Fix temporal resolution connection_ratio_out_in

[gnawin]

Fixes #1146

Checklist before merging

• Documentation is up-to-date
• Unit tests have been added/updated accordingly
• Code has been formatted according to SpineOpt's style
• Unit tests pass

[gnawin]

Hi @DillonJ @manuelma @Tasqu ,

t indice:
I changed t resolution from lowest to highest, and also modified the constraint accordingly. Does it make sense to you?

s indice:
To me, it works as expected, if s is only parent. But when there is parent and child, it does not. Could you give some instructions on where/how to make changes?

Thanks!

[gnawin]

I made some updates, could you have a further review @Tasqu? Thanks!

Another issue that I have not touched upon is the delays. Since delay basically introduces a separate resolution, we (@datejada and I) found out:

• if the delay is multiples of the resolutions from the node, it works.
• otherwise, it does not. What would be desired is that the constraint indices are the highest resolution between the nodes and the delay. If we decide to go that way, any idea of how the implementation works?

[Tasqu]

Another issue that I have not touched upon is the delays. Since delay basically introduces a separate resolution, we found out:

• if the delay is multiples of the resolutions from the node, it works.
• otherwise, it does not. What would be desired is that the constraint indices are the highest resolution between the nodes and the delay. If we decide to go that way, any idea of how the implementation works?

This might be tricky. I took a quick look, and I'm not sure why it wouldn't be working even if the delay isn't an exact match. My best guess is that it might have something to do with the to_time_slice and _to_time_slice functions, as those are used to figure out which time slice the delay should be pointing at.

Unfortunately, I most likely don't have time to look into this any further for now either :(