Add functions to compute Schultz and Gutman Index

[jangwon-yie]

It contains implementations of first, second kinds of Schultz index and their unit cases. Please give me any comments so that it will be merged to the main branch. Thanks.

[pep8speaks]

Hello @jangwon-yie! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

• In the file networkx/algorithms/__init__.py:

Line 53:1: F403 'from networkx.algorithms.schultz import ' used; unable to detect undefined names
Line 53:1: F401 'networkx.algorithms.schultz.' imported but unused

Comment last updated at 2019-11-07 11:57:18 UTC

[dschult]

The Schultz index, and Gutman index (second type of schultz index) are related to the Wiener Index and Kirchoff Index (https://hrcak.srce.hr/132323). All four (and others) are used as markers of chemicals based on the molecular network of atoms connected by chemical bonds.

The functions provided here can be improved and updated. The file structure, doc_strings and tests seem pretty good. I think they should be moved into the module wiener.py and probably connected to the kirchoff index if we ever get that. Maybe they should all go into a distance module as suggested in #6975 but for now we can put these into wiener.py.

[dschult]

I'll go ahead and approve this -- mostly to let you know that it is ready to be reviewed.
Do we want these functions? Is this the right place?

[ghost]

The functions seem useful and appropriate for NX.
I would advocate to move all functions to Distance measures.

[dschult]

These measures involve both distances between nodes and the degree of the nodes. Can you say why you tink they should all be contained within the ditance_measures.py module?

[rossbar]

IMO adding these actually motivates keeping a separate module for the related indices. From the user perspective it should be a moot point as users are encouraged to call indices from the top level namespace rather than go spelunking through the package hierarchy. In other words, it's largely an implementation detail that shouldn't hold anything up in this PR!

[rossbar]

I took the liberty of pushing up a few minor docstring tweaks, but otherwise LGTM. Thanks @jangwon-yie !

I'm certainly no expert, but I'm in favor of adding these indices - they're straightforward computations and seem to be generic measures of graph properties!