Skip to content

周伟奇 / bmw-ocr

Go to a project
Toggle navigation
Toggle navigation pinning
e3d719b8 by 周伟奇
Options

fix db bug

1 parent 1f46e609
Inline Side-by-side
Showing 1 changed file with 24 additions and 11 deletions
src/apps/doc/management/commands/ocr_process.py
...@@ -55,14 +55,14 @@ class Command(BaseCommand, LoggerMixin): ...@@ -55,14 +55,14 @@ class Command(BaseCommand, LoggerMixin):
55 def signal_handler(self, sig, frame): 55 def signal_handler(self, sig, frame):
56 self.switch = False # 停止处理文件 56 self.switch = False # 停止处理文件
57 57
58 def get_doc_object(self, task_str): 58 # def get_doc_object(self, task_str):
59 business_type, doc_id_str = task_str.split(consts.SPLIT_STR) 59 # business_type, doc_id_str = task_str.split(consts.SPLIT_STR)
60 doc_id = int(doc_id_str) 60 # doc_id = int(doc_id_str)
61 doc_class = HILDoc if business_type == consts.HIL_PREFIX else AFCDoc 61 # doc_class = HILDoc if business_type == consts.HIL_PREFIX else AFCDoc
62 # doc_info = doc_class.objects.filter(id=doc_id, status=DocStatus.INIT.value).values( 62 # # doc_info = doc_class.objects.filter(id=doc_id, status=DocStatus.INIT.value).values(
63 # 'id', 'metadata_version_id', 'application_id', 'document_name', 'document_scheme').first() 63 # # 'id', 'metadata_version_id', 'application_id', 'document_name', 'document_scheme').first()
64 doc = doc_class.objects.filter(id=doc_id).first() 64 # doc = doc_class.objects.filter(id=doc_id).first()
65 return doc, business_type 65 # return doc, business_type
66 66
67 def get_doc_info(self): 67 def get_doc_info(self):
68 task_str, is_priority = rh.dequeue() 68 task_str, is_priority = rh.dequeue()
...@@ -73,7 +73,12 @@ class Command(BaseCommand, LoggerMixin): ...@@ -73,7 +73,12 @@ class Command(BaseCommand, LoggerMixin):
73 self.cronjob_log.info('{0} [get_doc_info] [task={1}] [is_priority={2}]'.format( 73 self.cronjob_log.info('{0} [get_doc_info] [task={1}] [is_priority={2}]'.format(
74 self.log_base, task_str, is_priority)) 74 self.log_base, task_str, is_priority))
75 try: 75 try:
76 doc, business_type = self.get_doc_object(task_str) 76 # doc, business_type = self.get_doc_object(task_str)
77 business_type, doc_id_str = task_str.split(consts.SPLIT_STR)
78 doc_id = int(doc_id_str)
79 doc_class = HILDoc if business_type == consts.HIL_PREFIX else AFCDoc
80 doc = doc_class.objects.filter(id=doc_id).first()
81
77 if doc is None: 82 if doc is None:
78 self.cronjob_log.warn('{0} [get_doc_info] [doc not exist] [task_str={1}] [is_priority={2}]'.format( 83 self.cronjob_log.warn('{0} [get_doc_info] [doc not exist] [task_str={1}] [is_priority={2}]'.format(
79 self.log_base, task_str, is_priority)) 84 self.log_base, task_str, is_priority))
...@@ -627,7 +632,11 @@ class Command(BaseCommand, LoggerMixin): ...@@ -627,7 +632,11 @@ class Command(BaseCommand, LoggerMixin):
627 del unknown_summary 632 del unknown_summary
628 633
629 # 4.2 重构Excel文件 634 # 4.2 重构Excel文件
630 doc, business_type = self.get_doc_object(task_str) 635 # doc, business_type = self.get_doc_object(task_str)
636 business_type, doc_id_str = task_str.split(consts.SPLIT_STR)
637 doc_id = int(doc_id_str)
638 doc_class = HILDoc if business_type == consts.HIL_PREFIX else AFCDoc
639 doc = doc_class.objects.filter(id=doc_id).first()
631 doc_data_path = os.path.join(self.data_dir, business_type, consts.TMP_DIR_NAME, str(doc.id)) 640 doc_data_path = os.path.join(self.data_dir, business_type, consts.TMP_DIR_NAME, str(doc.id))
632 excel_path = os.path.join(doc_data_path, '{0}.xlsx'.format(doc.id)) 641 excel_path = os.path.join(doc_data_path, '{0}.xlsx'.format(doc.id))
633 # src_excel_path = os.path.join(doc_data_path, 'src.xlsx') 642 # src_excel_path = os.path.join(doc_data_path, 'src.xlsx')
...@@ -639,7 +648,11 @@ class Command(BaseCommand, LoggerMixin): ...@@ -639,7 +648,11 @@ class Command(BaseCommand, LoggerMixin):
639 with lock: 648 with lock:
640 if task_str in res_dict: 649 if task_str in res_dict:
641 del res_dict[task_str] 650 del res_dict[task_str]
642 doc, business_type = self.get_doc_object(task_str) 651 # doc, business_type = self.get_doc_object(task_str)
652 business_type, doc_id_str = task_str.split(consts.SPLIT_STR)
653 doc_id = int(doc_id_str)
654 doc_class = HILDoc if business_type == consts.HIL_PREFIX else AFCDoc
655 doc = doc_class.objects.filter(id=doc_id).first()
643 doc.status = DocStatus.PROCESS_FAILED.value 656 doc.status = DocStatus.PROCESS_FAILED.value
644 doc.save() 657 doc.save()
645 self.cronjob_log.warn('{0} [process failed (res to wb)] [task={1}] [error={2}]'.format( 658 self.cronjob_log.warn('{0} [process failed (res to wb)] [task={1}] [error={2}]'.format(
......
Styling with Markdown is supported
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!