todos.txt

Create a fetch query caching engine
-----------------------------------

-> create a lokijs based database for fetched results. the object holds properties to quickly access the cache.

for performance values that define the fetch request are stored on many but not to many separate properties

those should be (so far):

the database that was used (main node)
the entity that was fetched
the where - statements that were used
the return statement that was used
also be able to add custom identifiers that are used in insert-clearing (as written below)

and after fetch:
the ids that were fetched (also mapped ones)
the mapping that was used

and on that object the fetched data is stored to deliver without contacting the database

-> when fetched using cache
1) find the cache-entries with the same entity name
2) check if the where - statements and return-statements fit. if there are more that 50 (or whatever threshold is defined) to check then compare incrementally (first word is same, 2nd word is same, 3rd word ...)
3) if found, return result without contacting database

-> when object is inserted / updated / deleted
remove all cached that fit database, entity
and if given one or many custom identifier only the ones that match it

cache should be able to scale horizontally

Create an output caching engine
-------------------------------