Use more array programming style to improve firesquid speed

[DylanVerstraete]

For some object we already use "array programming" meaning we collect all creates/updates inside a batch of blocks and save to the database later.

For farms, nodes, contracts, we still save one by one due to relations they have. In a future version, we should convert this to array programming style.

For example, you can read block by block and save entities to the db inbetween
or with batch processing you can: read X amount of blocks, save entities to be saved in memory and save to the DB whenever you processed the X amount of blocks => higher performance

[renauter]

More insights from @DylanVerstraete

Should be: we collect all creates/updates inside a batch of blocks and save to the database later
Array style programming could be phrased as batch style processing
For example, you can read block by block and save entities to the db in between or with batch processing you can: read X amount of blocks, save entities to be saved in memory and save to the DB whenever you processed the X amount of blocks => higher performance
Since every "save to the db" opens a connection to the postgres and has some overhead, opening the connection only once and saving an entire bunch of entities is therefor more performant