doc: add some documentation how to add an index for scientific metadata

[cfelder]

Is your feature request related to a problem? Please describe.
We've done sone benchmarks using SampleDB and recognised lower performance for certain fields, e.g. using scientific metadata b/c they are not indexed.

Describe the solution you'd like
It would be nice to have some documentation about how to add an index either to the data model or using a mongodb command to an existing system or to have a small administrative utility to do so as mentioned on the SciCat annual meeting today.

[stephan271]

I agree that needs to be added. Do you have an example document and field that you would like to have indexed ? W could use it as a concrete example in the documenattion. Are you refering to the current loopback based backend or to the upcoming nestjs based backend ?

[cfelder]

For the existing models may the checksum of a file (chk) would be good candidate. For scientific metadata may wavelength is a candidate.

Our test have been carried out w/ the loopback backend.