HDF5 dependence moved outside of Genome object

EricR86 commented 3 years ago

This PR moves HDF5 specifics out of the Genome object into newer chromosome list objects. The idea is this work will easily allow other file formats to be used with the Genome object properties/context manager.

The changes were done on a best effort basis to match the original intention of certain methods. Whether or not they should be changed or updated in some fashion would be best discussed in this PR.

EricR86 commented 3 years ago

@michaelmhoffman This is ready for your review

michaelmhoffman commented 3 years ago

Please let me know when ready to review.

EricR86 commented 2 years ago

@michaelmhoffman changes are ready for your review. There's a few things of note:

The Chromosome and HDF5 specifics were moved out to separate files. __init__.py was only going to grow larger and now it is much easier to see and reason which modules have what external depencies. For example, __init__.py no longer has any references to PyTables.
Supercontig has not been refactored out of its HDF5 dependency as was discussed in a code review. Hence there are currently some PyTables imports in the _chromosome.py file since it houses Supercontig as-is
The .attrs property was removed from the base Chromosome and moved into the HDF5 implementation. mins, max, etc. will need to be implemented (or error report) for other implementations.

Other than that all feedback currently in the PR should have been included.

hoffmangroup / genomedata

HDF5 dependence moved outside of Genome object #61