Skip to content

Interval logic#15

Draft
weymouth wants to merge 37 commits intomainfrom
IntervalLogic
Draft

Interval logic#15
weymouth wants to merge 37 commits intomainfrom
IntervalLogic

Conversation

@weymouth
Copy link
Owner

No description provided.

weymouth and others added 30 commits July 15, 2025 08:45
Added a slowly varying function γ to complex_path and nsp. Also cleaned up the finite_ranges logic.
Almost have the set difference version working, but I'm filtering out empty sets which have open ends that I need to chase to infinity. Almost?
Very clean and efficient and accurate...
They reintroduce high frequency w-waves.
Added more tests and found some issues for very large or small values of x,y. Now throws an error when M is empty (means the points are too far apart). Also added a fallback for when M doesn't contain 0 since we can't use γ in that case.
The integration of the panel helps, but the frequency is still far too high. Maybe there is a way to double panel integrate, like a BEM. But the single panel integral is not enough when z=0
Adding t=0 as a stationary point (even when it isn't) ensures f_m is used there. I also strip Duals before call finite_ranges. This obviates the fallback.

I also changed the damping function to t't*(1+t't) so it never contributes to the phase.
And a not-yet-working example to invert for q!
Test for correctness, but still working
I dispatched based on the size of c, which lets me clean up code and use SVectors. I eventually found the bug in the IBP, but it's allocating like crazy.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant