Add generic base converter

[groteworld]

This commit adds a base converter that contains all the common functions used by all converters. New converters simple can inherit from the BaseConverter class and add the following:

1. variable - name: name of the format (e.g "csv")
2. variable - extension: extension of the format (e.g. ".csv")
3. function - write_output: does the format specific writing

Each child class has access to the variables and functions in base converter such as, but not limited to:

• variable - args: the arguments or defaults from the command
• varibale - headers: the columns to be returned
• function - get_issues: returns list of issues

This could be made even more concise by having the base return back a list of issues based off of the args, and not requirings the use of the args variable or helper functions like is_pull_request, but I think that doesn't allow for opinionated configurations.

[groteworld]

Totally agree and updated.

[sigmavirus24]

I'm going to test this and accept it. I might take this a slightly different direction (see also https://youtu.be/OMPfEXIlTVE).

[groteworld]

My eyes are open! So you're thinking something closer to....

csv.py

def csv_writer():
    # csv writing code

def main():
    Converter('csv', '.csv', csv_writer)

converter.py

class Converter():
    def __init__(self, name='BaseConverter', extension='.txt', writer=self.default_writer, **kwargs):
        self.name = name
        self.extension = extension
        self.writer = writer

        # ....

       self.writer()

    def default_writer(self):
       # default writer stuff

wrote this in this comment so unsure on "truthiness"

[sigmavirus24]

So I think we have two things we care about (right now)

1. Formatting
2. Writing

I don't think we should conflate the two. The Writer class (whatever we call it) can handle default file names, file extensions (by proxy), and handling the file-pointer. The Formatter class can receive the issues and format them appropriately.

I think both classes will need to be able to register arguments on the ArgumentParser instance, so we'll have something like:

class Converter(object):
    def __init__(self, formatter=None, writer=None):
        if formatter is None:
            raise ValueError('..')
        if writer is None:
            raise ValueError('..')

        self.formatter = formatter
        self.writer = writer
        self.parser = self.make_parser()
        # These next two lines can be in their own method too
        self.formatter.register_arguments(self.parser)
        self.writer.register_arguments(self.parser)
        self.args = self.parser.parse_args()
        self.make_client()  # Setups up our instance of github3.GitHub

If we want, make_parser can read information from self.writer by either using it implicitly or by explicitly passing it in. Make sense?

[sigmavirus24]

Also, don't feel as if you have to do this. I'm getting paid to work on this for now =P

[groteworld]

NP, just helping out the discussion.

[sigmavirus24]

And I appreciate it :)