Unit testing for Dataset and Dimension

[ssomnath]

sidpy.Dataset is a core class in sidpy and needs to be fairly well tested as soon as possible.

[ssomnath]

I added ideas for testing Dataset in this file.

[ssomnath]

The latest PR merge added a fair bit of unit tests for these two classes. I think we need more unit testing for Dataset and Dimension to plug in remaining holes. Please consider using reports from coveralls.

Currently, the following things need additional tests:

• sidpy.Dataset:
  • h5_dataset property
  • closing the HDF5 file containing the data
  • copy() is untested
  • view_metadata()
  • view_original_metadata()
  • plot() <- consider verifying which Visualizer object is used for now
  • get_extent()
  • labels property setting / resetting
  • invalid values for value, quantity, modality, source, metadata, original_metadata, data_descriptor, data_type properties
• sidpy.Dimension:
  • variations on dimension_type especially UNKOWN
  • cases where __eq__ tests as False
  • We should be calling super's __eq__ for validating the values

[ssomnath]

@gduscher - I added a function called validate_dataset_properties() to test_dataset.py that should be fleshed out and called within each (wherever appropriate) unit test for comprehensive testing. The current rigor of the unit tests is insufficient, especially for such an important class.

We would need to do something similar for Dimension as well. As in, we would need a validate_dimension_properties() that all unit tests could simply call for a thorough validation.