如何用Python正则表达式拆分发言文本中的发言人和内容？

正则表达式模式 #

使用以下正则（需开启多行模式（MULTILINE）和点匹配换行模式（DOTALL））：

^([^:]+?)\s*:\s*(.*?)(?=\n\n[^:]+?:|\Z)

正则各部分说明 #

• ^：匹配每行的起始位置（多行模式下生效）
• ([^:]+?)：捕获发言人信息，匹配到第一个冒号前的所有内容（非贪婪模式避免过度匹配）
• \s*:\s*：匹配冒号及前后任意数量的空白字符（适配原文本中冒号前后的空格差异）
• (.*?)：捕获发言内容，非贪婪模式匹配到下一个分隔符前的所有内容
• (?=\n\n[^:]+?:|\Z)：正向预查，触发匹配结束的条件是「两个换行+新的发言人标识」或「文本末尾」

实际使用示例（Python） #

import re

text = """JOHN SMITH, GLOBAL HEAD OF YOUTUBE : Good morning, good 
afternoon, everyone . Before I hand over to facebook, I want to give a quick reminder of the reporting 
changes that have taken effect this filming of a tv show.  
 
 
BOBBY DUDE, GROUP FROM FACEBOOK:     Thanks, john smith lets talk about movies and films we watch when we are bored parents. 
"""

pattern = re.compile(r'^([^:]+?)\s*:\s*(.*?)(?=\n\n[^:]+?:|\Z)', re.M | re.S)
matches = pattern.findall(text)

# 输出拆分结果并翻译
for idx, (speaker, content) in enumerate(matches, 1):
    # 翻译发言人信息
    speaker_cn = speaker.replace("JOHN SMITH, GLOBAL HEAD OF YOUTUBE", "约翰·史密斯，YouTube全球负责人")\
                        .replace("BOBBY DUDE, GROUP FROM FACEBOOK", "鲍比·杜德，Facebook团队成员")
    # 翻译发言内容
    content_cn = content.replace("Good morning, good afternoon, everyone . Before I hand over to facebook, I want to give a quick reminder of the reporting changes that have taken effect this filming of a tv show.",
# 正则表达式拆分发言人和发言内容方案

### 适用的正则模式
针对你提供的文本格式，可使用以下正则表达式（启用多行模式）：
```regex
(?m)^([A-Z][A-Z\s,]+?)\s*:\s*(.*?)(?=\n[A-Z][A-Z\s,]+:|\Z)

模式逐段解释 #

• (?m)：启用多行模式，让^匹配每一行的开头，确保只从行首识别发言人
• ^([A-Z][A-Z\s,]+?)：匹配发言人信息，从行首大写字母开始，匹配大写字母、空格、逗号，用非贪婪模式避免过度匹配到无关内容
• \s*:：匹配冒号前的任意数量空格（包括0个，兼容不同格式的冒号写法）
• \s*：匹配冒号后的任意数量空格，清理发言内容前的冗余空格
• (.*?)：非贪婪匹配发言内容，避免把后续发言人信息包含进来
• (?=\n[A-Z][A-Z\s,]+:|\Z)：正向预查，确保内容截止到下一个行首的发言人标记（大写字母开头+冒号）或文本结束位置

代码示例（Python） #

import re

text = """JOHN SMITH, GLOBAL HEAD OF YOUTUBE : Good morning, good 
afternoon, everyone . Before I hand over to facebook, I want to give a quick reminder of the reporting 
changes that have taken effect this filming of a tv show.  
 
 
BOBBY DUDE, GROUP FROM FACEBOOK:     Thanks, john smith lets talk about movies and films we watch when we are bored parents. 
"""

pattern = r'(?m)^([A-Z][A-Z\s,]+?)\s*:\s*(.*?)(?=\n[A-Z][A-Z\s,]+:|\Z)'
matches = re.findall(pattern, text, re.DOTALL)

for idx, (speaker, content) in enumerate(matches, 1):
    # 清理内容中的多余空行和首尾空格
    cleaned_content = re.sub(r'\n\s*\n', '\n', content).strip()
    print(f'string{idx}: (speaker = {speaker.strip()}, text = {cleaned_content} )')

输出结果 #

string1: (speaker = JOHN SMITH, GLOBAL HEAD OF YOUTUBE, text = Good morning, good afternoon, everyone . Before I hand over to facebook, I want to give a quick reminder of the reporting changes that have taken effect this filming of a tv show. )
string2: (speaker = BOBBY DUDE, GROUP FROM FACEBOOK, text = Thanks, john smith lets talk about movies and films we watch when we are bored parents. )

内容的提问来源于stack exchange，提问作者DreadPirateRoberts