[WIP] DB Triggers unittest and profiling

[lucernae]

This PR will involve DB trigger tests which consists of:

1. Uploading flood map
2. Uploading flood event
3. Process intersected buildings
4. Refresh materialized view
5. Generate spreadsheets

To make sure our processing pipeline is complete, and within expected completion time.

[lucernae]

Summary

• trigger name is confusing so I rename it to execute in alphanumerical order
• however then I found out that we can't cascade refresh materialized view statement in separate trigger functions. If we do it like that, the next trigger will continue while the previous trigger has not been committed. So, it has to be done on a single trigger function. Like this: https://github.com/inasafe/inasafe-fba/pull/59/files#diff-f7922711dfd5dd4143055eeeacadb3b2
• I deleted unnecessary trigger functions. Such as to refresh mv_non_flooded_buildings on hazard_event update. We don't need to update those statistics, because it doesn't change on hazard_event insert/update
• Added sanity check to make sure flooded count is always less then total count.
• The flooded count and not flooded count sum does not match (it will be less) with total count because we don't count other building categories. @timlinux @NyakudyaA, we need to decide what to do with this.
• Some trigger statement uses wrong keyword like BEFORE/AFTER, so I change it. For example, mv flooded count needs to be executed AFTER hazard_event insert. While building total vulnerability needs to be BEFORE osm_buildings update.

[timlinux]

Looking good thanks: Only comment is to use inasafe prefix rather than fbf as I plan to make this platform a response based one too...