Feature/collections class

[granrothge]

Provides a start of scans collections class that allows pcolor plots and waterfall plots of multiple runs.

[pseudocubic]

First of all, thanks for the PR! I think that this will make a good addition to neutronpy.

I have briefly looked over it and my first impression is good. I will spend some time with it next week to give more complete comments and maybe submit some suggested changes to you. A couple of things that captured my attention are the docstrings, which will have to be complete before a merge, the class definition, which should be capitalized and inherit object (i.e. the "new" Python 3 style class definition), i.e. Scans(object), and the name of the package.

To address the last point, I'm not sure if scancollection is the right name (at the very least, to be consistent with the rest of the package it should be scan_collection). With the single class there it might be more appropriate as a module within the data package, but if we can foresee collections of other objects, like materials or instruments for example, then maybe a package named something like containers (collections would have been my choice, but better to not overlap with the standard library), with the submodule scans like you have it. I am certainly open to suggestions/discussion on this point.

[codecov-io]

Current coverage is 84.32% (diff: 89.18%)

Merging #91 into master will increase coverage by 0.12%

@@             master        #91   diff @@
==========================================
  Files            44         45     +1   
  Lines          3780       3854    +74   
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
+ Hits           3183       3250    +67   
- Misses          597        604     +7   
  Partials          0          0          

Powered by Codecov. Last update 983fe90...2a29af8

[granrothge]

As to the questions of naming I can change it to scan_collection. I can not think of a case where one would like a collection that can not be fit by the scans_collection and the scan. Say for an example if I wanted to do a waterfall plot as function of doping concentration. I think I would add a column for doping to the scan, then a variation by material would then be a scan collection. I think a collection of objects that would contain scans and other things might be better handled by a higher level collection yet, (maybe a NOSQL database), because your processing chain would likely be vary different. Say for example if you wanted to constrain an sqw model to both scan data and a magnetic susceptibility measurement, I think that putting the susceptibility data into a scan would add a lot of overhead.

I'll look into improving the doc strings.

[granrothge]

Let me know how you want the module named and I will make the change. I think the doc string are now ready to go.

[pseudocubic]

Thanks for updating the docstrings. Could you also add a one-liner in the top level docstring of the file?

The class definition should also be changed to class Scans(object):

After considering it some more, I think that this would would be better as an addition to the data package, i.e. in neutronpy/data/scans.py and add from .scans import Scans in the __init__.py. I would still keep the tests in a separate file like you have it (but rename to test_scans.py), instead of combining it with the data tests.

Could you also rebase to make a more simple commit history?

Otherwise I think this is pretty close to being ready for merge. I will take care of merging the docstrings into the sphinx-built documentation, and updating the changelog myself once we have merged.

[granrothge]

Ok so this should be good now.

[pseudocubic]

The directory neutronpy/scancollection and the file tests/test_scancollection.py are still left over after your changes, but once those are gone I should have no problem to merge the PR.

[granrothge]

the combination of the rebase, fetch and merge, brought things back that I had removed. Sorry. It should be good now.

[granrothge]

Finally, this should be ready.

[pseudocubic]

Thanks for the contribution and all your hard work! I have some minor bug fixes to include, and need to make sure the docs get generated, but hopefully I can get an incremental release onto the anaconda cloud in a few days.