Editing/testing the Demo notebook

[drphilmarshall]

Here are my notes!

• I did some lintering of the python towards pep8, and made the docstrings numpydoc compliant. You should check to see if your editor can run pylint so you get warnings about your code - I started doing this year and am enjoying it! I love standards.
• I changed the name of the main class to Empiricist, calling it a "worker" class. This is because we are asking it to perform functions for us - so it is modeling a person doing calculations. The method names now make sense: "Empiricist! get_SN!"
• At first, I got an error when I tried to fit the model and test the BIC, because I had not installed XDGMM. I updated the README and demo notebook to point this out.
• The get_local_SB method belongs in the Empiricist class - the reason you needed it was because the current Empiricist.get_data method is not modular enough. I did not do this refactoring but will issue it.
• I kept your shortcut code in the notebook, since others will probably appreciate it! You can now Run All Cells
• I wasn't keen on appending to a list and then arraying it at the end, so tried to do it with numpy but failed. np.empty() does not actually make an empty array! I think the right thing to do is make an array of the correct shape up front, and then fill it. This now done. In th eprocess I fixed get_SN so that the output is always (n_SN, 3) in shape, even if n_SN=1

[drphilmarshall]

Oops, forgot to update the tests... BTW my version of the notebook is here. I edited the text too, to make it terser. Also I found that the distributions of radial positions are not significantly different - try a few different training /test splits and you'll see.

[drphilmarshall]

Alright! I think my work is done. Travis now runs the notebook as well as the unit tests, check out the .travis.yml file :-) Reproducible astronomy!