Configure regex validation to be case insensitive

[luizpericolo]

I know it is not possible to express case insensitivity flag in the actual regex string in python, so I guess that cerberus could use another attribute to manage this, for instance: regex_case.

[funkyfuture]

you could fiddle around with the anyof-rule (or similar) and lower/upper-methods of strings. could you?

[luizpericolo]

This does not sound optimal since I'd have to have two versions of my regex, the lower and the upper case version, if I were to use anyof as you suggested. This is bad because If anything on the regex changes I have to change it on the other version aswell.

[luizpericolo]

I'm debugging the regex validation process and I realize that my suggestion might not be feasible.
Maybe you could add a regex_case attribute to the Validator constructor, but I don't know if that does really make sense, specially if you would like some regex to be case sensitive and others not.

I'd be willing to contribute to this, but I just can't think of a good solution for this problem.

[luizpericolo]

Maybe change the regex key mapping from str to dict would be a good solution.
Something along the lines of:

data_format = {
    'name': {
        'type': 'string',
        'regex': {
            'pattern': r'^[a-z0-9]$', 
            'case': 'insensitive'
        }
    }
}

The problem then would be how to maintain backwards compatibility to the current syntax of regex validation.

[funkyfuture]

maintaining backward compatibility wouldn't be a problem. (duck-typing per instance-tests, you'll find a lot of these in the code.)
however, i suggest to name the second parameter flags and let that be an integer like the parameter with the same name for re.compile.