Naming of the keyword arguments of simple_ring

[lfarv]

The recent simple_ring function has some arguments in common with the existing sigma_matrix function. However, their names are slightly different. Apart from the mnemonic inconvenience, this prevents from using the same dictionary as input for both functions (I encountered the problem when trying to compare the "design" Σ-matrix with the output of tracking a population through simple_ring: I need both functions).

So I would like to rename the simple_ring arguments coherently with those of sigma_matrix: beta_x -> betax alpha_x -> alphax emit_x -> emitx sigma_dp -> espread For the last one, strictly speaking, it's indeed a momentum spread (not energy), so we could envisage adding an alias in sigma_matrix (mspread or even keeping sigma_dp). But I would not like to alias all the arguments…

Next: simple_ring could accept the same input as sigma_matrix:

• twiss_in, using R if present (this can define a coupled lattice), or alpha and beta (for uncoupled lattices),
• otherwise alphax, … as now
• one could even accept a Σ-matrix, which can replace alpha*, beta*, emit*, sigma_dp*.

Third: we miss the Matlab simple ring: atsimplering

Otherwise, simple_ring worked nicely in all my tests!

[lfarv]

Anybody willing to do something about that ? Or shall I do it ?

[swhite2401]

I think @lcarver said he would do it, but most likely not for matlab...

[lfarv]

Fine. Apart from naming, better being done before the function gets widely used, all my other suggestions are neither critical nor urgent, just ideas!

For Matlab, same: not urgent, I might do it some time later…

[lcarver]

Sorry, I was off yesterday. I will do it for your first 2 points. But yes, I will not do the MATLAB! Although I am not sure of the following point:

"one could even accept a Σ-matrix, which can replace alpha, beta, emit, sigma_dp.`"

This is not clear to me.

[lfarv]

This is not clear to me.

The Σ-matrix (beam matrix as returned by ohmienvelope, for instance) contains all the optics and emittance information. So, with Σ and tunes, you can get the M₆₆ matrix and the emittances. Adding the damping times, you have everything.

Here, since the longitudinal motion is treated apart, I think we should use a 4D twiss_in input, generate a 4x4 transfer map and add the 44, 45, 54 and 55 terms as you do it now.

So to start from the R matrix in twiss_in, the easiest is first to build Σ = ε₁.R₁ + ε₂.R₂, then use a_matrix to get the A matrix, use the tunes to generate the rotation matrix S and finally build M₄₄ = A.S.A^-1

If starting directly from Σ, a_matrix will also return the needed emittances.

[lcarver]

fixed in #655