Following http://www.artima.com/weblogs/viewpost.jsp?thread=240845, I've written the decorator extend_to_separated_input:
class extend_to_separated_input(object):
def __init__(self, separator):
assert isinstance(separator, str), "The separator must be a string."
self.separator = separator
def __call__(self, parse_function):
def wrapped_function(raw_conditions, *args, **kwargs):
raw_conditions = raw_conditions.split(self.separator)
return [parse_function(raw_condition, *args, **kwargs) for raw_condition in raw_conditions]
return wrapped_function
This decorator works for globally defined functions, for example,
import parse
fmt = "{word1} {word2}"
@extend_to_separated_input(separator="&&")
def _parse(string):
return parse.parse(fmt, string).named
string = "foo bar && hello world"
print _parse(string)
prints the result
[{'word1': 'foo', 'word2': 'bar '}, {'word1': ' hello', 'word2': 'world'}]
which is what I would expect. (Here I've used the parse module). In my actual use case, however, the _parse method is a class method, not a global method. However, if I try to do this:
class TwoWords(object):
fmt = "{word1} {word2}"
@classmethod
@extend_to_separated_input(separator="&&")
def _parse(cls, raw_condition):
return parse.parse(cls.fmt, raw_condition).named
string = "foo bar && hello world"
print TwoWords._parse(string)
I get an
AttributeError: type object 'TwoWords' has no attribute 'split'
The problem is that it is picking up the class as the first argument. I've tried to adapt the definition to def wrapped_function(cls, raw_conditions, *args, **kwargs), but this leads to a TypeError because of insufficient arguments.
How I adapt the decorator so that it works for class methods?
