• Going to include a keyword argument to swap between the new way (where histnorm = 'probability density'), and the original (histnorm = 'probability')
• Need to add new tests for this functionality

@Kully @cldougl @theengineear @empet
Added an extra parameter called histnorm to deal with the two different types of distplots (the seaborn-like method/the original and probability density method)
I've kept the default for histnorm as probability for backwards compatibility.

@yankev The seaborn distplot is equivalent to plotly distplot corresponding to histnorm=probability density. Leaving the default as probability it's not a good idea because a lot of plots on Plotly cloud used wrong this key and people that take as examples these older plots can think that this version leads to a histogram comparable with an estimate of the probbaility density associated to data.

@yankev Here is the proof that seaborn distplot works like Plotly distplot with histnorm=probability density (see the last plot):http://nbviewer.jupyter.org/gist/empet/1bf38aea056f1d60481e600fc68c21be?flush_cache=true

@empet oh I see, thanks. Initially I was trying to keep some backwards compatibility, but it appears as though setting histnorm = 'probability density' is the correct approach.

@yankev Yes, this is the recommended default value for histnorm in distplot.

@yankev tests look great, thanks for adding those! 💃 me after pulling those histnorm values out into some constants somewhere!