[FEATURE] Enhance cost computing mechanism in AbstractCalciteIndexScan

[qianheng-aws]

Is your feature request related to a problem?
The current cost computing mechanism in AbstractCalciteIndexScan should be out-of-date after we've introduced so many push down rules for variety cases.

Some bad cases:

1. Prefer more scripts than less script with additional project operator: Push down project operator with non-identity projections into scan #4279 (comment)
2. Prefer scan(agg+project) + agg than scan(agg)
3. ...

The shortcoming of current mechanism:

1. Mix rows estimation with the cost computing in one place
2. Lack of the delicate cost computing for scripts

What solution would you like?
A clear and concise description of what you want to happen.

What alternatives have you considered?
A clear and concise description of any alternative solutions or features you've considered.

Do you have any additional context?
Add any other context or screenshots about the feature request here.