Custom fieldname support

[jonafato]

Add support for validating CSVs that don't include a header row. The implementation allows a user to define a sequence of field names and passes these through to the underlying DictReader instances. The fieldnames attribute is None by default, which retains the existing behavior of inferring field names from a header row in the source CSV.

Resolves #88.

---

Example usage:

from vladiate import Vlad
from vladiate.validators import UniqueValidator, SetValidator
from vladiate.inputs import LocalFile

class CustomFieldnamesValidator(Vlad):
    source = LocalFile('vampires.csv')
    fieldnames = ['Column A', 'Column B']
    validators = {
        'Column A': [
            UniqueValidator()
        ],
        'Column B': [
            SetValidator(['Vampire', 'Not A Vampire'])
        ]
    }

Column A,Column B
Vlad the Impaler,Not A Vampire
Dracula,Vampire
Count Chocula,Vampire

» vladiate

Validating CustomFieldnamesValidator(source=LocalFile('vampires.csv'))
Passed! :)

[di]

Makes sense to me!