Bug report

The __init__ method for scale.LogTransformBase takes a required positional argument nonpos, but it's derived classes are constructed without that required argument in the inverted method.

Code for reproduction

import matplotlib.scale as scale
tform = scale.InvertedLog10Transform() #OK
tform = tform.inverted() #not OK

Actual outcome

TypeError: __init__() missing 1 required positional argument: 'nonpos'

Expected outcome

proper construction of transform object.

Matplotlib version

• Operating system: ubuntu 16.04
• Matplotlib version: 2.0.0
• Matplotlib backend (print(matplotlib.get_backend())): Qt5Agg
• Python version: 3.6
• Jupyter version (if applicable):
• Other libraries: