Skip to content

周伟奇 / part_of_F3_OCR

Go to a project
Toggle navigation
Toggle navigation pinning
57ff6bf4 by 周伟奇
Options

add left

1 parent 43758c7e
Inline Side-by-side
Showing 1 changed file with 43 additions and 26 deletions
general_extractor/retriever.py
...@@ -110,37 +110,51 @@ class Retriever: ...@@ -110,37 +110,51 @@ class Retriever:
110 if self.find_keys_list[key_idx] is None: 110 if self.find_keys_list[key_idx] is None:
111 return 111 return
112 112
113 _, _, find_key_str, suffix_key, key_x0, key_y0, key_x1, key_y1 = self.find_keys_list[key_idx] 113 _, _, find_key_str, suffix_key, key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, \
114 114 key_x3_src, key_y3_src = self.find_keys_list[key_idx]
115
115 for scope_key_idx in scope_tuple[:-1]: 116 for scope_key_idx in scope_tuple[:-1]:
116 if self.find_keys_list[scope_key_idx] is None: 117 if self.find_keys_list[scope_key_idx] is None:
117 continue 118 continue
118 key_scope = self.find_keys_list[scope_key_idx][6] # left x1 119 key_scope_tuple = (
119 break 120 self.find_keys_list[scope_key_idx][6], self.find_keys_list[scope_key_idx][7]) # left x1, y1
121 break
120 else: 122 else:
121 key_scope = None 123 key_scope_tuple = None
122 124
123 if isinstance(if_startswith, str): 125 # if isinstance(if_startswith, str):
124 if isinstance(suffix_key, str): 126 # if isinstance(suffix_key, str):
125 # TODO suffix_key校验与修正 127 # # TODO suffix_key校验与修正
126 # TODO 目前只考虑了split的情况 128 # # TODO 目前只考虑了split的情况
127 return suffix_key, key_x0, key_y0, key_x1, key_y1, () 129 # if isinstance(length, int):
130 # if -3 < length - len(suffix_key) < 3:
131 # return suffix_key, (
132 # key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src,
133 # key_y3_src), ()
134 # else:
135 # return suffix_key, (
136 # key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src,
137 # key_y3_src), ()
128 138
129 height = key_y1 - key_y0 139 # 坐标系转换
140 cos, sin = self.get_theta(key_x0_src, key_y0_src, key_x1_src, key_y1_src)
141 key_x0, key_y0, key_x1, key_y1, key_x2, key_y2, key_x3, key_y3 = self.rebuild_coord(
142 (key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src), cos,
143 sin)
144
145 height = key_y2 - key_y0
130 y_min = key_y0 - (top_or_left * height) 146 y_min = key_y0 - (top_or_left * height)
131 y_max = key_y1 + (bottom_or_right * height) 147 y_max = key_y2 + (bottom_or_right * height)
132 148
133 width = key_x1 - key_x0 149 width = key_x2 - key_x0
134 x_max = key_x0 - (offset * width) 150 x_max = key_x0 - (offset * width)
135 x_min = x_max - (width * scope_tuple[-1]) if key_scope is None else key_scope 151 x_min = x_max - (width * scope_tuple[-1]) if key_scope_tuple is None else \
152 self.rebuild_xy(*key_scope_tuple, cos, sin)[0]
136 153
137 all_find_value_list = [] 154 all_find_value_list = []
138 for go_key_idx, ((x0, y0, _, _, x1, y1, _, _), text) in go_res.items(): 155 for go_key_idx, ((x0, y0, x1, y1, x2, y3, x3, y3), text) in go_res.items():
139 cent_x = x0 + ((x1 - x0) / 2) 156 cent_x, cent_y = self.rebuild_xy(x0 + ((x2 - x0) / 2), y0 + ((y2 - y0) / 2), cos, sin)
140 cent_y = y0 + ((y1 - y0) / 2)
141 # if go_key_idx == '98' and key_idx == 34: 157 # if go_key_idx == '98' and key_idx == 34:
142 # print(key_scope)
143 # print('-------------')
144 # print(cent_x) 158 # print(cent_x)
145 # print(cent_y) 159 # print(cent_y)
146 # print('-----------') 160 # print('-----------')
...@@ -153,15 +167,15 @@ class Retriever: ...@@ -153,15 +167,15 @@ class Retriever:
153 # print(x_max) 167 # print(x_max)
154 # print(y_min) 168 # print(y_min)
155 # print(y_max) 169 # print(y_max)
156 # print('===============')
157 if x_min < cent_x < x_max and y_min < cent_y < y_max: 170 if x_min < cent_x < x_max and y_min < cent_y < y_max:
158 all_find_value_list.append((text, x0, y0, x1, y1, (go_key_idx, ))) 171 all_find_value_list.append((text, x0, y0, x1, y1, x2, y2, x3, y3, (go_key_idx,)))
159 172
160 if len(all_find_value_list) == 0: 173 if len(all_find_value_list) == 0:
161 return 174 return
162 elif len(all_find_value_list) == 1: 175 elif len(all_find_value_list) == 1:
163 return all_find_value_list[0] 176 return all_find_value_list[0]
164 else: 177 else:
178 # TODO choice时的坐标转换?
165 choice_value = getattr(self, 'choice_{0}'.format(choice_method))(all_find_value_list, length) 179 choice_value = getattr(self, 'choice_{0}'.format(choice_method))(all_find_value_list, length)
166 return choice_value 180 return choice_value
167 181
...@@ -176,7 +190,8 @@ class Retriever: ...@@ -176,7 +190,8 @@ class Retriever:
176 if self.find_keys_list[key_idx] is None: 190 if self.find_keys_list[key_idx] is None:
177 return 191 return
178 192
179 _, _, find_key_str, suffix_key, key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src = self.find_keys_list[key_idx] 193 _, _, find_key_str, suffix_key, key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, \
194 key_x3_src, key_y3_src = self.find_keys_list[key_idx]
180 195
181 for scope_key_idx in scope_tuple[:-1]: 196 for scope_key_idx in scope_tuple[:-1]:
182 if self.find_keys_list[scope_key_idx] is None: 197 if self.find_keys_list[scope_key_idx] is None:
...@@ -197,8 +212,9 @@ class Retriever: ...@@ -197,8 +212,9 @@ class Retriever:
197 return suffix_key, (key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src), () 212 return suffix_key, (key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src), ()
198 213
199 # 坐标系转换 214 # 坐标系转换
200 cos, sin = self.get_theta(x0, y0, x1, y1) 215 cos, sin = self.get_theta(key_x0_src, key_y0_src, key_x1_src, key_y1_src)
201 key_x0, key_y0, key_x1, key_y1, key_x2, key_y2, key_x3, key_y3 = self.rebuild_coord((key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src), cos, sin) 216 key_x0, key_y0, key_x1, key_y1, key_x2, key_y2, key_x3, key_y3 = self.rebuild_coord(
217 (key_x0_src, key_y0_src, key_x1_src, key_y1_src, key_x2_src, key_y2_src, key_x3_src, key_y3_src), cos, sin)
202 218
203 height = key_y2 - key_y0 219 height = key_y2 - key_y0
204 y_min = key_y0 - (top_or_left * height) 220 y_min = key_y0 - (top_or_left * height)
...@@ -206,7 +222,8 @@ class Retriever: ...@@ -206,7 +222,8 @@ class Retriever:
206 222
207 width = key_x2 - key_x0 223 width = key_x2 - key_x0
208 x_min = key_x2 + (offset * width) 224 x_min = key_x2 + (offset * width)
209 x_max = x_min + (width * scope_tuple[-1]) if key_scope_tuple is None else self.rebuild_xy(*key_scope_tuple, cos, sin)[0] 225 x_max = x_min + (width * scope_tuple[-1]) if key_scope_tuple is None else self.rebuild_xy(
226 *key_scope_tuple, cos, sin)[0]
210 227
211 all_find_value_list = [] 228 all_find_value_list = []
212 for go_key_idx, ((x0, y0, x1, y1, x2, y3, x3, y3), text) in go_res.items(): 229 for go_key_idx, ((x0, y0, x1, y1, x2, y3, x3, y3), text) in go_res.items():
......
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!