client: Speed up row access by index

[dave-tucker]

This adds a new api to the cache - RowByMdoel which will calculate the
will look up by the UUID of the model, or by the matching the model
against the table indexes. This is much faster as it doesn't need to
iterate across the entire cache.

The Get API has been updated to use this new function

Fixes: #150

Signed-off-by: Dave Tucker dave@dtucker.co.uk

[coveralls]

Pull Request Test Coverage Report for Build 990617578

• 20 of 25 (80.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.1%) to 76.823%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
cache/cache.go	15	20	75.0%

Totals
Change from base Build 990592167:	0.1%
Covered Lines:	3202
Relevant Lines:	4168

---

💛 - Coveralls

[amorenoz]

In order to really verify that the data contained in model is valid and nondefault, consider using mapperInfo.getValidIndexes() (which btw, includes the "_uuid")

[dave-tucker]

That would reqiure exporting that function. The loop in getValidIndexes looks complicated/expensive for every time we Get from the cache - especially as we already know the indexes for the table are valid as we got them from the schema directly. I've pushed an update that is ~functionally equivalent - RowByModel can cope with the UUID, or other indexes but the implementation is a bit different.