PyAT 'PassMethod QuadLinearRadPass: library, module or trackFunction not found'

[space-budgie]

Issue

I am trying to analyze a lattice written in Matlab with PyAT. I have generated the lattice in Matlab, exported the accelerator as a .mat file and loaded the lattice in PyAT. I can do 4D tracking without problems, but when I execute lattice.enable_6d() and then try to track particles, I get the error:

RuntimeError: PassMethod QuadLinearRadPass: library, module or trackFunction not found

Potential Cause

I believe that .enable_6d() changes all pass methods that end with Pass to end with RadPass instead.

https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/lattice/elements.py#L164-L166

It seems like QuadLinearPass is a valid pass method, but QuadLinearRadPass is not.

https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/load/utils.py#L126-L141

I also cannot find any QuadLinearRadPass in the at/integrators directory.

Temporary fix

Looking at at/atintegrators/QuadLinearPass.c, it appears that QuadLinearPass already supports 6D calculations. Hence, I have inserted the code

if self.PassMethod == "QuadLinearPass":
    return "QuadLinearPass"

above line 164 of https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/lattice/elements.py.

This seems to work. I can enable 6D and track particles without problems. However, this solution is cleary very ad hoc. I also don't know if this tracks the particle correctly from a physics point of view.

[lcarver]

Hello,

You can do

Ring.enable_6d(quadrupole_pass='QuadLinearPass')

Which should do exactly what you want.

Cheers,

Lee

On 2024-09-14 20:39, space-budgie wrote:

ISSUE

I am trying to analyze a lattice written in Matlab with PyAT. I have generated the lattice in Matlab, exported the accelerator as a .mat file and loaded the lattice in PyAT. I can do 4D tracking without problems, but when I execute lattice.enable_6d() and then try to track particles, I get the error:

RuntimeError: PassMethod QuadLinearRadPass: library, module or trackFunction not found

POTENTIAL CAUSE

I believe that .enable_6d() changes all pass methods that end with Pass to end with RadPass instead.

https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/lattice/elements.py#L164-L166

It seems like QuadLinearPass is a valid pass method, but QuadLinearRadPass is not.

https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/load/utils.py#L126-L141

I also cannot find any QuadLinearRadPass in the at/integrators directory.

TEMPORARY FIX

Looking at at/atintegrators/QuadLinearPass.c, it appears that QuadLinearPass already supports 6D calculations. Hence, I have inserted the code

if self.PassMethod == "QuadLinearPass": return "QuadLinearPass"

above line 164 of https://github.com/atcollab/at/blob/5e3ca23b90725c24f74fd21a1e75058c9af59b9a/pyat/at/lattice/elements.py.

This seems to work. I can enable 6D and track particles without problems. However, this solution is cleary very ad hoc.

-- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you are subscribed to this thread.Message ID: @.***>

Links:

[1] https://github.com/atcollab/at/issues/835 [2] https://github.com/notifications/unsubscribe-auth/AD2N7UUFUGFUQAGOLFUQGYTZWR7FPAVCNFSM6AAAAABOHCF72WVHI2DSMVQWIX3LMV43ASLTON2WKOZSGUZDMNJVG43TANA

[lfarv]

Hello @space-budgie. QuadLinearPass is a simplistic passmethod which does a matrix-based transfer through the element. To model synchrotron radiation, you need a method based on a drift-kick sequence (radiation occurs in kicks). That's why there cannot be a QuadLinearRadPass method. QuadLinearPass was the default for quadrupoles a long time ago, but it has now changed for the more accurate StrMPoleSymplectic4Pass. Maybe you Matlab lattice was created a long time ago. For now, you have two options:

1. Use @lcarver's way of turning radiation on: you'll neglect synchrotron radiation in quadrupoles, which is good enough when running on-momentum,
2. Better option: convert your quadrupoles to use StrMPoleSymplectic4Pass: you get a more accurate tracking, at the cost of a small speed reduction, and you can turn radiation on in quads.