Fix abstraction for EBBO test

[fhenneke]

On thing that is not yet clear to me is what the correct abstraction for the EBBO test should be.
Generally I think we should compare buy and sell amounts of an executed order with buy and sell amounts of alternative executions.

The buy and sell amounts for the executed order can always be read from on chain data. The buy and sell amount for alternative executions can be obtained differently. Examples for alternative executions include solutions proposed by other solvers and solutions computed after the fact by Quasimodo or other solvers.

This would mean that the different EBBO test would differ in how they compute buy and sell amounts. They could then use the same code for the comparison itself.