Question

I might have pretty basic question about regex. I have the following regex, which when hardcoded int the application work fine, but then I read it with ConfigParser doesn’t seem to work:

r"\[[+-]?\d+(?:\.\d+)?\]"

The way how I read it is :

Config = ConfigParser.ConfigParser()
Config.read("test.conf")
test_regex = Config.get("test","test_regex")
search_pattern = re.compile(test_regex)

test_result = search_pattern.findall(text_to_parse)

The part of the test.conf

[test]
test_regex=r"\[[+-]?\d+(?:\.\d+)?\]"

The input for testing might be something as follows:

text_to_parse = " Here is the [TEST-DONE]" // Success: my regex is extracting [TEST-DONE]
text_to_parse = " Here is the some text" // Failure my regex returns empty list

Any solution for this issue?
Thanks a lot,
Serhiy.

EDIT: was my attention error, as I mention in the comment to answer, but the solution of remotion of the r from regex when it’s in the file helped a lot.

Answer 1

You can use ast.literal_eval to parse the string according to the Python rules:

>>> import ast
>>> ast.literal_eval(conf.get("test", "test_regex"))
'\\[[+-]?\\d+(?:\\.\\d+)?\\]'

But it’s easier to just change your config file to contain the unescaped regex in the first place:

[test]
test_regex=\[[+-]?\d+(?:\.\d+)?\] 

>>> conf.get("test", "test_regex")
'\\[[+-]?\\d+(?:\\.\\d+)?\\]'

That said, the regex doesn’t seem to do what you think it does. It matches:

• opening bracket
• followed by optional + or – sign
• followed by a number of digits
• optionally followed by a dot and more digits
• followed by closing bracket

Example:

>>> re.findall(r'\[[+-]?\d+(?:\.\d+)?\]', 'foo [+10] bar [-3.5]')
['[+10]', '[-3.5]']

Of course there will not be any matches in both of your example strings, because they don’t contain the pattern!