Separate project for the library components - VERY low priority!

[vetlewi]

Some parts of OMpy such as the Matrix, Vector, etc. are fairly general and could easily be reused by other projects.
In the future it will probably make sense to use tools such as Dask/Apache Spark/etc. to do sorting and correlations from the raw data. Having a package with the data representation used in OMpy ready could make it easier to transition to such tools.

I therefore suggest that once OMpy is stable one could think about separating some of the library functionality of OMpy into its own package:

1. OMpy - Oslo Method algorithms (Unfolding/FirstGen/Decomposition/Normalization)
2. OCLpy - Containers, utilities, etc. (vector/matrix/matrix-operations/etc.)