Improve query planning for attribute functions

Improve query planning for attribute functions

Currently functions fall back to full table scans. Some of them could be optimized. For example,
strToUpper(foo) = 'FOO'
could be expanded to something we can query plan correctly:
foo = 'FOO' OR foo = 'foo' OR foo = 'Foo' ... // etc
That might not be feasible for longer words of course...
List of functions here: http://docs.geoserver.org/latest/en/user/filter/function_reference.html

Original JIRA Issue: https://geomesa.atlassian.net/browse/GEOMESA-1990

Key: GEOMESA-1990
Type: Improvement
Priority: Major
Status: To Do
Resolution: Unresolved
Reporter: Emilio Lahr-Vivaz
Created: Tue, 22 Aug 2017 13:43:23 -0400
Updated: Wed, 16 May 2018 16:14:14 -0400

Comments:

557058:da7f6dfa-f0a1-468b-aec5-2cb61273feb5 (Wed, 16 May 2018 16:14:14 -0400):

The logic is all wired through already. The method that needs to be implemented is:
org.locationtech.geomesa.filter.FilterHelper#extractFunctionBounds

Start with org.locationtech.geomesa.filter.FilterHelper#extractFunctionBounds, which the issue identifies as the unimplemented method, and consult the linked GeoServer filter function reference for the functions involved. Done means supported attribute functions produce query bounds instead of falling back to full table scans, while accounting for cases where expansion is impractical.