Making stubs great again

[fredbi]

@casualjim

Please don't ditch this repo: it is awesome, albeit unfinished work.
I'll try to revive it. I managed to make it up and running and I am starting to figure out the many shortcomings. That's quite a heavy task.

Here is a todo that came up after a couple hours tinkering with the package.

What remains to be done. Feel free to comment.

• workable faker for numbers and integers, according to format (format given as Arg to ValueGenerator)
• faker for date, date-time and duration formats (duration is simple starting from a random int and using random alternate text marshaller,
  date is not so complex: take time.Now() and subtract a random number of days then marshal as text
• relocate the reseed at top level (as top level option)
• taking into account validations (planned dev, but no actual implementation)
• abide by generation flags to control when generating a passing case and a non passing case (currently not implemented)
• faker for slices (with option to produce a parsable string using CollectionFormat)
• a testable fixture generator to instruct a test client (or a unit testing program)
• a spec walker (may reuse the one under development with validate) to generate test cases for parameters, headers and responses
• a test case orchestrator to plan for a full coverage of validation points
• rule-based option detection for simple types: the idea is to find a smart way to select the ValueGenerator according to rules
  • simple types and formats should provide straightforward rules
• a fuzzy-er based on title and description tags, along the lines of what meqqa suggests
• the idea is to detect proper options according to hints in title/doc
• investigate go native solution for NLP (meqqa use Python)
• limit the use of x-datagen extension tag
• generate combinatorial test suite with required / non required / allowsEmpty items
• find a way to customize the faker dictionary

[shahzaibaziz]

@fredbi I am in for contribution