-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest_tmp.py
2115 lines (1816 loc) · 77.6 KB
/
test_tmp.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
# -*- coding:utf-8 -*-
import math
import copy
from functools import reduce
import argparse
import numpy as np
import util
from apphelper.image import order_point, calc_iou, xy_rotate_box
from crnn import crnnRec
from eval.invoice.eval_invoice import evaluate
def get_right_marge(data):
# 排除备注中的信息，以图片的中间为限
max_r = 0
for dd in data:
tem_r = float(dd['cx']) + float(dd['w']) / 2
if tem_r > max_r:
max_r = tem_r
return max_r + 200
def get_min_distance_index(tem,indexes,near):
x = float(tem['cx']) + float(tem['w']) / 2
y = float(tem['cy']) + float(tem['h']) / 2
distance_min = 100000
ii = 0
for i in indexes:
x_tem = float(near[i]['cx']) - float(near[i]['w']/2)
y_tem = float(near[i]['cy']) + float(near[i]['h']/2)
distance_tem = math.sqrt(math.pow((x-x_tem),2) + math.pow((y-y_tem),2))
if distance_tem < distance_min and len(near[i]['text']) and calc_axis_iou(near[i],tem,1) > 0.3:
ii = i
distance_min = distance_tem
# if distance_tem < 30 and re.search(r'[\u4e00-\u9fa5]{6,}',near[i]['text']):
# return i
return ii
def format(data):
return data
def extend_3c(img):
img = img.reshape(img.shape[0], img.shape[1], 1)
img = np.concatenate((img, img, img), axis=2)
return img
def debug(idx, img_paths, imgs, output_root):
if not os.path.exists(output_root):
os.makedirs(output_root)
col = []
for i in range(len(imgs)):
row = []
for j in range(len(imgs[i])):
# img = cv2.copyMakeBorder(imgs[i][j], 3, 3, 3, 3, cv2.BORDER_CONSTANT, value=[255, 0, 0])
row.append(imgs[i][j])
res = np.concatenate(row, axis=1)
col.append(res)
res = np.concatenate(col, axis=0)
img_name = img_paths[idx].split('/')[-1]
print(idx, '/', len(img_paths), img_name)
cv2.imwrite(output_root + img_name, res)
def write_result_as_txt(image_name, bboxes, path):
filename = util.io.join_path(path, 'res_%s.txt' % (image_name))
lines = []
for b_idx, bbox in enumerate(bboxes):
values = [int(v) for v in bbox]
line = "%d, %d, %d, %d, %d, %d, %d, %d\n" % tuple(values)
lines.append(line)
util.io.write_lines(filename, lines)
def polygon_from_points(points):
"""
Returns a Polygon object to use with the Polygon2 class from a list of 8 points: x1,y1,x2,y2,x3,y3,x4,y4
"""
resBoxes = np.empty([1, 8], dtype='int32')
resBoxes[0, 0] = int(points[0])
resBoxes[0, 4] = int(points[1])
resBoxes[0, 1] = int(points[2])
resBoxes[0, 5] = int(points[3])
resBoxes[0, 2] = int(points[4])
resBoxes[0, 6] = int(points[5])
resBoxes[0, 3] = int(points[6])
resBoxes[0, 7] = int(points[7])
pointMat = resBoxes[0].reshape([2, 4]).T
return plg.Polygon(pointMat)
# -----------------------------------------------------------------------------------
import os
import sys
import pathlib
__dir__ = pathlib.Path(os.path.abspath(__file__))
sys.path.append(str(__dir__))
sys.path.append(str(__dir__.parent.parent))
# project = 'DBNet.pytorch' # 工作项目根目录
# sys.path.append(os.getcwd().split(project)[0] + project)
import time
import cv2
import torch
from dbnet.data_loader import get_transforms
from dbnet.models import build_model
from dbnet.post_processing import get_post_processing
def resize_image(img, short_size):
height, width, _ = img.shape
if height < width:
new_height = short_size
new_width = new_height / height * width
else:
new_width = short_size
new_height = new_width / width * height
new_height = int(round(new_height / 32) * 32)
new_width = int(round(new_width / 32) * 32)
resized_img = cv2.resize(img, (new_width, new_height))
return resized_img
class Pytorch_model:
def __init__(self, model_path, post_p_thre=0.7, gpu_id=None):
'''
初始化pytorch模型
:param model_path: 模型地址(可以是模型的参数或者参数和计算图一起保存的文件) model_path='/home/share/gaoluoluo/dbnet/output/DBNet_resnet18_FPN_DBHead/checkpoint/model_latest.pth'
:param gpu_id: 在哪一块gpu上运行
'''
self.gpu_id = gpu_id
if self.gpu_id is not None and isinstance(self.gpu_id, int) and torch.cuda.is_available():
self.device = torch.device("cuda:%s" % self.gpu_id)
else:
self.device = torch.device("cpu")
# print('device:', self.device)
checkpoint = torch.load(model_path, map_location=self.device)
# print("checkpoint:",checkpoint)
config = checkpoint['config']
# print(checkpoint['config'])
config['arch']['backbone']['pretrained'] = False
self.model = build_model(config['arch'])
self.post_process = get_post_processing(config['post_processing'])
self.post_process.box_thresh = post_p_thre
self.img_mode = config['dataset']['train']['dataset']['args']['img_mode']
self.model.load_state_dict(checkpoint['state_dict'])
self.model.to(self.device)
self.model.eval()
self.transform = []
for t in config['dataset']['train']['dataset']['args']['transforms']:
if t['type'] in ['ToTensor', 'Normalize']:
self.transform.append(t)
self.transform = get_transforms(self.transform)
def predict(self, img, is_output_polygon=False, short_size: int = 1024):
'''
对传入的图像进行预测，支持图像地址,opecv 读取图片，偏慢
:param img_path: 图像地址
:param is_numpy:
:return:
'''
# assert os.path.exists(img_path), 'file is not exists'
# img = cv2.imread(img_path, 1 if self.img_mode != 'GRAY' else 0)
if self.img_mode == 'RGB':
img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
h, w = img.shape[:2] # 2550 3507
img = resize_image(img, short_size)
# 将图片由(w,h)变为(1,img_channel,h,w)
tensor = self.transform(img)
tensor = tensor.unsqueeze_(0)
tensor = tensor.to(self.device)
batch = {'shape': [(h, w)]}
with torch.no_grad():
if str(self.device).__contains__('cuda'):
torch.cuda.synchronize(self.device)
start = time.time()
preds = self.model(tensor)
if str(self.device).__contains__('cuda'):
torch.cuda.synchronize(self.device)
box_list, score_list = self.post_process(batch, preds, is_output_polygon=is_output_polygon)
box_list, score_list = box_list[0], score_list[0]
if len(box_list) > 0:
if is_output_polygon:
idx = [x.sum() > 0 for x in box_list]
box_list = [box_list[i] for i, v in enumerate(idx) if v]
score_list = [score_list[i] for i, v in enumerate(idx) if v]
else:
idx = box_list.reshape(box_list.shape[0], -1).sum(axis=1) > 0 # 去掉全为0的框
box_list, score_list = box_list[idx], score_list[idx]
else:
box_list, score_list = [], []
t = time.time() - start
return preds[0, 0, :, :].detach().cpu().numpy(), box_list, score_list, t
def init_args():
import argparse
parser = argparse.ArgumentParser(description='DBNet.pytorch')
# parser.add_argument('--model_path', default=r'/home/share/gaoluoluo/dbnet/output/DBNet_resnet50_FPN_DBHead/checkpoint/model_latest.pth', type=str)
# parser.add_argument('--model_path', default=r'/home/share/gaoluoluo/模型/model50_ch_epoch510_latest.pth', type=str)
parser.add_argument('--model_path', default=r'./models/detective/model50_ch_epoch248_latest.pth', type=str)
# parser.add_argument('--input_folder', default='/home/share/gaoluoluo/invoices_gao_true/img/', type=str, help='img path for predict')
parser.add_argument('--input_folder', default='./data/img_not_use/', type=str, help='img path for predict')
# parser.add_argument('--input_folder', default='/home/share/gaoluoluo/complete_ocr/data/img_debug/', type=str, help='img path for predict')
parser.add_argument('--img_correct',default='./invoices_gao_true/corre_gao/',type=str,help='img_correct path for predict')
# parser.add_argument('--input_folder',default='/home/share/gaoluoluo/dbnet/test/test_corre_input', type=str, help='img path for predict')
# parser.add_argument('--input_folder', default='/home/share/gaoluoluo/complete_ocr/data/images/tmp', type=str,help='img path for predict')
parser.add_argument('--output_folder', default='./dbnet/test/test_output', type=str, help='img path for output')
parser.add_argument('--gt_txt', default='/home/share/gaoluoluo/complete_ocr/data/txt_not_use', type=str, help='img 对应的 txt')
parser.add_argument('--thre', default=0.1, type=float, help='the thresh of post_processing')
parser.add_argument('--polygon', action='store_true', help='output polygon or box')
parser.add_argument('--show', action='store_true', help='show result')
parser.add_argument('--save_result', action='store_true', help='save box and score to txt file')
parser.add_argument('--evaluate', nargs='?', type=bool, default=True,
help='evalution')
args = parser.parse_args()
return args
def test():
import pathlib
from tqdm import tqdm
import matplotlib.pyplot as plt
from dbnet.utils.util import show_img, draw_bbox, save_result, get_file_list
args = init_args()
# print(args)
os.environ['CUDA_VISIBLE_DEVICES'] = str('0')
# 初始化网络 0.1
model = Pytorch_model(args.model_path, post_p_thre=args.thre, gpu_id=0)
img_folder = pathlib.Path(args.input_folder) # dbnet/test/input/
# i = 0
total_frame = 0.0
total_time = []
precisions = []
names = []
names_tem = []
# iii = 1
for img_path in tqdm(get_file_list(args.input_folder, p_postfix=['.jpg','.png'])): # img_path /home/share/gaoluoluo/dbnet/test/input/2018实验仪器发票.jpg
# print("img_path:",img_path) /home/share/gaoluoluo/dbnet/test/input/2018实验仪器发票.jpg
print("\nimg_path:", img_path)
start_time = time.time()
from angle import angle_corre
img = angle_corre(img_path)# 调整图片角度
# img = cv2.imread(img_path)
img_path = args.img_correct + img_path.split('/')[-1] # After correct angle img path
names.append(img_path.split('/')[-1])
# names_tem.append(str(iii)+img_path.split('.')[-1])
# iii += 1
print("img_path:",img_path)
preds, boxes_list, score_list, t = model.predict(img, is_output_polygon=args.polygon)
img_path1 = img_path
boxes_list_save = boxes_list
box = []
for i in range(0,len(boxes_list)):
for j in range(0,len(boxes_list[0])):
b = []
b.append(np.float32(boxes_list[i][j][0]))
b.append(np.float32(boxes_list[i][j][1]))
box.append(b)
boxes_list = boxes_list.reshape(-1,2)
boxes_list = box
#---合框只能处理一个框被分成两个部分
i = 4
points_kuang = []
remove_mark = []
max_X = -1
while(i<=len(boxes_list)):
points = boxes_list[i-4:i]
for _,p in enumerate(points):
if p[0] > max_X:
max_X = p[0]
i = i+4
points = np.array(points)
points_kuang.append(points)
for i in range(10,len(points_kuang)-10):
point3 = points_kuang[i][2]
start_point = i - 8
end_point = i + 8
if start_point < 0:
start_point = 0
if end_point > len(points_kuang):
end_point = len(points_kuang)
if i not in remove_mark:
for j in range(start_point,end_point):
point4 = points_kuang[j][3]
min_dis = math.sqrt(math.pow((point3[0] - point4[0]),2) + math.pow((point3[1] - point4[1]),2))
Y_cha = math.fabs(point3[1] - point4[1])
if min_dis < 15 and point4[0] > max_X / 2 and Y_cha < 25 and j not in remove_mark and i != j: # 10 reasonable
# if min_dis < 0:
point1_1 = points_kuang[i][0]
point1_2 = points_kuang[j][0]
x_min = min(point1_1[0],point1_2[0])
y_min = min(point1_1[1],point1_2[1])
point3_2 = points_kuang[j][2]
x_max = max(point3[0],point3_2[0])
y_max = max(point3[1],point3_2[1])
points_kuang[i][0,0] = x_min
points_kuang[i][0,1] = y_min
points_kuang[i][1,0] = x_max
points_kuang[i][1,1] = y_min
points_kuang[i][2,0] = x_max
points_kuang[i][2,1] = y_max
points_kuang[i][3,0] = x_min
points_kuang[i][3,1] = y_max
remove_mark.append(j)
# print("i=",i," j=",j)
# print('x:',point4[0])
break
remove_mark = sorted(remove_mark,reverse=True)
# print("---------------------------")
for _,i in enumerate(remove_mark):
# print('x:',points_kuang[i][3][0])
del points_kuang[i]
boxes_list_save = points_kuang # 决定保存的画框图片是否是 合框之后的图片
#---
i = 0;
rects = []
while(i<len(points_kuang)):
points = points_kuang[i]
rect = cv2.minAreaRect(points) # 4个点 -> d cx cy w h
rec = []
rec.append(rect[-1])
rec.append(rect[1][1])
rec.append(rect[1][0])
rec.append(rect[0][0])
rec.append(rect[0][1])
rects.append(rec)
i += 1
ori_img = cv2.imread(img_path1)
result = crnnRec(cv2.cvtColor(ori_img, cv2.COLOR_BGR2RGB), rects)
# result = crnnRec_tmp(cv2.cvtColor(ori_img, cv2.COLOR_BGR2RGB), rects)
result = list(reversed(result))
dur = time.time() - start_time
print("dur:",dur)
# 保存画框图片---------
# img = draw_bbox(img(img_path)[:, :, ::-1], boxes_list_save)
img = draw_bbox(cv2.imread(img_path)[:, :, ::-1], boxes_list_save)
if args.show:
show_img(preds)
show_img(img, title=os.path.basename(img_path))
plt.show()
# 保存结果到路径
os.makedirs(args.output_folder, exist_ok=True)
img_path = pathlib.Path(img_path)
output_path = os.path.join(args.output_folder, img_path.stem + '_result.jpg') # /home/share/gaoluoluo/dbnet/test/output/2018实验仪器发票_result.jpg
pred_path = os.path.join(args.output_folder,img_path.stem + '_pred.jpg') # /home/share/gaoluoluo/dbnet/test/output/2018实验仪器发票_pred.jpg
cv2.imwrite(output_path, img[:, :, ::-1])
cv2.imwrite(pred_path, preds * 255)
save_result(output_path.replace('_result.jpg', '.txt'), boxes_list_save, score_list, args.polygon)
# --------
total_frame += 1
total_time.append(dur)
result = formatResult(result)
# 计算准确率
if args.evaluate:
file = os.path.join(args.gt_txt,img_path.stem + '.txt')
errfile = os.path.join(args.output_folder, 'error_txt/' + img_path.stem + '-errors.txt')
canfile = os.path.join(args.output_folder, 'candidate_txt/' + img_path.stem + '-candidate.txt')
if os.path.exists(file):
precision = evaluate(file, errfile,canfile,result)
print('precision:' + str(precision[0]) + '%')
precisions.append(precision[0])
for i in range(0,len(precisions)):
# print(names[i],precisions[i])
print(precisions[i])
if len(precisions):
mean = np.mean(precisions)
print("names:",names)
# print("names:",names_tem)
print("precisions:",precisions)
time_mean = np.mean(total_time)
print('mean precision:' + str(mean) + '%')
print('mean time:' + str(time_mean) + 's')
return 0;
#--------------------------------------------------
import re
def isTitle(text):
# return re.search(r'(货物|劳务|服.?名称|规.?型号|^单[位价]|^数|.?额$|^税.$|项目|类型|车牌|通行日期)', text) != None
# return re.search(r'(货物|劳务|服.?名称|规.?型号|^单[位价]|^数|.?额$|税率|税额|项目|类型|车牌|通行日期)', text) != None
return re.search(r'(货物|劳务|服.?名称|规.?型号|^单[位价]|^数量|税率|税额)', text) != None
return re.search(r'(货物|劳务|服.?名称|规.?型号|^单[位价]|^数量|.?额$|税率|税额|项目|类型|车牌|通行日期)', text) != None
def isSummary(text):
# pattern = re.compile(r'[￥|Y|羊]\d+?\.?\d*')
# pattern = re.compile(r'(￥|Y|羊)[0-9]+(.[0-9]{2})?$')
pattern = re.compile(r'(￥|Y|羊)[0-9]+(.[0-9]{2})$')
return text==r'计' or text == r'合' or text == r'合计' or pattern.search(text) != None
def get_content_boundary(data):
#
#--- 过滤左边的无用信息
left = [d for i, d in enumerate(data) if isTitle(d['text'])]
left = sorted(left,key=lambda x: x['cx'] - x['w'] / 2)
left = left[0]
cx_tem = float(left['cx'] - float(left['w']) / 2) # 标题的左
if re.search(r'(货物|劳务|服.?名称)', left['text']) != None:
for i in range(len(data)-1,-1,-1):
cx = float(data[i]['cx']) + float(data[i]['w']) / 2 # 其他的右
if(0 < cx_tem -cx):
# print(data[i])
data.pop(i)
#----
#----
while len(data):
left_tem = min(data, key = lambda x : x['cx'] - x['w'] / 2)
if len(left_tem['text']) < 3 :
data.remove(left_tem)
else :
break
while len(data):
right_tem = max(data, key = lambda x : x['cx'] + x['w'] / 2)
if len(right_tem['text']) < 3 :
data.remove(right_tem)
else :
break
#----
title = [i for i, d in enumerate(data) if isTitle(d['text'])]
# print(data[31]['text'])
s = title[0]
e = title[-1]
title.extend([i+s-4 for i,d in enumerate(data[s-4:s]) if abs(d['cy'] - data[s]['cy']) < 10])
title.extend([i+e+1 for i,d in enumerate(data[e+1:e+6]) if abs(d['cy'] - data[e]['cy']) < 10])
s = min(title)
e = max(title)
lf = min(data[s:e+1], key=lambda x:x['cx']-x['w']/2)
summary = [i for i, d in enumerate(data) if isSummary(d['text'])]
s = summary[0]
summary.extend([i+s-4 for i, d in enumerate(data[s-4:s]) if abs(d['cy'] - data[summary[-1]]['cy']) < 10])
s = min(summary)
start = e
# print("start:",start)
end = s
# print("end:",end)
# rt = [d for d in data[start:end] if re.match(r'\d+\.?\d*$', d['text'])]
rt = max(data[start:end], key=lambda x: x['cx']+x['w']/2)
left = lf['cx']-lf['w']/2 # 20为经验值
right = rt['cx']+rt['w']/2 # 80 为经验值
return (start, end, left, right)
def check(data, placeholder, dir = 0):
try:
i = None
if isinstance(placeholder,list):
for pld in placeholder:
f = [x for x,d in enumerate(data) if d == pld]
if len(f):
i = f[-1]
break
else:
f = [x for x,d in enumerate(data) if d == placeholder]
if len(f):
i = f[-1]
return i
except:
return None
LEFT_MARGIN = 70
RIGHT_MARGIN = 20
def parseLine(line, boundary, isvehicle=False):
xmin, xmax = boundary
copyed = copy.deepcopy(line)
copyed = preprocess_line(copyed) # 主要处理的是框框交叉,并对cx排序
if isvehicle:
# ratio, price = get_ratio(copyed, xmax)
ratio, price, tax = get_ratio_price_tax(copyed, xmax)
# title
title = get_title(copyed, xmin)
sdate, edate, price, ratio = get_date(copyed, price, ratio)
platenum, cartype, sdate, edate = get_platenum_cartype(copyed, sdate, edate)
return postprocess_line(title, platenum, cartype, sdate, edate, price, ratio, tax)
else:
# ratio
# ratio, price = get_ratio(copyed, xmax)
ratio, price, tax = get_ratio_price_tax(copyed, xmax)
# title
title = get_title(copyed, xmin) # jin
# tax
# tax = get_tax(copyed, xmax, ratio)
# prices
#
specs, amount, uprice, price, ratio = get_numbers(copyed, price, ratio,tax)
#specs
specs,unit = get_specs_unit(copyed, specs)
return postprocess_line(title, specs, unit, amount, uprice, price, ratio, tax)
def preprocess_line(line):
line = sorted(line, key=lambda x: x['cx'] - x['w'] / 2)
res = []
i = 0
j = 1
while i < len(line) and j < len(line):
x = line[i]
y = line[j]
x1 = x['cx'] + x['w'] / 2 # 左的右边
y1 = y['cx'] - y['w'] / 2 # 右的左边
if abs(x1 - y1) < 8: # 说明是同一个框 合并
x['w'] = y['cx'] + y['w']/2 - x['cx'] + x['w']/2
x['cx'] = y['cx'] + y['w']/2 - x['w']/2
x['text'] = x['text'] + y['text']
j = j + 1
else: # 不是同一个框，继续向you遍历
res.append(x)
i = j
j = j + 1
res.append(line[i])
return res
def is_number(s):
try:
float(s)
return True
except ValueError:
pass
return False
def postprocess_line(title, specs, unit, amount, uprice, price, ratio, tax):
if tax != '***' and is_number(price) and ratio and (not tax or not is_number(tax) or float(tax) > float(price) or not '.' in tax) :
tax = '%.2f' % (float(price) / 100 * float(ratio[:-1]))
return [title, specs, unit, amount, uprice, price, ratio, tax]
def get_title(line, xmin):
title = ''
candidates = [d for d in line if abs(d['cx'] - d['w'] / 2 - xmin) < LEFT_MARGIN]
if len(candidates):
candidates = sorted(candidates, key=lambda x:x['cy'])
for c in candidates:
title += c['text']
line.remove(c)
if title and not title.startswith('*'):
title = '*' + title
return title
def get_ratio_price_tax(line, xmax):
ratio = ''
price = ''
tax = ''
pat = re.compile(r'(\-?[\dBG]+\.?[\dBG]{2})*(([\dBG]+[\%])|([\u4e00-\u9fa5]+税))$')
pat2 = re.compile(r'(\-?[\dBG]+\.[\dBG]{2})([\dBG]{1,2}[8])')
ratioitem = None
for i in range(len(line)-1, -1, -1):
text = line[i]['text']
m = pat.match(text)
if m:
price, ratio = (m.group(1), m.group(2)) if m.group(1) else ('', m.group(2))
else:
if len(line) - i < 3: # reduce
m = pat2.match(text)
if m:
price, ratio = (m.group(1), m.group(2))
ratio = ratio[:-1] + '%'
if ratio:
ratioitem = line.pop(i)
break
if not ratio:
numbers = sorted([i for i,d in enumerate(line) if re.match(r'([\dBG]+\.?[\dBG]{2})*', d['text'])], key=lambda x:line[x]['cx']-line[x]['w']/2)
if len(numbers)>=3:
i = numbers[-2]
d = line[i]
m = re.match(r'(\d{1,2})\D+', d['text']) # 1-2 number and !number char eg 11%
if m:
ratio = m.group(1)
ratioitem = line.pop(i)
if re.search(r'税$', ratio):
tax = '***'
else:
if ratioitem:
taxes = [l for l in line if l['cx'] > ratioitem['cx']]
if len(taxes):
tax = taxes[0]['text']
line.remove(taxes[0])
if not tax:
x = 1000
for i,d in enumerate(line):
if abs(d['cx'] + d['w'] / 2 - xmax) < x:
x = abs(d['cx'] + d['w'] / 2 - xmax)
idx = [i for i, d in enumerate(line) if abs(d['cx'] + d['w'] / 2 - xmax) < RIGHT_MARGIN]
if len(idx):
idx = idx[0]
ln = line[idx]
tax = ln['text']
line.pop(idx)
if len(tax) > 2 and tax.find('.') == -1:
tax = tax[:-2] + '.' + tax[-2:]
if len(price) and not '.' in price:
if tax and ratio:
while float(price) > float(tax):
prc = price[:-2] + '.' + price[-2:]
f_tax = float(tax)
f_ratio = float(ratio[:-1])
f_price = float(prc)
if abs(f_price * f_ratio / 100.0 - f_tax) > 0.1 and f_ratio < 10:
ratio = price[-1] + ratio
price = price[:-1]
else:
price = prc
break
else:
price = price[:-2] + '.' + price[-2:]
if price and ratio and not tax:
tax = str(round(float(price) * float(ratio[:-1]) / 100.0, 2))
return ratio, price, tax
def get_tax(line, xmax, ratio):
tax = ''
if re.search(r'税$', ratio):
tax = '***'
idx = [i for i, d in enumerate(line) if abs(d['cx'] + d['w'] / 2 - xmax) < RIGHT_MARGIN]
if len(idx):
idx = idx[0]
ln = line[idx]
tax = ln['text']
line.pop(idx)
if len(tax) > 2 and tax.find('.') == -1:
tax = tax[:-2] + '.' + tax[-2:]
return tax
def get_ratio(line, xmax):
ratio = ''
price = ''
pat = re.compile(r'(\-?[\dBG]+\.?[\dBG]{2})*(([\dBG]+[\%])|([\u4e00-\u9fa5]+税))$')
for i in range(len(line)-1, -1, -1):
text = line[i]['text']
m = pat.match(text)
if m:
price, ratio = (m.group(1), m.group(2)) if m.group(1) else ('', m.group(2))
if ratio:
line.pop(i)
break
if not ratio:
numbers = sorted([i for i,d in enumerate(line) if re.match(r'([\dBG]+\.?[\dBG]{2})*', d['text'])], key=lambda x:line[x]['cx']-line[x]['w']/2)
if len(numbers)>=3:
i = numbers[-2]
d = line[i]
m = re.match(r'(\d{1,2})\D+', d['text'])
if m:
ratio = m.group(1) + '%'
line.pop(i)
return ratio, price
def get_numbers(line, price, ratio,tax):
specs = ''
amount = ''
uprice = ''
pattern = re.compile(r'\-?[\dBG:]+\.?[\dBG]*$')
numbers = []
for d in line:
if pattern.match(d['text']):
d['text'] = d['text'].replace('B','8').replace('G', '6').replace(':','')
numbers.append(d)
if len(numbers):
for n in numbers:
line.remove(n)
# preprocess_number(numbers)
numbers = sorted(numbers, key=lambda x: x['cx'] - x['w'] / 2)
if not ratio and re.match(r'^\d{2,3}$', numbers[-1]['text']):
ratio = numbers[-1]['text']
ratio = ratio[:-1] + '%'
numbers = numbers[0:-1]
if not price:
# print("*price:",price)
# print("len:",len(numbers[-1]['text']))
if(len(numbers)):
price = numbers[-1]['text']
m = re.match(r'(\d+\.\d{2})\d*(\d{2})$', price)
if m and not ratio:
price = m.group(1)
ratio = m.group(2) + '%'
numbers = numbers[0:-1]
numlen = len(numbers)
if numlen == 3:
specs = numbers[0]['text']
amount = numbers[1]['text']
uprice = numbers[2]['text']
elif numlen == 2:
num1 = numbers[0]['text']
num2 = numbers[1]['text']
if abs(float(num1) * float(num2) - float(price)) < 0.01:
specs = ''
amount = num1
uprice = num2
elif abs(float(num2) - float(price)) < 0.01:
specs = num1
amount = '1'
uprice = num2
else:
specs, amount, uprice = get_amount_uprice(price, num2, num1)
elif numlen == 1:
specs = ''
num = numbers[0]['text']
if abs(float(num) - float(price)) < 0.01:
amount = '1'
uprice = num
else:
specs, amount, uprice = get_amount_uprice(price, num)
if not amount:
if uprice:
amount = str(int(float(price) / float(uprice) + 0.5))
else:
amount = num
if not ratio and price and tax:
ratio = str(int(float(tax) * 100 / float(price) + 0.5)) + '%'
return specs, amount, uprice, price, ratio
def get_date(line, price, ratio):
sdate = ''
edate = ''
pattern = re.compile(r'\-?[\dBG:]+\.?[\dBG]*$')
numbers = []
for d in line:
if pattern.match(d['text']):
d['text'] = d['text'].replace('B','8').replace('G', '6').replace(':','')
numbers.append(d)
if len(numbers):
for n in numbers:
line.remove(n)
# preprocess_number(numbers)
numbers = sorted(numbers, key=lambda x: x['cx'] - x['w'] / 2)
if not ratio and re.match(r'^\d{2,3}$', numbers[-1]['text']):
ratio = numbers[-1]['text']
ratio = ratio[:-1] + '%'
numbers = numbers[0:-1]
if not price:
price = numbers[-1]['text']
m = re.match(r'(\d+\.\d{2})\d*(\d{2})$', price)
if m and not ratio:
price = m.group(1)
ratio = m.group(2) + '%'
numbers = numbers[0:-1]
numlen = len(numbers)
if numlen == 2:
sdate = numbers[0]['text']
edate = numbers[1]['text']
elif numlen == 1:
edate = numbers[0]['text']
return sdate, edate, price, ratio
def get_platenum_cartype(line, sdate, edate):
platenum = ''
cartype = ''
pattern = re.compile(r'([\u4e00-\u9fa5]+)(\d{8,})$')
if len(line) == 2:
platenum = line[0]['text']
cartype = line[1]['text']
elif len(line) == 1:
if not sdate:
cartype = line[0]['text']
else:
platenum = line[0]['text']
if cartype and not sdate:
m = pattern.match(cartype)
if m:
cartype, sdate = m.group(1), m.group(2)
if len(sdate) > 8 and not edate:
edate = sdate[8:]
sdate = sdate[:8]
return platenum, cartype, sdate, edate
def preprocess_number(numbers):
number = [i for i,n in enumerate(numbers) if n['text'].find(':')>-1]
adds = []
removes = []
for i in number:
d = numbers[i]
text = d['text']
splits = text.split(':')
d1 = dict(d)
d1['text'] = splits[0]
d1['w'] = d['w'] * len(d1['text']) / len(text)
d1['cx'] = d['cx'] - d['w'] / 2 + d1['w'] / 2
d2 = dict(d)
d2['text'] = splits[1]
d2['w'] = d['w'] * len(d2['text']) / len(text)
d2['cx'] = d['cx'] + d['w'] / 2 - d2['w'] / 2
removes.append(d)
adds.extend([d1, d2])
for d in removes:
numbers.remove(d)
numbers.extend(adds)
def get_amount_uprice(price, upricecand, amtcand=None):
price = float(price)
specs = ''
amount = ''
uprice = ''
copyprice = upricecand
dotplace = upricecand.find('.')
if dotplace > 0:
upricecand = upricecand[:dotplace] + upricecand[dotplace + 1:]
if amtcand:
upr = str(math.trunc(float(price) / float(amtcand) * 100))
idx = upricecand.find(upr)
if idx >= 0:
amount = amtcand
upricecand = upricecand[idx:]
dot = len(upr[:-2])
uprice = upricecand[:dot] + '.' + upricecand[dot:]
if not uprice:
end = dotplace - 1 if dotplace else len(upricecand) - 2
for idx in range(2, end):
amt = int(upricecand[:idx])
upr = upricecand[idx:]
if not amt:
break
calcupr = price / amt
if calcupr < 1:
break
dot = str(calcupr).find('.')
if dot > len(upr):
break
upr = float(upr[0:dot] + '.' + upr[dot:])
if abs(upr - calcupr) < 1:
amount = str(amt)
uprice = str(upr)
break
if not uprice:
m = re.match(r'(\d+0+)([1-9]\d*\.\d+)', copyprice)
if m:
amount = m.group(1)
uprice = m.group(2)
else:
uprice = copyprice
if amtcand:
amount = amtcand
else:
if amtcand:
specs = amtcand
return specs,amount,uprice
def get_specs_unit(line, specs):
unit = ''
linelen = len(line)
if linelen == 2:
specs = line[0]['text']
unit = line[1]['text']
if linelen == 1:
text = line[0]['text']
if specs:
unit = text
else:
if len(text) == 1:
unit = text
else:
specs = text
return specs,unit
def is_wrapped_title(data, line, boundary):
res = False
xmin = boundary[0]
dx = abs(data['cx'] - data['w'] / 2 - xmin) # 该data的左边与最左边的差的绝对值
text = data['text'] #
if len(text) == 0: # 自己加的
return res
if dx < LEFT_MARGIN and text[0] != '*': # 偏左并且开头不是*
res = True
return res
def check_title(line, data, start, end, boundary):
xmin = boundary[0]
lf = min(line, key=lambda d:d['cx'] - d['w'] / 2)
dx = abs(lf['cx'] - lf['w'] / 2 - xmin)
if dx > LEFT_MARGIN:
for d in data[start:end]:
dx = abs(d['cx'] - d['w'] / 2 - xmin)
if dx < LEFT_MARGIN:
iou = [calc_axis_iou(d,l,1) for l in line]
if np.mean(iou) > 0.3:
line.append(d)
data.remove(d)
break
def check_wrap_title(res, wraptitles, line=None):
title = res[-1][0]
wraplen = len(wraptitles)
if wraplen:
idx = 0
if not line:
wrap = wraptitles[:]
else:
wrap = []
ref = min(line, key=lambda x:x['cx']-x['w']/2)
for i,w in enumerate(wraptitles):
if w['cy'] < ref['cy']:
wrap.append(w)
else:
break
if len(wrap):
del wraptitles[0:len(wrap)]
title = reduce(lambda a,b:a+b, [w['text'] for w in wrap], title)
res[-1][0] = title
def get_basic_boundary(data):
indexes = [i for i, d in enumerate(data) if re.search(r'开.?日期|票日期|校.?码|20.?.?年?.?.?月', d['text'])]
# print(data[71])
if len(indexes):
end = max(indexes)
# if end > 10 :
# end = 10
else:
end = 8
lt = min(data[:end+1]+data[-10:], key=lambda x:x['cx']-x['w']/2)
rt = max(data[:end+1]+data[-10:], key=lambda x:x['cx']+x['w']/2)
left = lt['cx'] - lt['w'] / 2
right = rt['cx'] + rt['w'] / 2
return (0, end, left, right)