Introduce formatted string generator

[mfalesni]

Introduce fauxfactory.generate() which takes a string with formatting similar to "".format() but instead of inserting strings, it randomizes the values based on what is located inside the braces.

[mfalesni]

I wonder why does it hate line 807 in python 2.6. I think doing # NOQA or whatever works is OK here.

[Ichimonji10]

Using # noqa isn't the right thing to do here, because flake8 is not complaining. The tool which is complaining is pylint. Old versions of pylint complain about the use of * and **, and it appears that the Python 2.6 environment uses an old version of pylint. (To be expected.) The appropriate comment is # pylint:disable=star-args.

[Ichimonji10]

An example of the formatter is::

    'test_vm_{alpha:12|.lower}'

See literal blocks.

[mfalesni]

@Ichimonji10 the pylint comment as you said triggered an error in all of the python versions. I have put the NOQA comment there to see if that one solves the issue. I also updated the example.

[mfalesni]

Doesn't help either.

[elyezer]

I think this is a good addition and a good start. The next step would be allowing one or more generators to be specified.

We need to take care of not reinventing some behaviours implemented by other libraries like rstr [1]

[1] https://pypi.python.org/pypi/rstr/2.1.3

[elyezer]

Thinking better... I think that we can have the same behaviour without the need of any additional logic:

generate("test_vm_{alpha:12|.lower}")

# is the same as

'test_vm_{0}'.format(gen_alpha(12).lower())

With that said, by adding this kind of generator will not bring any real improvement because we can mix the python string formatting with the generators without the overhead of parsing the generator string.

Also the default string formatting allows generating not just one but many formats in the same string like:

'test_vm_{0}_{1}'.format(
    gen_alpha(5).lower(),
    gen_integer(3)
)

[Ichimonji10]

The format() method already specifies a mini-language. See examples here. The mini-language uses the { and } characters to denote what should be interpolated. Having our own similar-but-incompatible formatting mini-language seems like a bad idea to me.

[omaciel]

@mfalesni After reviewing the PR and after considering the ultimate goal for it, I tend to agree with what @elyezer mentioned above. I feel that we'd be re-inventing the wheel and that perhaps this could be something that you can implement in your own code to handle your specific requirement?

[mfalesni]

ok