list.py 10.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235
  1. # This Python file uses the following encoding: utf-8
  2. # @author runhey
  3. # github https://github.com/runhey
  4. import cv2
  5. import random
  6. import numpy as np
  7. from random import randint
  8. from ppocronnx.predict_system import BoxedResult
  9. from module.atom.ocr import RuleOcr
  10. from module.atom.image import RuleImage
  11. from module.logger import logger
  12. class RuleList:
  13. def __init__(self, folder: str, direction: str, mode: str, roi_back: tuple, size: tuple, array: list[str]) -> None:
  14. """
  15. 初始化
  16. :param folder:
  17. :param direction:
  18. :param mode:
  19. :param roi_back:
  20. :param size:
  21. :param array:
  22. """
  23. self.folder = folder
  24. self.is_vertical = direction == "vertical"
  25. self.is_image = mode == "image"
  26. self.is_ocr = mode == "ocr"
  27. self.roi_back: list = list(roi_back)
  28. self.size: list = list(size)
  29. # 这个表示一个区域中一次所能显示的最大数量
  30. self.max_show: int = self.roi_back[3] // self.size[1] if self.is_vertical else self.roi_back[2] // self.size[0]
  31. self.array: list = array
  32. # 以下是需要计算的
  33. self.appear_area: list = None # 出现的区域
  34. self.is_bottom = False # 表示是否已经滑动到底部了
  35. self._target = None # 目标
  36. self.targets = {} # 目标列表 只是针对image
  37. def swipe_pos(self, number: int=2, after: bool=True) -> tuple:
  38. """
  39. 返回滑动的起始位置和终点位置
  40. :param number:
  41. :param after:
  42. :return:
  43. """
  44. center: tuple = (self.roi_back[0] + self.roi_back[2]) // 2, (self.roi_back[1] + self.roi_back[3]) // 2
  45. distance: int = self.size[1] * number if self.is_vertical else self.size[0] * number
  46. 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)
  47. 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)
  48. x1, y1, x2, y2 = None, None, None, None
  49. if self.is_vertical:
  50. if after:
  51. # 竖直方向向下滑动, 那就是起始点是在下面 终点是在上面
  52. x1, y1 = center[0]+random_start, center[1] + distance//2
  53. x2, y2 = center[0]+random_end, center[1] - distance//2
  54. else:
  55. # 竖直方向向上滑动, 那就是起始点是在上面 终点是在下面
  56. x1, y1 = center[0]+random_start, center[1] - distance//2
  57. x2, y2 = center[0]+random_end, center[1] + distance//2
  58. else:
  59. if after:
  60. # 水平方向向右滑动, 那就是起始点是在右面 终点是在左面
  61. x1, y1 = center[0] + distance//2, center[1]+random_start
  62. x2, y2 = center[0] - distance//2, center[1]+random_end
  63. else:
  64. # 水平方向向左滑动, 那就是起始点是在左面 终点是在右面
  65. x1, y1 = center[0] - distance//2, center[1]+random_start
  66. x2, y2 = center[0] + distance//2, center[1]+random_end
  67. if x1 is None or y1 is None or x2 is None or y2 is None:
  68. raise Exception("滑动位置计算错误")
  69. return int(x1), int(y1), int(x2), int(y2)
  70. def target_check(self, name: str) -> bool:
  71. """
  72. 检查输入的名称
  73. :param name:
  74. :return:
  75. """
  76. # 首先获取构建类的信息
  77. # roi_front = None
  78. # for item in self.array:
  79. # if item["itemName"] == name:
  80. # roi_front = item["roiFront"].split(",")
  81. # break
  82. # if roi_front is None:
  83. # logger.error(f'Not found {name} in {self.array}')
  84. # return False
  85. file = self.folder + "/" + name + ".png"
  86. # 保证匹配的目标 是正确的
  87. if self.is_image:
  88. if self._target is None or isinstance(self._target, RuleOcr):
  89. self._target = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back, method="Template matching", threshold=0.8, file=file)
  90. # 如果不对的话,就重新构建
  91. elif self._target.name != name:
  92. self._target = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back, method="Template matching", threshold=0.8, file=file)
  93. else:
  94. pass
  95. elif self.is_ocr:
  96. if self._target is None or isinstance(self._target, RuleImage):
  97. self._target = RuleOcr(roi=self.roi_back, area=(0, 0, 10, 10), mode="Full", method="Default",
  98. keyword=name, name=name)
  99. elif self._target.name != name:
  100. self._target.name = name
  101. self._target.keyword = name
  102. else:
  103. logger.error(f'Not found {name} in {self.array}')
  104. return False
  105. def targets_check(self, targets: list):
  106. """
  107. 检查输入的图片名是缓存有
  108. :param targets:
  109. :return:
  110. """
  111. for item in targets:
  112. if item in self.targets:
  113. continue
  114. # 如果还没有缓存的话就先缓存
  115. file = self.folder + "/" + item + ".png"
  116. self.targets[item] = RuleImage(roi_front=self.roi_back, roi_back=self.roi_back,
  117. method="Template matching", threshold=0.8, file=file)
  118. def image_appear(self, image: np.array, name: str) -> bool | tuple:
  119. """
  120. 判断是否出现了某个图片
  121. :param image: 屏幕的截图
  122. :param name:
  123. :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
  124. """
  125. if self.is_image and isinstance(name, str):
  126. self.target_check(name)
  127. appear = self._target.match(image)
  128. if appear:
  129. return self._target.coord()
  130. else:
  131. return False
  132. elif self.is_image and isinstance(name, list):
  133. self.targets_check(name)
  134. for item in name:
  135. appear = self.targets[item].match(image)
  136. if appear:
  137. return self.targets[item].coord()
  138. return False
  139. else:
  140. logger.error(f'Mode is not image')
  141. return False
  142. def ocr_appear(self, image: np.array, name: str):
  143. """
  144. 判断是否出现了某个文字,
  145. :param image: 屏幕的截图
  146. :param name:
  147. :return: 如果是出现的返回可以点击的坐标 x , y
  148. 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,
  149. """
  150. if self.is_image:
  151. return False
  152. self.target_check(name)
  153. # 开始一次ocr的检测
  154. boxed_results: list[BoxedResult] = self._target.detect_and_ocr(image)
  155. if not boxed_results:
  156. logger.warning(f'Not angy result in image')
  157. return 0, 0
  158. # 判断所有的结果,给获取的列表建立引索
  159. # index_list = self._target.filter(boxed_results, name)
  160. # logger.info(index_list)
  161. box = None
  162. for item in boxed_results:
  163. if item.ocr_text == name and item.score > RuleOcr.score:
  164. box = item.box
  165. break
  166. if box is not None:
  167. 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]
  168. x = rec_x + rec_w // 2 + self.roi_back[0]
  169. y = rec_y + rec_h // 2 + self.roi_back[1]
  170. logger.info(f'Ocr {name} appear in current screen, do not need to scroll')
  171. return x, y
  172. # if index_list and len(index_list) >= 1:
  173. # box = boxed_results[index_list[0]].box
  174. # 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]
  175. # return rec_x+self.roi_back[0], rec_y+self.roi_back[1], rec_w, rec_h
  176. # 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
  177. else:
  178. keyword_list: list = [item.ocr_text for item in boxed_results if item.score > RuleOcr.score]
  179. # 判断识别处理的文字是否属于要验证识别的文字(array)
  180. keyword_list = [keyword for keyword in keyword_list if keyword in self.array]
  181. logger.info(f'After list: {keyword_list}')
  182. if not keyword_list:
  183. logger.warning(f'Not found {name} in {self.array}')
  184. return 2
  185. start_index = self.array.index(keyword_list[0])
  186. end_index = self.array.index(keyword_list[-1])
  187. if name in keyword_list:
  188. distance_start = 0
  189. distance_end = 0
  190. current_index = self.array.index(name)
  191. distance_start = 0
  192. distance_end = 0
  193. if current_index < start_index or current_index > end_index:
  194. distance_start = current_index - start_index
  195. distance_end = current_index - end_index
  196. # 如果目标的是在当前的显示的前面, 则distance为负的
  197. # 如果目标的是在当前的显示的后面, 则distance为正的
  198. # 所有前面返回的是负的,后面返回的是正的
  199. if distance_start == 0 and distance_end == 0:
  200. logger.error(f'{name} not found in {keyword_list}')
  201. return (distance_start + distance_end) // 2
  202. if __name__ == '__main__':
  203. L_N33AME = RuleList(folder="./tasks/Orochi/res", direction="vertical", mode="ocr", roi_back=(160, 130, 317, 500),
  204. size=(301, 86),
  205. 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"])
  206. image = cv2.imread("D:/watu_list_text204237.png")
  207. print(L_N33AME.ocr_appear(image, "柒层"))
  208. print(L_N33AME.ocr_appear(image, "神罚"))