Network TMLE about zepid HOT 3 CLOSED

For all the interference functionality #15 #14 #29 will need to have NetworkX somewhere in the background. Since not everyone uses methods to handle interference, I would prefer to have NetworkX as an extra. For adding extras, see https://setuptools.readthedocs.io/en/latest/setuptools.html#declaring-extras-optional-features-with-their-own-dependencies

from zepid.

Before doing this, I have to figure out stochastic interventions as described by van der Laan.

Also will need to rewrite TMLE slightly. For stochastic interventions, need to have iterative process of Qn fitting (unlike the present TMLE which converges in a single step)

from zepid.

Due to various other requirements (statsmodels 0.10.0) and the less broad use-case for interference methods, this will become its own library. This applies to all interference causal inference methods

from zepid.