Closes #423

This PR introduces the pybel.io.sbgnml I/O module, which first consumes SBGN-ML and converts it into a facile JSON-like data structure (which is generally reusable) then reasons over it and converts to BEL.

• SBGN parser
• BEL converter

The code also has some grounding mechanisms built in to deal with the often missing identifiers from the COVID-Pathways content, but I might remove this to make the tool more generally reusable. I'm not really sure who else generates SBGN-ML if not from CellDesigner (and more specifically, coming out of MINERVA)

DEMO NOTEBOOK

Example SBGN-ML Content

For an example, see the content posted by @cannin https://cannin.github.io/covid19-sbgn. The SBGN-ML file (disregard the SIF and SIF SVG, since those are derived resources).

He notes on his site that these SBGN-ML files were converted from CellDesigner using https://github.com/sbgn/cd2sbgnml. It's clear that there's some lost information due to cd2sbgnml (or there are strange pockets of incredibly low quality curation), so it will be a future step to write a CellDesigner importer later (see #440).

For example, this script can be run (after installing this branch in development mode) with the following code in the python REPL:

import json
from urllib.request import urlretrieve
from pybel.io.sbgnml.sbgnml import parse

url = 'https://cannin.github.io/covid19-sbgn/ER_Stress_Cov19.xml.sbgn'
path = 'ER_Stress_Cov19.xml'
urlretrieve(url, path)

rv = parse(path)
with open('ER_Stress_Cov19.json', 'w') as file:
    json.dump(rv, file, indent=2)