Expand docscrape to support more numpy sections

Please see [numpy format](https://numpydoc.readthedocs.io/en/latest/format.html#see-also)

```
class NumpyDocString(Mapping):
"""Parses a numpydoc string to an abstract representation """
sections = { 'Summary': [''], 'Extended': [], 'Parameters': [], 'Returns': [] }

def __init__(self, docstring, config=None):
docstring = textwrap.dedent(docstring).split('\n')
self._doc = Reader(docstring)
self._parsed_data = copy.deepcopy(self.sections)
self._parse()
self['Parameters'] = {o.name:o for o in self['Parameters']}
if self['Returns']: self['Returns'] = self['Returns'][0]
self['Summary'] = dedent_lines(self['Summary'], split=False)
self['Extended'] = dedent_lines(self['Extended'], split=False)
```

The issue points to the numpydoc format reference and the NumpyDocString entry point; start by comparing the parser's supported sections with that reference. Done means the additional sections are represented in the abstract result while preserving the existing Summary, Extended, Parameters, and Returns handling.

Written by the indexing model from the issue text.