The workflow ci.yml is referencing action actions/checkout using references v1. However this reference is missing the commit a6747255bd19d7a757dbdda8c654a9f84db19839 which may contain fix to the some vulnerability.
The vulnerability fix that is missing by actions version could be related to:
(1) CVE fix
(2) upgrade of vulnerable dependency
(3) fix to secret leak and others.
Please consider to update the reference to the action.

• use standard matrices to track particles
• write functions to conveniently create different initial Particle Distributions

For example a bend element B does not show its length, because it is defined in the base class 'element'

PEP518 defines a way to specifies build deps.

GCC and CFFI is needed to build the package.

See https://stackoverflow.com/questions/18023546/marking-cython-as-a-build-dependency

Python 3.8 now has a functools.cached_property

from functools import cached_property

class Foo:
    def __init__(self):
        self._x = 0

    @cached_property
    def x(self):
        print("Increment x!")
        return self._x

Create a new foo:

>>> foo = Foo()
>>> foo.x
Increment x!
1
>>> foo.x
1

The cache can be invalidated with:

>>> del foo.x
>>> foo.x
Increment x!
2

take a look at jsonpickle

Maybe it is better to use Decimal floating point numbers instead of the default binary floating point numbers. Another possiblity would be to use mm and an int type to store the length.

For the length attribute, the wrong representation of decimal numbers can become a problem: When accumulating the element ending position for a FODO lattice

{
  "name": "FODO_RING",
  "description": "This is the simplest possible strong focusing lattice.",
  "elements": {
    "D1": ["Drift", {"length": 0.55}],
    "Q1": ["Quadrupole", {"length": 0.2, "k1": 1.2}],
    "Q2": ["Quadrupole", {"length": 0.4, "k1": -1.2}],
    "B1": ["Dipole", {"length": 1.5, "angle": 0.392701, "e1": 0.1963505, "e2": 0.1963505}]
  },
  "sub_lattices": {
    "FODO": ["Q1", "D1", "B1", "D1", "Q2", "D1", "B1", "D1", "Q1"]
  },
  "lattice": ["FODO", "FODO", "FODO", "FODO", "FODO", "FODO", "FODO", "FODO"]
}

the length of the first cell gets calculated correctly. Cells beginning with the 6th cell have the wrong length:

First cell:
------------
end 0.75 D1
end 2.25 B1
end 2.8 D1
end 3.1999999999999997 Q2
end 3.75 D1
end 5.25 B1
end 5.8 D1
end 6.0 Q1
...
6th cell:
----------
end 30.200000000000003 Q1
end 30.750000000000004 D1
end 32.25 B1
end 32.8 D1
end 33.199999999999996 Q2
...
8th cell:
end 47.799999999999976 D1
end 47.99999999999998 Q1

The problem is, that this is inconsistent with the length attribute of the ring:

>>> ap.Lattice.from_file("fodo_ring.json").length
48.0

which, returns the "correct" length. This is because the Lattice object will add the length of all its sub_lattices (which have length 6.0) instead of adding the length of all its elements (like shown above).

return None for Twiss parameter if no solution

chromaticity is always recomputed ...

Maybe it would be the best to just pass n_kicks as parameter to TransferMatrices and do not allow it to be changed after an instance was created. This would make the implementation easier.

Tests:

• test should use as little as possible kicks to be fast

Portfolio like DIN A4 PDF with all important properties of Lattice:

• nice formatted table of elements and cells and all there attributes (length, angle, k1, k2, ...)
• plot of all Twiss parameter
• Tunes fractional and integer (also in kHz)
• natural chromaticity
• floorplot (maybe in later iterations)

The API Documenation should be update automatically whenever the codes changes.

Options:
autoapi
sphinx-apidoc
autosummary

check https://lattice-summaries.netlify.app/bessy2/bessy2_stduser-2019-05-07-v_1/madx

• Load lattice from file
• Save lattice from file
• Change magnets
• Linear Beam Dynamics
• Create Necktie plot
• Matrix Tracking
• Tune from particle oscillation

Was not fixed by #88
Supersedes #35

Sphinx-Gallery examples can be executed. This does currently not work on the readthedocs cli because of the C extensions:

readthedocs log

ModuleNotFoundError: No module named 'apace._clib'

https://github.com/andreasfelix/apace/blob/cb0edb9a766eee65eb09fcba339660eca0925dde/docs/conf.py#L45-L46

Solutions

• execute files locally and somehow push them to readthedocs
• make it work on readthedocs

Run automated tests for twiss parameter:

• Use Pytests
• check if Tunes are correct
• check beta, alpha, gamma functions for individual points

⚠️ Currently blocked by PyPy as it currently only supports Python 3.6.

Use Postponed Evaluation of Annotations for Cell in the classes module.

This will become default behavior in Python 4.0 and is available for Python 3.7+ by importing the future module:

from __future__ import annotations

Currently blocked by github actions, as it does not support pypy3.6.9

Optimally there would be different draw functions, which could be used for close-up/far away lattice plots

Draw elements like in WIlle page 228

different elements can not have the same name. If, then calculations error occurs in MatrixCalculation without warning or errror message.

add dict which returns a tuple of positions for a given element! similar to indices

Currently an instance of the Twiss class is linked a given Lattice object. If the Lattice object changes so will the Twiss object.

This can be counter-intuitive if one wants to save some reference values before optimizing a given lattice and can lead to mistakes:

twiss = ap.Twiss(lattice)
ref_beta_x = twiss.beta_x

# do the opitmization
beta_x = twiss.beta_x

Now ref_beta_x is bound to the same underlying numpy.ndarray as beta_x. This might be error prone.

One solution to that would be to always allocate a new numpy.ndarray when calculating th twiss parameter instead of reusing the old one. Test how large the overhead would be!

Discovered in #37 but master has the same problem:

self.element_n_kicks = {Drift: 1, Dipole: 1, Quadrupole: 1} will cause the test test_particle_trajectory to fail.