Replacing Curve object with pd.DataFrame

[patrick-reinhard]

Hi @kwinkunks, please have a look at this draft PR covering:

• Moving out quality and plotting functionality from class methods to dedicated quality.py and plot.py modules.
• Moving out reader/writer logic from class methods to dedicated las.py module.
• Replacing subclassed np.array Curve class with a custom class object that is composed (has a) pd.DataFrame and meta data attributes. If you want to operate on the data you need to operate on curve.df instead on the curve directly.
• Add constructor methods Well.from_datasets()., keeping Well.from_lasio() and Well.from_las() intact.
• Moved test example files to assets subdirectory. Added conftest.py that constructs a test well that can be called by all tests in the tests directory, speeding up test ensemble execution.

To do:

• Adapt interfaces for new Curve object, repo internally and with striplog.
• Replace replication of pandas functionality, mostly around the concept of basis
• Add tests

There's still a lot to do but it would be helpful if you can already have a first glance at the new Curve and datasets structure.