Naming convention for Python regular expressions?

Question

Is there an accepted naming convention for regular expressions in Python? Or if there's not, what are some suggestions on how to name them?

Usually I name them something like look_for_date or address_re but I've read several places that using a suffix like '_re' in a variable name isn't good. To me it seems like the regex needs something to indicate it's a regex, since if you named it just date or address, you wouldn't be able to do stuff like this, which seems intuitive:

date = date_re.match(text)

Answer 1

Compiled regular expressions are generally constants, so should have an UPPER_CASE_WITH_UNDERSCORES name per PEP 8. I tend to name them for what they would match; to give an example from some code I wrote recently:

import re

VALID_CLOSURE_PATTERN = re.compile(r'''
    ^\d{2}    # starts with two digits 0-9
    [NY]{4}$  # followed by four Y/N characters
''', re.IGNORECASE + re.VERBOSE)


class RoadClosure(object):

    def __init__(self, ..., closure_pattern):
        """Initialise the new instance."""
        if not VALID_CLOSURE_PATTERN.match(closure_pattern):
            raise ValueError('invalid closure pattern: {!r}'.format(closure_pattern))
       ...

...

I think this makes it pretty clear what's going on, VALID_CLOSURE_PATTERN communicates "this describes what we would consider to be a valid closure pattern" and a line like:

if not VALID_CLOSURE_PATTERN.match(closure_pattern):

describes what it's actually doing in close to plain English. So in your case, you might write:

date = VALID_DATE.match(text)