grayscale1, cv.TM_CCOEFF_NORMED ) threshold = 0.1 min_val, max_val, min_loc, max_loc = cv.minMaxLoc(result) print(min_val, max_val, min_loc, max_loc) if threshold >= max_val: print(max_val, max_loc) return True else: print(max_val, max_loc) return Falsedef findclosestspot(): print("Executing findclosestspot()") top_left = 1, 227 bottom_right = 515, 270 screenshot = aut.screenshot(region=(top_left[0], top_left[1], bottom_right[0] - top_left[0], bottom_right[1] - top_left[1])) screenshot_cv = cv.cvtColor(np.array(screenshot), cv.COLOR_RGB2BGR) grayscale = cv.cvtColor(screenshot_cv, cv.COLOR_BGR2GRAY) template = cv.imread(str("picture2.png"), cv.IMREAD_GRAYSCALE) result = cv.matchTemplate(grayscale, template, cv.TM_CCOEFF_NORMED) threshold = 0.8 min_val, max_val, min_loc, max_loc = cv.minMaxLoc(result) print(min_val, max_val, min_loc, max_loc) if max_val >= threshold: locs = np.where(result >= threshold) best_match_loc = None best_distance = float('inf') # Target coordinates target_x, target_y = 263, 191 # Iterate over all locations above threshold and find the closest one to the target coordinates for pt in zip(*locs[::-1]): center_x = pt[0] + (template.shape[1] // 2) center_y = pt[1] + (template.shape[0] // 2) distance = ((center_x - target_x) ** 2 + (center_y - target_y) ** 2) ** 0.5 # Check if this match is closer than the previous best match and not overlapping with it if distance < best_distance: overlapping = False for existing_loc in locs: existing_center_x = existing_loc[0] + (template.shape[1] // 2) existing_center_y = existing_loc[1] + (template.shape[0] // 2) existing_distance = ((existing_center_x - center_x) ** 2 + (existing_center_y - center_y) ** 2) ** 0.5 if existing_distance < 10: # Adjust this threshold value as needed overlapping = True break if not overlapping: best_match_loc = pt best_distance = distance if best_match_loc is not None: # Click at the center of the best match location in actual screen coordinates max_loc_screen_x = top_left[0] + best_match_loc[0] + 10 max_loc_screen_y = top_left[1] + best_match_loc[1] + 15 aut.click(max_loc_screen_x, max_loc_screen_y, clicks=2) aut.click(max_loc_screen_x, max_loc_screen_y, clicks=2) # Click twice on the best match else: print("No suitable match found. Performing default action.") aut.scroll(-5000) aut.moveTo(262, 212) aut.click() time.sleep(10) else: print("Max value not above threshold. Performing default action.") aut.moveTo(262, 212) aut.click()I have tried many versions of
if not confirmclosestspot(): findclosestspot() but it always ignores the findclosestspot() function and keeps doing whatever code I put below.