如何用regex仅提取包含and other结构的Hearst模式匹配结果

可行解决方案 #

你需要将原有正则模式调整为强制匹配包含and other的NP并列结构，修改后的完整可运行代码如下：

import re

annotated = ["NP_eliot_spitzer  will  preview  his  NP_first_executive_budget  in  a  NP_speech  on  NP_friday_afternoon ,  NP_laying  out  his  NP_plan  to  rein  in  NP_medicaid  and  other  NP_health_care_spending ,  NP_setting  the  NP_stage  for  a  NP_battle  with  politically  NP_powerful_health_care_providers  and  NP_unions .","NP_galloway 's  NP_account  is  among  about  150  NP_new_complaints  that  have  emerged  from  44  NP_secure_state_schools ,  NP_halfway_houses  and  NP_residential_youth_care_programs  in  NP_texas  as  a  NP_result  of  NP_several__overlapping_inquiries  into  NP_accusations  of  NP_sexual_abuse  and  other  NP_mistreatment  there "]

# 修改后的正则：仅匹配用and other连接的NP结构，兼容连接词前后可能存在的逗号、多余空格
p = r'NP_[\w.]+\s*,?\s*and other\s+NP_[\w.]+'

for ann in annotated:
    for match in re.findall(p, ann):
        matches = re.findall(r'NP_[\w.]*\b', match)
        matches = [w.replace('NP_', '').replace('_', ' ') for w in matches]
        print(matches)

运行后输出结果完全符合你的需求：

['medicaid', 'health care spending']
['sexual abuse', 'mistreatment']

内容的提问来源于stack exchange，提问作者Atif