Skip to content

周伟奇 / bmw-ocr

Go to a project
Toggle navigation
Toggle navigation pinning
9f51de96 by 冯轩
Options

merge 5075-chenyao

2 parents d6fc968c ff8e7380
Inline Side-by-side
Showing 4 changed files with 81 additions and 5 deletions
src/apps/doc/consts.py
......@@ -99,6 +99,7 @@ RES_SUCCESS = '识别成功'
RES_SUCCESS_OTHER = '识别成功（其他类）'
RES_SUCCESS_EMPTY = '识别成功（空数据）'
RES_SUCCESS_FINANCIAL_STATEMENT = '识别成功（财务报表类）'
RES_SUCCESS_DOWN_PAYMENT = '识别成功(首付款支付承诺书类)'
RES_FAILED = '识别失败'
RES_FAILED_1 = '识别失败(阶段1)'
RES_FAILED_2 = '识别失败(阶段2)'
......@@ -2586,6 +2587,13 @@ FINANCIAL_EXPLANATION_CLASSIFY_LIST = [100]
# 财报情况说明sheet名称
FINANCIAL_EXPLANATION_SHEET_NAME = "财报情况说明"
# 首付款支付承诺书分类标签
DOWN_PAYMENT_CLASSIFY_LIST = [96]
# 首付款支付承诺书sheet名称
DOWN_PAYMENT_SHEET_NAME = "首付款支付承诺书"
NEW_FILE_COMPARE_SET = [96]
# Jira-4562 - 银行流水首页提取关键词
INCOME_KEYWORDS_LIST = ["养老金", "社保", "代发工资", "工资入账", "奖金", "养老保险", "代发", "工资"]
INCOME_KEYWORDS_DICT = {
......
src/apps/doc/management/commands/ocr_process.py
......@@ -1768,6 +1768,8 @@ class Command(BaseCommand, LoggerMixin):
financial_statement_dict = {}
# 添加财报情况说明的处理
financial_explanation_dict = {}
# 添加首付款支付承诺书的处理
down_payment_dict = {}
res_list = []
interest_keyword = Keywords.objects.filter(
type=KeywordsType.INTEREST.value, on_off=True).values_list('keyword', flat=True)
......@@ -1835,6 +1837,31 @@ class Command(BaseCommand, LoggerMixin):
if "stamp" in ocr_data:
stamp = ocr_data.get("stamp", "")
financial_explanation_dict["stamp"] = stamp
elif classify in consts.DOWN_PAYMENT_CLASSIFY_LIST:
# 添加到 res_list 中
res_list.append((pno, ino, part_idx, consts.RES_SUCCESS_DOWN_PAYMENT))
# 1-金融机构名称
financial_org_name = ocr_data.get("financial_org_name", "")
down_payment_dict["financial_org_name"] = financial_org_name
# 2-主借人姓名
main_borrower_name = ocr_data.get("main_borrower_name", "")
down_payment_dict["main_borrower_name"] = main_borrower_name
# 3-主借人证件号码
main_borrower_id_no = ocr_data.get("main_borrower_id_no", "")
down_payment_dict["main_borrower_id_no"] = main_borrower_id_no
# 4-申请编号
apply_no = ocr_data.get("apply_no", "")
down_payment_dict["apply_no"] = apply_no
# 5-抵押/租赁合同名称
contract_name = ocr_data.get("contract_name", "")
down_payment_dict["contract_name"] = contract_name
# 6-承诺人签字
promisor_signature = ocr_data.get("promisor_signature", "")
down_payment_dict["promisor_signature"] = promisor_signature
# 7-承诺人签字日期
promisor_signature_date = ocr_data.get("promisor_signature_date", "")
down_payment_dict["promisor_signature_date"] = promisor_signature_date
elif classify in consts.OTHER_CLASSIFY_SET: # 其他类
res_list.append((pno, ino, part_idx, consts.RES_SUCCESS_OTHER))
continue
......@@ -2021,7 +2048,7 @@ class Command(BaseCommand, LoggerMixin):
# src_excel_path = os.path.join(doc_data_path, 'src.xlsx')
# wb.save(src_excel_path)
#need_follow表示在上传edms时文件名是否要添加"关注"两字
count_list, need_follow = wb.rebuild(merged_bs_summary, license_summary, res_list, doc.document_scheme, contract_result, doc.metadata, financial_statement_dict, financial_explanation_dict)
count_list, need_follow = wb.rebuild(merged_bs_summary, license_summary, res_list, doc.document_scheme, contract_result, doc.metadata, financial_statement_dict, financial_explanation_dict, down_payment_dict)
wb.save(excel_path)
except Exception as e:
......@@ -2122,7 +2149,7 @@ class Command(BaseCommand, LoggerMixin):
license_summary[consts.BS_CLASSIFY] = bs_rebuild
# 比对
if len(license_summary) > 0:
if len(license_summary) > 0 or classify in consts.NEW_FILE_COMPARE_SET:
if doc.document_scheme != consts.DOC_SCHEME_LIST[2]:
# if len(license_summary) > 0 and doc.document_scheme != consts.DOC_SCHEME_LIST[2]:
try:
......@@ -2547,7 +2574,7 @@ class Command(BaseCommand, LoggerMixin):
self.online_log.info('{0} [stop safely]'.format(self.log_base))
@transaction.atomic
def atomicSaveDBAFC(self,result_class,doc,license_summary,ic_merge,rp_merge,task_str, financial_statement_dict, financial_explanation_dict):
def atomicSaveDBAFC(self,result_class,doc,license_summary,ic_merge,rp_merge,task_str, financial_statement_dict, financial_explanation_dict, down_payment_dict):
with transaction.atomic('afc'):
res_obj = result_class.objects.using('afc').select_for_update().filter(application_id=doc.application_id).first()
self.online_log.info('{0} [sql lock AFC application_id={1} doc_id={2}]'.format(self.log_base, doc.application_id,doc.id))
......@@ -2563,6 +2590,10 @@ def atomicSaveDBAFC(self,result_class,doc,license_summary,ic_merge,rp_merge,task
if res_obj is not None:
if financial_explanation_dict:
res_obj.fss_ocr = json.dumps([financial_explanation_dict])
# 首付款支付承诺书存入数据库
if res_obj is not None:
if down_payment_dict:
res_obj.dp_ocr = json.dumps([down_payment_dict])
for classify, field in consts.RESULT_MAPPING.items():
if not hasattr(res_obj, field):
continue
......@@ -2588,7 +2619,7 @@ def atomicSaveDBAFC(self,result_class,doc,license_summary,ic_merge,rp_merge,task
return res_obj
@transaction.atomic
def atomicSaveDBHIL(self,result_class,doc,license_summary,ic_merge,rp_merge, task_str, financial_statement_dict, financial_explanation_dict):
def atomicSaveDBHIL(self,result_class,doc,license_summary,ic_merge,rp_merge, task_str, financial_statement_dict, financial_explanation_dict, down_payment_dict):
with transaction.atomic('default'):
res_obj = result_class.objects.using('default').select_for_update().filter(application_id=doc.application_id).first()
self.online_log.info('{0} [sql lock HIL application_id={1} doc_id={2}]'.format(self.log_base, doc.application_id,doc.id))
......@@ -2604,6 +2635,10 @@ def atomicSaveDBHIL(self,result_class,doc,license_summary,ic_merge,rp_merge, tas
if res_obj is not None:
if financial_explanation_dict:
res_obj.fss_ocr = json.dumps([financial_explanation_dict])
# 首付款支付承诺书存入数据库
if res_obj is not None:
if down_payment_dict:
res_obj.dp_ocr = json.dumps([down_payment_dict])
for classify, field in consts.RESULT_MAPPING.items():
if not hasattr(res_obj, field):
continue
......
src/apps/doc/models.py
......@@ -338,6 +338,7 @@ class AFCOCRResult(models.Model):
fsm_activited = models.IntegerField(null=False, default=0, verbose_name="fsm激活状态 1：激活")
fs_ocr = models.TextField(null=True, verbose_name="财务报表")
fss_ocr = models.TextField(null=True, verbose_name="财务情况说明书")
dp_ocr = models.TextField(null=True, verbose_name="首付款支付承诺书")
update_time = models.DateTimeField(auto_now=True, verbose_name='修改时间')
......@@ -383,6 +384,7 @@ class HILOCRResult(models.Model):
fsm_activited = models.IntegerField(null=False, default=0, verbose_name="fsm激活状态 1：激活")
fs_ocr = models.TextField(null=True, verbose_name="财务报表")
fss_ocr = models.TextField(null=True, verbose_name="财务情况说明书")
dp_ocr = models.TextField(null=True, verbose_name="首付款支付承诺书")
update_time = models.DateTimeField(auto_now=True, verbose_name='修改时间')
create_time = models.DateTimeField(auto_now_add=True, verbose_name='创建时间')
......@@ -426,6 +428,7 @@ class AFCSEOCRResult(models.Model):
fsm_activited = models.IntegerField(null=False, default=0, verbose_name="fsm激活状态 1：激活")
fs_ocr = models.TextField(null=True, verbose_name="财务报表")
fss_ocr = models.TextField(null=True, verbose_name="财务情况说明书")
dp_ocr = models.TextField(null=True, verbose_name="首付款支付承诺书")
update_time = models.DateTimeField(auto_now=True, verbose_name='修改时间')
create_time = models.DateTimeField(auto_now_add=True, verbose_name='创建时间')
......@@ -469,6 +472,7 @@ class HILSEOCRResult(models.Model):
fsm_activited = models.IntegerField(null=False, default=0, verbose_name="fsm激活状态 1：激活")
fs_ocr = models.TextField(null=True, verbose_name="财务报表")
fss_ocr = models.TextField(null=True, verbose_name="财务情况说明书")
dp_ocr = models.TextField(null=True, verbose_name="首付款支付承诺书")
update_time = models.DateTimeField(auto_now=True, verbose_name='修改时间')
create_time = models.DateTimeField(auto_now_add=True, verbose_name='创建时间')
......
src/apps/doc/ocr/wb.py
......@@ -874,6 +874,33 @@ class BSWorkbook(Workbook):
ws.append(row)
def down_payment_rebuild(self, down_payment_dict):
"""
Desc:
重构首付款支付承诺书sheet
"""
# 如果 down_payment_dict 为空, 则不创建sheet
if not down_payment_dict:
return
# 如果 down_payment_dict 不为空, 则创建sheet
ws = self.create_sheet(consts.DOWN_PAYMENT_SHEET_NAME)
english_chinese_dict = {
"financial_org_name": "渠道",
"main_borrower_name": "姓名",
"main_borrower_id_no": "证件号码",
"apply_no": "合同编号",
"contract_name": "合同名称",
"promisor_signature": "承诺人签字-电子",
"promisor_signature_date": "承诺人签字日期-电子"
}
for dp_key, dp_value in down_payment_dict.items():
if dp_key in english_chinese_dict.keys():
row = [english_chinese_dict[dp_key], str(dp_value)]
ws.append(row)
else:
row = [english_chinese_dict[dp_key], ""]
ws.append(row)
@staticmethod
def remove_yuan(amount_key_set, key, src_str):
if key in amount_key_set and isinstance(src_str, str):
......@@ -973,7 +1000,7 @@ class BSWorkbook(Workbook):
if len(self.sheetnames) > 1:
self.remove(self.get_sheet_by_name('Sheet'))
def rebuild(self, bs_summary, license_summary, res_list, document_scheme, contract_result, metadata, financial_statement_dict, financial_explanation_dict):
def rebuild(self, bs_summary, license_summary, res_list, document_scheme, contract_result, metadata, financial_statement_dict, financial_explanation_dict, down_payment_dict):
res_count_tuple = self.res_sheet(res_list)
count_list = [(consts.MODEL_FIELD_BS, len(bs_summary))]
......@@ -983,12 +1010,14 @@ class BSWorkbook(Workbook):
self.bs_rebuild(bs_summary, res_count_tuple, metadata)
self.financial_rebuild(financial_statement_dict)
self.financial_explanation_rebuild(financial_explanation_dict)
self.down_payment_rebuild(down_payment_dict)
else:
self.bs_rebuild(bs_summary, res_count_tuple, metadata)
self.license_rebuild(license_summary, document_scheme, count_list)
self.contract_rebuild(contract_result, True)
self.financial_rebuild(financial_statement_dict)
self.financial_explanation_rebuild(financial_explanation_dict)
self.down_payment_rebuild(down_payment_dict)
self.move_res_sheet()
self.remove_base_sheet()
return count_list, self.need_follow
......
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!