Add support for addition of arbitrary metadata to DAISY entities

[pinarpink]

We need to be able to add arbitrary metadata to DAISY entitiies (Project, Study/Cohort, Dataset).

This functionality is initially needed to be able to associate scientific metadata with DAISY datasets. For this case the metadata will be a JSON BLOB.

Assignee needs to discuss options with the End2End team before embarking on the implementation. typed/untyped metadata, K-V pairs, so on.

[jLebioda]

The ways I can see to implement #294:

• add a TextField to CoreTrackedModel (parent class of Project/Dataset/Cohort/Partner) to store the complete collection of key-values as raw string. A disadvantage is that we must ensure that it contains a valid JSON, but it’s a bit more flexible.
• add a JSONField to CoreTrackedModel to store the complete collection of key-values as JSONb. Slightly less flexible (Postgresql-specific, and will always need a valid JSON), but offers validation on DB level. Slightly slower to write (converts text to some binary blob/probably negligible overhead), but more efficient in storing the data.
• add M2M key-value relationships (a single entry for every Object-Key-Value tuple - will probably be most efficient and flexible in terms of what it offers, but way harder to maintain compared to the other two.

I will implement the first option if there are no comments/objections

[jLebioda]

Closing this one, further upgrades will be conducted in #300 and #302