Add support for using a random seed for repeatability

[kephale]

This adds a random seed to all_features2, and propagates some of those changes in the code.

This also adds a test to show that the results are the same with the same random seed and different with different seeds.

[kephale]

Current output:

======================================================================
FAIL: test_repeatability (__main__.TestAllFeatures2Repeatability)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/Users/kharrington/git/anmartinezs/polnet/tests/test_repeatability.py", line 116, in test_repeatability
    self.assertEqual(hashes1, hashes2, "Key output files are not identical between runs")
AssertionError: {'tom[16 chars]c': '734d4fd6d885a7857d2aa6c8d024af39', 'tomos[230 chars]7e6'} != {'tom[16 chars]c': '0e29104d55f8ac4f2186df77bd8df8cd', 'tomos[230 chars]02d'}
  {'tomos/poly_den_0.vtp': 'd021f5e9f13952d8d5df43a551ce7cab',
   'tomos/poly_skel_0.vtp': '0530b5fa41f6d6a102d7234b653c2f6c',
-  'tomos/tomo_den_0.mrc': '734d4fd6d885a7857d2aa6c8d024af39',
-  'tomos/tomo_lbls_0.mrc': 'b133b6d87b336f82c430fd70cf43ff60',
-  'tomos_motif_list.csv': 'ea6ec9517f678e747a5bb5f75ecff7e6'}
+  'tomos/tomo_den_0.mrc': '0e29104d55f8ac4f2186df77bd8df8cd',
+  'tomos/tomo_lbls_0.mrc': '1304d2534f288bcb8b6e013987c0b34d',
+  'tomos_motif_list.csv': '4d4258aa2eb3f5f317414f749e3f602d'} : Key output files are not identical between runs

[anmartinezs]

Hi kephale

It is nice you're making progresses with repeatability. I am not experienced in such tests, but I guess you're comparing if the output files are exactly the same between different runs. However, when these files are writen on disk maybe some unique information can be added to the file headers, such as time, that makes these files not exactly the same although the data contained are the same.

Best

[kephale]

@anmartinezs aha you got it! Now this compares the array contents of mrcs, and checks the csv outputs while ignoring some columns (e.g. differences in the mrc file paths for run1 and run2).

The test passes now!

[kephale]

@Fran-AM Great, sorry for the slow response! Thank you for catching those bugs. I've made fixes now.