Extracting elements inside and between brackets

I have the following string and I would like to extract the elements (xx="yy") and what's between the brackets. Here's an example:

[caption id="get this" align="and this" width="and this" caption="and this"]this too please[/caption]

I've tried the following code but I'm quite a noob with regex.

re.sub(r'\[caption id="(.*)" align="(.*)" width="(.*)" caption="(.*)"\](.*)\[\/caption\]', "tokens: %1 %2 %3 %4 %5", self.content开发者_运维知识库, re.IGNORECASE) 

Thanks a lot in advance!

---

It's probably not working for you because .* is greedy. Try [^"]* in its place. [^"] means the set of all characters except the quote character. Also, as you've pointed out in the comments, the token syntax, is \\n, not %n. Try this:

re.sub(r'\[caption id="([^"]*)" align="([^"]*)" width="([^"]*)" caption="([^"]*)"\](.*)\[\/caption\]', "tokens: \\1 \\2 \\3 \\4 \\5", self.content, re.IGNORECASE)

Do the contents of the caption tag span multiple lines? If they do .* won't capture the newlines. You'll need to us something like [^\x00]* instead. [^\x00] means the set of all charchters except the null character.

re.sub(r'\[caption id="([^"]*)" align="([^"]*)" width="([^"]*)" caption="([^"]*)"\]([^\x00]*)\[\/caption\]', "tokens: \\1 \\2 \\3 \\4 \\5", self.content, re.IGNORECASE)

On the off chance that your strings can actually legitimately contain null characters, you would need to use the re.DOTALL flag instead.

---

You might be able to leverage the power of Python's standard SGML/HTML/XML parsing modules: if it is safe to substitute "[]" for "<>", then you can do this substitution in order to produce valid XML, and do the parsing with the standard library XML parsing functions:

import string
from xml.etree import ElementTree as ET

text = '[caption id="get this" align="and this" width="and this" caption="and this"]this too please[/caption]'
xml_text = string.translate(text, string.maketrans('[]', '<>'))  # Conversion to XML
parsed_text = ET.fromstring(xml_text)  # Parsing

# Extracted information
print "Text part:", parsed_text.text
print "Values:", parsed_text.attrib.values()

This correctly print:

Text part: this too please
Values: ['and this', 'and this', 'get this', 'and this']

The advantage of this approach is that (1) it uses a standard module that many people know; (2) it explicitly shows what you want to do; and (3) you can readily extract more information, handle more complicated values (including values that contain double quotes…), etc.

---

You could try something like this?

re = '[caption id="get this" align="and this" width="and this" caption="and this"]this too please[/caption]'
re.gsub(/([a-z]*)=\"(.*?)\"/i) do |m|
    puts "#{$1} = #{$2}
end