investigate different default ordering for dataset response in opensearch

[rsmith013]

I would suggest sorting by the following facets in the order below:

1. Processing Level (ideally custom, so e.g. L4 is higher priority than L2, but 'unknown' is lowest priority)
2. Data Type
3. Product
4. Product Version (highest first)
5. Alphabetical (default if no other facets - possibly reverse?)

I think that should work for most of the ECV's, but would be good to check exactly what that would produce before we finalise it

[rsmith013]

Move query ordering to:

• datatype (asc),
• processing_level (desc)
• productString(asc)
• productVersion(desc)
• title(asc)