Add python library for loading/parsing output

[powerjg]

I've investigated a number of code generation libraries, and none of them seem to be able to deal with our schema. However, I think that writing a simple python library that we manually keep in sync with the schema will not be too much of an overhead.

In fact, I can imagine implementing new features in the python library based on non-standard specializations and then backporting these specializations into the schema.

[powerjg]

An initial library has been added in python/. There's still work to do, though:

• Add comments

• Add a stats-schema/python/tests directory with unittests

• Add a note in the readme on how to run mypy

• Add unittests that load stats-schema/tests/*.json and make sure they work/fail as they should

• Add documentation on the different init functions and the load functions

• Add distribution type

• Add str and/or repr functions for everything

• Add dump functions

• Clean up style (e.g., line length)?

[powerjg]

I tested the following, and it works:

from simstats import *
import json
with open('../tests/examples/v-full.json') as f:
  d = json.load(f)
stats = SimStat.load(d)
print(stats.my_system.cpus[0].tlb.data_hits.value)