| from utils.screen_utils import ScreenController |
| from models.vision_model import VisionModel |
| import time |
|
|
| class Controller: |
| def __init__(self): |
| self.screen_controller = ScreenController() |
| self.vision_model = VisionModel() |
| |
| def execute_command(self, command): |
| """执行自然语言命令""" |
| try: |
| |
| action_plan = self.vision_model.interpret_command(command) |
| if not action_plan: |
| return False, "无法解析命令" |
| |
| |
| screen = self.screen_controller.capture_screen() |
| |
| |
| location = self.vision_model.get_element_location(screen, action_plan['target']) |
| if not location: |
| return False, f"无法找到目标元素：{action_plan['target']}" |
| |
| |
| success = self._perform_action(action_plan['action'], location, action_plan.get('params', {})) |
| |
| return success, "操作执行成功" if success else "操作执行失败" |
| |
| except Exception as e: |
| return False, f"执行命令时出错：{str(e)}" |
| |
| def _perform_action(self, action, location, params): |
| """执行具体操作""" |
| x, y = location.get('x'), location.get('y') |
| |
| if not (x and y): |
| return False |
| |
| if action == 'click': |
| return self.screen_controller.click_position(x, y) |
| elif action == 'type': |
| self.screen_controller.click_position(x, y) |
| return self.screen_controller.type_text(params.get('text', '')) |
| elif action == 'press': |
| return self.screen_controller.press_key(params.get('key', '')) |
| elif action == 'move': |
| return self.screen_controller.move_to(x, y) |
| elif action == 'drag': |
| target_x = params.get('target_x') |
| target_y = params.get('target_y') |
| if target_x and target_y: |
| self.screen_controller.move_to(x, y) |
| return self.screen_controller.drag_to(target_x, target_y) |
| elif action == 'scroll': |
| return self.screen_controller.scroll(params.get('clicks', 0)) |
| |
| return False |
