python复合正则表达式提取不同文档中不同标签之间的文本

我有两种类型的文档，其中包含我要复制和保存的文本。一种文档类型具有由名为 TAGSTART 和 TAGEND 的标记分隔的有趣文本。另一种文档类型具有由 CORESTART 和 COREEND 分隔的有趣文本。以下是两个示例:

intro intro intro intro intro intro 
BEGIN A This is where some text starts 
That is not interesting or wanted 
CORESTART save text A save text A save text A save text A  
save text A save text A save text A save text A save text A  
save text A COREEND 
This is an addendum that is not needed but is just in the way 
END A outro outro outro outro outro outro  
outro outro outro outro outro outro outro  

.

intro intro intro intro intro intro 
INIT B This is where some text starts 
That is not interesting or wanted 
TAGSTART B save text B save text B save text B save text B  
save text B save text B save text B save text B save text B  
save text B TAGEND B 
This is an addendum that is not needed but is just in the way 
TERM B outro outro outro outro outro outro  
outro outro outro outro outro outro outro  

这个 python 脚本适用于第一种类型的文件

import os 
import re 
import codecs 
# walk the directory tree 
rootDir = '.' 
for dirName, subdirs, files in os.walk(rootDir): 
    #    exclude hidden files and directories 
    files = [f for f in files if not f[0] == '.'] 
    subdirs[:] = [d for d in subdirs if not d[0] == '.'] 
    for fname in files: 
         if fname.endswith(('.txt', '.TXT')): 
            #    create the full path 
            filename = os.path.join(dirName, fname) 
            with codecs.open(filename, encoding='utf-8', errors='ignore') as infile, codecs.open('SAVED.txt', 'a',encoding='utf-8') as outfile:  
                stuff = infile.read() 
                saveTEXT = '\n' + ''.join(re.findall(r"CORESTART(.+?)COREEND", stuff, re.DOTALL|re.MULTILINE)) + '\n' 
                outfile.write(saveTEXT) 
                infile.close() 
                outfile.close() 

如果我将正则表达式更改为

      saveTEXT = '\n' + ''.join(re.findall(r"TAGSTART B(.+?)TAGEND B", stuff, re.DOTALL|re.MULTILINE)) + '\n' 

我可以从第二种文件中得到我想要的东西。然而，复合正则表达式失败了:

      saveTEXT = '\n' + ''.join(re.findall(r"CORESTART|TAGSTART B(.+?)COREEND|TAGEND B", stuff, re.DOTALL|re.MULTILINE)) + '\n' 

什么都找不到。我尝试将原始正则表达式括在括号中，但随后出现错误，正则表达式需要一个字符串但找到了一个元组。我尝试用\b 关闭正则表达式中的单词以指示单词边界，如下所示

       saveTEXT = '\n' + ''.join(re.findall(r"\bCORESTART B\b|\bTAGSTART B\b(.+?)\bCOREEND B\b|\bTAGEND B\b", stuff, re.DOTALL|re.MULTILINE)) + '\n' 

但这也是空的。当我为原始字符串尝试这个时，我的想法完全被震撼了:

[\bCORESTART\b|\bTAGSTART B\b](.+?)[\bCOREEND\b|\bTAGEND B\b] 

我可以就我忽略的内容提供一些指导吗？我的脑子坏了。

请您参考如下方法:

@bobblebubble's regex in the deleted answer如果你允许一些小的偏差(比如“CORESTART”后面可能跟着一个你不想从比赛中得到的空格+“B”)，这是正确的方法。也就是说，我建议将 (?: B)? 添加到 (TAG|CORE)START B(.+?)(\1END B) 正则表达式:

(TAG|CORE)START(?: B)?(.+?)(\1END(?: B)?) 

参见 regex demo

您还必须使用 re.finditer 提取字符串，因为 re.findall 将提取所有捕获组值。

IDEONE demo :

import re 
p = re.compile(r'(TAG|CORE)START(?: B)?(.+?)(\1END(?: B)?)', re.DOTALL) 
test_str = "intro intro intro intro intro intro\nBEGIN A This is where some text starts\nThat is not interesting or wanted\nCORESTART save text A save text A save text A save text A \nsave text A save text A save text A save text A save text A \nsave text A COREEND\nThis is an addendum that is not needed but is just in the way\nEND A outro outro outro outro outro outro \noutro outro outro outro outro outro outro \n.\n\nintro intro intro intro intro intro\nINIT B This is where some text starts\nThat is not interesting or wanted\nTAGSTART B save text B save text B save text B save text B \nsave text B save text B save text B save text B save text B \nsave text B TAGEND B\nThis is an addendum that is not needed but is just in the way\nTERM B outro outro outro outro outro outro \noutro outro outro outro outro outro outro " 
print([x.group(2) for x in p.finditer(test_str)]) 

请注意 re.MULTILINE 在您的正则表达式中是多余的，因为此标志重新定义了开始匹配的 ^ 和 $ anchor 的行为行的开始和结束而不是整个字符串(分别)。因此，我将其排除在正则表达式声明之外。