Skip to content

周伟奇 / bmw-ocr

Go to a project
Toggle navigation
Toggle navigation pinning
33ea2687 by 周伟奇
Options

add bs verify

1 parent 8e146fd9
Inline Side-by-side
Showing 3 changed files with 44 additions and 8 deletions
src/apps/doc/consts.py
......@@ -1558,7 +1558,7 @@ BL_COMPARE_LOGIC = {
'registeredCapital': ('注册资本', 'se_rmb_compare', {}, '营业执照注册资本与系统不一致'),
}
BL_COMPARE_LOGIC = {
SME_BL_COMPARE_LOGIC = {
'legalRepName': ('经营者姓名', 'se_name_compare', {}, '营业执照法定代表人与系统不一致'),
'businessLicenseDueDate': ('营业期限', 'se_date_compare', {'ocr_split': True, 'long': True, 'ocr_replace': True, 'today': True}, '公司营业期限疑似过期'),
}
......
src/apps/doc/management/commands/ocr_process.py
......@@ -168,6 +168,12 @@ class Command(BaseCommand, LoggerMixin):
words = cell.get('words')
ws.cell(row=r1 + 1, column=c1 + 1, value=words)
# 真伪
verify_info = []
verify_dict = sheet.get('verify', {})
if verify_dict.get('verify_res') == 'fake':
verify_info.extend(verify_dict.get('verify_info', []))
# ['户名', '卡号', '页码', '回单验证码', '打印时间', '起始时间', '终止时间']
summary = sheet.get('summary')
card = summary[1]
......@@ -183,6 +189,7 @@ class Command(BaseCommand, LoggerMixin):
pt_list = role_dict.setdefault('print_time', [])
sd_list = role_dict.setdefault('start_date', [])
ed_list = role_dict.setdefault('end_date', [])
verify_list = role_dict.setdefault('verify', [])
if summary[3] is not None:
code_list.append((summary[2], summary[3]))
if summary[4] is not None:
......@@ -191,6 +198,10 @@ class Command(BaseCommand, LoggerMixin):
sd_list.append(summary[5])
if summary[6] is not None:
ed_list.append(summary[6])
if len(verify_info) > 0:
verify_list.append(
(pno, ino, '、'.join(verify_info))
)
else:
card_dict = bs_summary.setdefault(card, {})
card_dict['count'] = card_dict.get('count', 0) + 1
......@@ -203,6 +214,7 @@ class Command(BaseCommand, LoggerMixin):
pt_list = card_dict.setdefault('print_time', [])
sd_list = card_dict.setdefault('start_date', [])
ed_list = card_dict.setdefault('end_date', [])
verify_list = card_dict.setdefault('verify', [])
if summary[0] is not None:
role_list.append(summary[0])
role_set.add(summary[0])
......@@ -214,6 +226,10 @@ class Command(BaseCommand, LoggerMixin):
sd_list.append(summary[5])
if summary[6] is not None:
ed_list.append(summary[6])
if len(verify_info) > 0:
verify_list.append(
(pno, ino, '、'.join(verify_info))
)
if cells_exists:
res_list.append((pno, ino, part_idx, consts.RES_SUCCESS))
......@@ -833,6 +849,7 @@ class Command(BaseCommand, LoggerMixin):
merged_bs_summary[main_card]['role'].extend(bs_summary[card]['role'])
merged_bs_summary[main_card]['role_set'].update(bs_summary[card]['role_set'])
merged_bs_summary[main_card]['code'].extend(bs_summary[card]['code'])
merged_bs_summary[main_card]['verify'].extend(bs_summary[card]['verify'])
merged_bs_summary[main_card]['print_time'].extend(bs_summary[card]['print_time'])
merged_bs_summary[main_card]['start_date'].extend(bs_summary[card]['start_date'])
merged_bs_summary[main_card]['end_date'].extend(bs_summary[card]['end_date'])
......@@ -882,7 +899,8 @@ class Command(BaseCommand, LoggerMixin):
# 'classify': [],
# 'confidence': [],
# 'role': [],
# 'code': [('page', 'code')],
# 'code': [('page', 'code'), ],
# 'verify': [(pno, ino, reason_str), ]
# 'print_time': [],
# 'start_date': [],
# 'end_date': [],
......@@ -896,7 +914,8 @@ class Command(BaseCommand, LoggerMixin):
# 'classify': 0,
# 'confidence': [],
# 'role': '户名',
# 'code': [('page', 'code')],
# 'code': [('page', 'code'), ],
# 'verify': [(pno, ino, reason_str), ]
# 'print_time': [],
# 'start_date': [],
# 'end_date': [],
......@@ -921,6 +940,7 @@ class Command(BaseCommand, LoggerMixin):
# summary_dict['confidence'].extend(summary['confidence'])
summary_dict['role'] = summary['role']
summary_dict['code'].extend(summary['code'])
summary_dict['verify'].extend(summary['verify'])
summary_dict['print_time'].extend(summary['print_time'])
summary_dict['start_date'].extend(summary['start_date'])
summary_dict['end_date'].extend(summary['end_date'])
......@@ -952,6 +972,7 @@ class Command(BaseCommand, LoggerMixin):
# card_summary['confidence'].extend(summary['confidence'])
card_summary['sheet'].extend(summary['sheet'])
card_summary['code'].extend(summary['code'])
card_summary['verify'].extend(summary['verify'])
card_summary['print_time'].extend(summary['print_time'])
card_summary['start_date'].extend(summary['start_date'])
card_summary['end_date'].extend(summary['end_date'])
......
src/apps/doc/ocr/wb.py
......@@ -20,6 +20,7 @@ class BSWorkbook(Workbook):
self.meta_sheet_title = 'Key info'
self.blank_row = (None,)
self.code_header = ('页数', '电子回单验证码')
self.verify_header = ('页数', '图片序号', '检测内容')
self.date_header = ('打印时间', '起始日期', '终止日期', '流水区间结果')
self.interest_keyword_header = ('结息关键词', '记账日期', '金额')
self.salary_keyword_header = ('收入关键词', '记账日期', '金额')
......@@ -310,19 +311,28 @@ class BSWorkbook(Workbook):
month_mapping.setdefault(item, []).append(
(ws.title, idx_list[i] + min_row, idx_list[i + 1] + min_row - 1, day_mean))
def build_metadata_rows(self, confidence, code, print_time, start_date, end_date, res_count_tuple):
def build_metadata_rows(self, confidence, code, verify_list, print_time, start_date, end_date, res_count_tuple):
if start_date is None or end_date is None:
timedelta = None
else:
timedelta = (end_date - start_date).days
verify_res = '正常'
if len(verify_list) > 0:
verify_res = '疑似伪造'
metadata_rows = [
('流水识别置信度', confidence),
('流水检测结果', verify_res),
('图片总数', res_count_tuple[0]),
('识别成功', res_count_tuple[1]),
self.blank_row,
self.code_header,
]
metadata_rows.extend(code)
if len(verify_list) > 0:
metadata_rows.append(self.verify_header)
metadata_rows.extend(verify_list)
metadata_rows.extend(
[self.blank_row,
self.date_header,
......@@ -332,18 +342,21 @@ class BSWorkbook(Workbook):
)
return metadata_rows
def build_meta_sheet(self, role_name, card, confidence, code, print_time, start_date, end_date, res_count_tuple):
metadata_rows = self.build_metadata_rows(confidence, code, print_time, start_date, end_date, res_count_tuple)
def build_meta_sheet(self, role_name, card, confidence, code, verify_list, print_time, start_date, end_date, res_count_tuple):
metadata_rows = self.build_metadata_rows(confidence, code, verify_list, print_time, start_date, end_date, res_count_tuple)
if not isinstance(role_name, str):
role_name = consts.UNKNOWN_ROLE
ms = self.create_sheet('{0}{1}({2})'.format(self.meta_sheet_title, role_name, card))
for row in metadata_rows:
ms.append(row)
if res_count_tuple[0] != res_count_tuple[1]:
if len(verify_list) > 0:
for cell in ms[2]:
cell.fill = self.amount_fill
if res_count_tuple[0] != res_count_tuple[1]:
for cell in ms[3]:
cell.fill = self.amount_fill
for cell in ms[4]:
cell.fill = self.amount_fill
return ms
@staticmethod
......@@ -608,7 +621,8 @@ class BSWorkbook(Workbook):
# 'classify': 0,
# 'confidence': 0.9,
# 'role': '柳雪',
# 'code': [('page', 'code')],
# 'code': [('page', 'code'), ],
# 'verify': [(pno, ino, reason_str), ]
# 'print_time': 'datetime',
# 'start_date': 'datetime',
# 'end_date': 'datetime',
......@@ -661,6 +675,7 @@ class BSWorkbook(Workbook):
new_card,
confidence,
summary.get('code'),
summary.get('verify'),
summary.get('print_time'),
start_date,
end_date,
......
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!