| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242 |
- # This Python file uses the following encoding: utf-8
- # @author runhey
- # github https://github.com/runhey
- import cv2
- import random
- import numpy as np
- from random import randint
- from ppocronnx.predict_system import BoxedResult
- from module.atom.ocr import RuleOcr
- from module.atom.image import RuleImage
- from module.logger import logger
- class RuleList:
- def __init__(self, folder: str, direction: str, mode: str, roi_back: tuple, size: tuple, array: list[str]) -> None:
- """
- 初始化
- :param folder:
- :param direction:
- :param mode:
- :param roi_back:
- :param size:
- :param array:
- """
- self.folder = folder
- self.is_vertical = direction == "vertical"
- self.is_image = mode == "image"
- self.is_ocr = mode == "ocr"
- self.roi_back: list = list(roi_back)
- self.size: list = list(size)
- # 这个表示一个区域中一次所能显示的最大数量
- self.max_show: int = self.roi_back[3] // self.size[1] if self.is_vertical else self.roi_back[2] // self.size[0]
- self.array: list = array
- # 以下是需要计算的
- self.appear_area: list = None # 出现的区域
- self.is_bottom = False # 表示是否已经滑动到底部了
- self._target = None # 目标
- self.targets = {} # 目标列表 只是针对image
- @property
- def name(self):
- return f'RuleList[{self.__hash__()}]'
- def __hash__(self):
- return hash((self.folder, self.is_vertical, self.is_image, self.is_ocr, tuple(self.roi_back), tuple(self.size), tuple(self.array)))
- def swipe_pos(self, number: int=2, after: bool=True) -> tuple:
- """
- 返回滑动的起始位置和终点位置
- :param number:
- :param after:
- :return:
- """
- center: tuple = (self.roi_back[0] + self.roi_back[2]) // 2, (self.roi_back[1] + self.roi_back[3]) // 2
- distance: int = self.size[1] * number if self.is_vertical else self.size[0] * number
- random_start: int = randint(- self.size[0] // 4, self.size[0] // 4) if self.is_vertical else randint(- self.size[1] // 4, self.size[1] // 4)
- random_end: int = randint(- self.size[0] // 4, self.size[0] // 4) if self.is_vertical else randint(- self.size[1] // 4, self.size[1] // 4)
- x1, y1, x2, y2 = None, None, None, None
- if self.is_vertical:
- if after:
- # 竖直方向向下滑动, 那就是起始点是在下面 终点是在上面
- x1, y1 = center[0]+random_start, center[1] + distance//2
- x2, y2 = center[0]+random_end, center[1] - distance//2
- else:
- # 竖直方向向上滑动, 那就是起始点是在上面 终点是在下面
- x1, y1 = center[0]+random_start, center[1] - distance//2
- x2, y2 = center[0]+random_end, center[1] + distance//2
- else:
- if after:
- # 水平方向向右滑动, 那就是起始点是在右面 终点是在左面
- x1, y1 = center[0] + distance//2, center[1]+random_start
- x2, y2 = center[0] - distance//2, center[1]+random_end
- else:
- # 水平方向向左滑动, 那就是起始点是在左面 终点是在右面
- x1, y1 = center[0] - distance//2, center[1]+random_start
- x2, y2 = center[0] + distance//2, center[1]+random_end
- if x1 is None or y1 is None or x2 is None or y2 is None:
- raise Exception("滑动位置计算错误")
- return int(x1), int(y1), int(x2), int(y2)
- def target_check(self, name: str) -> bool:
- """
- 检查输入的名称
- :param name:
- :return:
- """
- # 首先获取构建类的信息
- # roi_front = None
- # for item in self.array:
- # if item["itemName"] == name:
- # roi_front = item["roiFront"].split(",")
- # break
- # if roi_front is None:
- # logger.error(f'Not found {name} in {self.array}')
- # return False
- file = self.folder + "/" + name + ".png"
- # 保证匹配的目标 是正确的
- if self.is_image:
- if self._target is None or isinstance(self._target, RuleOcr):
- self._target = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back, method="Template matching", threshold=0.8, file=file)
- # 如果不对的话,就重新构建
- elif self._target.name != name:
- self._target = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back, method="Template matching", threshold=0.8, file=file)
- else:
- pass
- elif self.is_ocr:
- if self._target is None or isinstance(self._target, RuleImage):
- self._target = RuleOcr(roi=self.roi_back, area=(0, 0, 10, 10), mode="Full", method="Default",
- keyword=name, name=name)
- elif self._target.name != name:
- self._target.name = name
- self._target.keyword = name
- else:
- logger.error(f'Not found {name} in {self.array}')
- return False
- def targets_check(self, targets: list):
- """
- 检查输入的图片名是缓存有
- :param targets:
- :return:
- """
- for item in targets:
- if item in self.targets:
- continue
- # 如果还没有缓存的话就先缓存
- file = self.folder + "/" + item + ".png"
- self.targets[item] = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back,
- method="Template matching", threshold=0.8, file=file)
- def image_appear(self, image: np.array, name: str) -> bool | tuple:
- """
- 判断是否出现了某个图片
- :param image: 屏幕的截图
- :param name:
- :return: 如果在当前的显示中,返回可以点击的坐标.如果不是出现就是返回False
- """
- if self.is_image and isinstance(name, str):
- self.target_check(name)
- appear = self._target.match(image)
- if appear:
- return self._target.coord()
- else:
- return False
- elif self.is_image and isinstance(name, list):
- self.targets_check(name)
- for item in name:
- appear = self.targets[item].match(image)
- if appear:
- return self.targets[item].coord()
- return False
- else:
- logger.error(f'Mode is not image')
- return False
- def ocr_appear(self, image: np.array, name: str):
- """
- 判断是否出现了某个文字,
- :param image: 屏幕的截图
- :param name:
- :return: 如果是出现的返回可以点击的坐标 x , y
- 如果是在当前显示的前面则返回小于零的int, 如果是在当前显示的后面则返回大于零的int,
- """
- if self.is_image:
- return False
- self.target_check(name)
- # 开始一次ocr的检测
- boxed_results: list[BoxedResult] = self._target.detect_and_ocr(image)
- if not boxed_results:
- logger.warning(f'Not angy result in image')
- return 0, 0
- # 判断所有的结果,给获取的列表建立引索
- # index_list = self._target.filter(boxed_results, name)
- # logger.info(index_list)
- box = None
- for item in boxed_results:
- if item.ocr_text == name and item.score > RuleOcr.score:
- box = item.box
- break
- if box is not None:
- rec_x, rec_y, rec_w, rec_h = box[0, 0], box[0, 1], box[1, 0] - box[0, 0], box[2, 1] - box[0, 1]
- x = rec_x + rec_w // 2 + self.roi_back[0]
- y = rec_y + rec_h // 2 + self.roi_back[1]
- logger.info(f'Ocr {name} appear in current screen, do not need to scroll')
- return x, y
- # if index_list and len(index_list) >= 1:
- # box = boxed_results[index_list[0]].box
- # rec_x, rec_y, rec_w, rec_h = box[0, 0], box[0, 1], box[1, 0] - box[0, 0], box[2, 1] - box[0, 1]
- # return rec_x+self.roi_back[0], rec_y+self.roi_back[1], rec_w, rec_h
- # 如果没有找到,那么应该不是在当前的页面,需要获取当前的信息判断是往前滑动还是往后滑动
- else:
- keyword_list: list = [item.ocr_text for item in boxed_results if item.score > RuleOcr.score]
- # 判断识别处理的文字是否属于要验证识别的文字(array)
- keyword_list = [keyword for keyword in keyword_list if keyword in self.array]
- logger.info(f'After list: {keyword_list}')
- if not keyword_list:
- logger.warning(f'Not found {name} in {self.array}')
- return 2
- start_index = self.array.index(keyword_list[0])
- end_index = self.array.index(keyword_list[-1])
- if name in keyword_list:
- distance_start = 0
- distance_end = 0
- current_index = self.array.index(name)
- distance_start = 0
- distance_end = 0
- if current_index < start_index or current_index > end_index:
- distance_start = current_index - start_index
- distance_end = current_index - end_index
- # 如果目标的是在当前的显示的前面, 则distance为负的
- # 如果目标的是在当前的显示的后面, 则distance为正的
- # 所有前面返回的是负的,后面返回的是正的
- if distance_start == 0 and distance_end == 0:
- logger.error(f'{name} not found in {keyword_list}')
- return (distance_start + distance_end) // 2
- if __name__ == '__main__':
- L_N33AME = RuleList(folder="./tasks/Orochi/res", direction="vertical", mode="ocr", roi_back=(160, 130, 317, 500),
- size=(301, 86),
- array=["壹层", "贰层", "叁层", "肆层", "伍层", "陆层", "柒层", "捌层", "玖层", "拾层", "悲鸣", "神罚"])
- image = cv2.imread("D:/watu_list_text204237.png")
- print(L_N33AME.ocr_appear(image, "柒层"))
- print(L_N33AME.ocr_appear(image, "神罚"))
|