Echo/fcb_photo_review
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

修正页码排序

Browse Source
This commit is contained in:
liuyebo
2024-10-18 10:49:01 +08:00
parent d85b3fff8f
commit 3f93bd476a
2 changed files with 12 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
photo_review/auto_photo_review.py
View File

@@ -421,7 +421,9 @@ def cost_task(pk_phhd, cost_list_ie_result):
cost_data["medical_expenses"] = handle_decimal(cost_data["medical_expenses_str"]) cost_data["medical_expenses"] = handle_decimal(cost_data["medical_expenses_str"])
if cost_list_ie_result.get(IE_KEY['page']): if cost_list_ie_result.get(IE_KEY['page']):
page_nums, page_count = parse_page_num(cost_list_ie_result[IE_KEY['page']]) page_nums, page_count = parse_page_num(cost_list_ie_result[IE_KEY['page']])
cost_data['page_nums'] = handle_original_data(','.join(page_nums)) if page_nums:
page_nums_str = [str(num) for num in page_nums]
cost_data['page_nums'] = handle_original_data(','.join(page_nums_str))
cost_data['page_count'] = handle_tiny_int(page_count) cost_data['page_count'] = handle_tiny_int(page_count)
save_or_update_ie(ZxIeCost, pk_phhd, cost_data) save_or_update_ie(ZxIeCost, pk_phhd, cost_data)
return cost_data return cost_data
@@ -583,11 +585,12 @@ def photo_review(pk_phhd, name):
cost_missing_page = {} cost_missing_page = {}
if cost_data.get('page_nums'): if cost_data.get('page_nums'):
page_nums = cost_data['page_nums'].split(',') page_nums = cost_data['page_nums'].split(',')
required_set = set(str(i) for i in range(1, cost_data['page_count'] + 1)) required_set = set(range(1, cost_data['page_count'] + 1))
page_set = set(page_nums) page_set = set([int(num) for num in page_nums])
cost_missing_page = required_set - page_set cost_missing_page = required_set - page_set
if cost_missing_page: if cost_missing_page:
cost_missing_page = sorted(cost_missing_page) cost_missing_page = sorted(cost_missing_page)
cost_missing_page = [str(num) for num in cost_missing_page]
page_description.append(f"《住院费用清单》，缺第{','.join(cost_missing_page)}页") page_description.append(f"《住院费用清单》，缺第{','.join(cost_missing_page)}页")
if page_description: if page_description:

8
util/data_util.py
View File

@@ -194,14 +194,16 @@ def parse_page_num(page_list):
for page in page_list: for page in page_list:
page_texts = [p.get('text', '') for p in page] page_texts = [p.get('text', '') for p in page]
join = ''.join(page_texts) join = ''.join(page_texts)
numbers = re.findall(r'\d+', join) numbers_str = re.findall(r'\d+', join)
# 过滤异常值 # 过滤异常值
numbers = [num for num in numbers if int(num) <= 30] numbers = [int(num) for num in numbers_str if int(num) <= 30]
if not numbers: if not numbers:
continue continue
pages.append(min(numbers)) pages.append(min(numbers))
total.append(max(numbers)) total.append(max(numbers))
return pages, int(max(total if total else ['1'])) if not pages:
return None, None
return pages, max(total)
def handle_tiny_int(num): def handle_tiny_int(num):