Clean up use of dataset `display_name`

[rokroskar]

There is some inconsistency with the use of dataset display name. Initially this was introduced to deal with really long dataset titles imported from external data repositories. However, the way it is used is inconsistent. For example, using renku dataset yields this:

renku dataset
ID                                    DISPLAY_NAME            VERSION    CREATED              CREATORS
------------------------------------  ----------------------  ---------  -------------------  ----------
2a2c6230-aeb2-41b0-89d5-f1cdd240d95a  maunaloa                           2019-01-14 12:29:45  R.Nijzink

But now to list the files I need to use a different name:

renku dataset ls-files maunaloa
ADDED    CREATORS    DATASET    PATH
-------  ----------  ---------  ------

renku dataset ls-files MaunaLoa
ADDED                CREATORS       DATASET    PATH
-------------------  -------------  ---------  --------------------------------------------------------
2019-11-04 10:55:23  Remko Nijzink  MaunaLoa   /<dataset path>/data/MaunaLoa/weekly_in_situ_co2_mlo.csv

So I somehow need to remember that the dataset was named MaunaLoa when it was created even though it's not possible for me to retrieve this information unless i look at the name of the directory for the dataset. We should standardize this behavior and make it easier for the user to refer to a dataset.

Perhaps a solution would be to ask the user what they want to name the dataset when they are importing it? That way we probably don't run the risk of absurdly long dataset names and the user will actually know under which name they can find it.