Merge remote-tracking branch 'origin/feature_add_insurance_sec_page' into feature/uat-tmp

@@ -2056,6 +2056,43 @@ class Command(BaseCommand, LoggerMixin):
                             # src_excel_path = os.path.join(doc_data_path, 'src.xlsx')
                             # src_excel_path = os.path.join(doc_data_path, 'src.xlsx')
                             # wb.save(src_excel_path)
                             # 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
                             #need_follow表示在上传edms时文件名是否要添加"关注"两字
+                            
+                            # 处理 保单 的后续页的章识别不到的问题
+                            if len(license_summary) > 0:
+                                # 如果是保单
+                                if consts.INSURANCE_CLASSIFY in license_summary.keys():
+                                    # 获取 license_list
+                                    license_list_tmp = license_summary.get(consts.INSURANCE_CLASSIFY, [])
+                                    if len(license_list_tmp) > 0:
+                                        # 章的有无
+                                        stamp_flag_list = [0] * len(license_list_tmp)
+                                        for license_list_tmp_idx, license_dict_tmp in enumerate(license_list_tmp):
+                                            if isinstance(license_dict_tmp, dict):
+                                                if "保单章" in license_dict_tmp.keys():
+                                                    if license_dict_tmp.get("保单章") is not None:
+                                                        if isinstance(license_dict_tmp.get("保单章"), str) and license_dict_tmp.get("保单章") == "有":
+                                                            stamp_flag_list[license_list_tmp_idx] = 1
+                                        # 计算 stamp_flag_list 中的 sum
+                                        stamp_flag_sum = sum(stamp_flag_list)
+                                        
+                                        # 检查 license_list_tmp 中的每个元素中 [被保险人姓名] 的值是否为空, 若为空, 则不是第一页
+                                        license_first_page_list = [0] * len(license_list_tmp)
+                                        for license_list_tmp_idx, license_dict_tmp in enumerate(license_list_tmp):
+                                            if isinstance(license_dict_tmp, dict):
+                                                if "被保险人姓名" in license_dict_tmp.keys():
+                                                    if license_dict_tmp.get("被保险人姓名") is not None:
+                                                        if isinstance(license_dict_tmp.get("被保险人姓名"), str) and license_dict_tmp.get("被保险人姓名") != "" and "保险" not in license_dict_tmp.get("被保险人姓名"):
+                                                            license_first_page_list[license_list_tmp_idx] = 1
+                                                            break
+                                        # 获取 license_first_page_list 中为 1 的索引
+                                        license_first_page_idx = [i for i, x in enumerate(license_first_page_list) if x == 1]
+                                        # 将 license_summary 中key为 INSURANCE_CLASSIFY 的 value 替换为 first_page 中的内容, 且只保留 1 页
+                                        if len(license_first_page_idx) > 0:
+                                            license_summary[consts.INSURANCE_CLASSIFY] = [license_list_tmp[license_first_page_idx[0]]]
+                                            # 如果 sum > 0, 说明有章, 无论是在第几页, 接下来将第一页的内容提取出来，只保留第一页
+                                            if stamp_flag_sum > 0:
+                                                license_summary[consts.INSURANCE_CLASSIFY][0]["保单章"] = "有"
+                            
                             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)
                             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)
                             wb.save(excel_path)
 
 

@@ -209,6 +209,7 @@ class Retriever:
             if pno_str == '-1':
             if pno_str == '-1':
                 is_last_pno = True
                 is_last_pno = True
                 pno_int_list = [int(pno_str) for pno_str in pdf_text_list.keys()]
                 pno_int_list = [int(pno_str) for pno_str in pdf_text_list.keys()]
+                if len(pno_int_list) > 0:
                     pno_str = str(max(pno_int_list)) 
                     pno_str = str(max(pno_int_list)) 
 
 
             # 搜索关键词
             # 搜索关键词