"""
Non-linear optics
"""
import numpy
from typing import Optional, Sequence
from scipy.special import factorial
from ..lattice import Element, Lattice
from ..tracking import internal_lpass
from .orbit import Orbit, find_orbit
from .linear import get_tune, get_chrom, linopt6
from .harmonic_analysis import get_tunes_harmonic
__all__ = ['detuning', 'chromaticity', 'gen_detuning_elem', 'tunes_vs_amp']
[docs]def tunes_vs_amp(ring: Lattice, amp: Optional[Sequence[float]] = None,
dim: Optional[int] = 0, nturns: Optional[int] = 512,
**kwargs) -> numpy.ndarray:
r"""Generates a range of tunes for varying x, y, or z amplitudes
Parameters:
ring: Lattice description
amp: Oscillation amplitude. If amp is None the
tunes are calculated from :py:func:`.linopt6`
else tracking is used
dim: Coordinate index of the oscillation
nturns: Number of turns
method: ``'laskar'`` or ``'fft'``. Default: ``'laskar'``
num_harmonics: Number of harmonic components to compute (before
mask applied)
fmin: Lower bound for tune
fmax: Upper bound for tune
hann: Turn on Hanning window. Default: :py:obj:`False`
Returns:
tunes (ndarray): Array of tunes qx, qy with shape (len(amp), 2)
"""
def _gen_part(ring, amp, dim, orbit, ld, nturns):
invx = numpy.array([[1/numpy.sqrt(ld['beta'][0]), 0],
[ld['alpha'][0]/numpy.sqrt(ld['beta'][0]),
numpy.sqrt(ld['beta'][0])]])
invy = numpy.array([[1/numpy.sqrt(ld['beta'][1]), 0],
[ld['alpha'][1]/numpy.sqrt(ld['beta'][1]),
numpy.sqrt(ld['beta'][1])]])
part = numpy.array([orbit, ] * len(amp)).T + 1.0e-6
part[dim, :] += amp
part = internal_lpass(ring, part, nturns=nturns)
sh = part.shape
partx = numpy.reshape(part[0, :], (sh[1], sh[3]))
partxp = numpy.reshape(part[1, :], (sh[1], sh[3]))
party = numpy.reshape(part[2, :], (sh[1], sh[3]))
partyp = numpy.reshape(part[3, :], (sh[1], sh[3]))
px = numpy.array([numpy.matmul(invx, [partx[i], partxp[i]])
for i in range(len(amp))])
py = numpy.array([numpy.matmul(invy, [party[i], partyp[i]])
for i in range(len(amp))])
return px[:, 0, :] - 1j*px[:, 1, :], py[:, 0, :] - 1j*py[:, 1, :]
l0, bd, _ = linopt6(ring)
orbit = l0['closed_orbit']
tunes = bd['tune']
if amp is not None:
partx, party = _gen_part(ring, amp, dim, orbit, l0, nturns)
qx = get_tunes_harmonic(partx, **kwargs)
qy = get_tunes_harmonic(party, **kwargs)
tunes = numpy.vstack((qx, qy)).T
return numpy.array(tunes)
[docs]def detuning(ring: Lattice,
xm: Optional[float] = 0.3e-4, ym: Optional[float] = 0.3e-4,
npoints: Optional[int] = 3,
nturns: Optional[int] = 512, **kwargs):
"""Computes the tunes for a sequence of amplitudes
This function uses :py:func:`tunes_vs_amp` to compute the tunes for
the specified amplitudes. Then it fits this data and returns
the detuning coefficiant dQx/dx, dQy/dx, dQx/dy, dQy/dy and the
qx, qy arrays versus x, y arrays
Parameters:
ring: Lattice description
xm: Maximum x amplitude
ym: Maximum y amplitude
npoints: Number of points in each plane
nturns: Number of turns for tracking
method: ``'laskar'`` or ``'fft'``. Default: ``'laskar'``
num_harmonics: Number of harmonic components to compute
(before mask applied)
fmin: Lower bound for tune
fmax: Upper bound for tune
hann: Turn on Hanning window. Default: :py:obj:`False`
Returns:
q0 (ndarray): qx, qy from horizontal and vertical amplitude scans.
Fit results with shape (2, 2)
q1 (ndarray): Amplitude detuning coefficients with shape (2, 2)
``[[dQx/dx, dQy/dx], [dQx/dy, dQy/dy]]``
x (ndarray): x amplitudes (npoints, )
q_dx (ndarray): qx, qy tunes as a function of x amplitude (npoints, 2)
y (ndarray): y amplitudes (npoints, )
q_dy (ndarray): qx, qy tunes as a function of y amplitude (npoints, 2)
"""
lindata0, _, _ = linopt6(ring)
gamma = (1 + lindata0.alpha * lindata0.alpha) / lindata0.beta
x = numpy.linspace(-xm, xm, npoints)
y = numpy.linspace(-ym, ym, npoints)
x2 = x * x
y2 = y * y
q_dx = tunes_vs_amp(ring, amp=x, dim=0, nturns=nturns, **kwargs)
q_dy = tunes_vs_amp(ring, amp=y, dim=2, nturns=nturns, **kwargs)
idx = numpy.isfinite(q_dx[:, 0]) & numpy.isfinite(q_dx[:, 1])
idy = numpy.isfinite(q_dy[:, 0]) & numpy.isfinite(q_dy[:, 1])
fx = numpy.polyfit(x2[idx], q_dx[idx], 1)
fy = numpy.polyfit(y2[idy], q_dy[idy], 1)
q0 = [[fx[1, 0], fx[1, 1]],
[fy[1, 0], fy[1, 1]]]
q1 = [[2 * fx[0, 0] / gamma[0], 2 * fx[0, 1] / gamma[0]],
[2 * fy[0, 0] / gamma[1], 2 * fy[0, 1] / gamma[1]]]
return numpy.array(q0), numpy.array(q1), x, q_dx, y, q_dy
[docs]def chromaticity(ring: Lattice, method: Optional[str] = 'linopt',
dpm: Optional[float] = 0.02, npoints: Optional[int] = 11,
order: Optional[int] = 3,
dp: Optional[float] = 0,
**kwargs):
r"""Computes the non-linear chromaticities
This function computes the tunes for the specified momentum offsets.
Then it fits this data and returns the chromaticity up to the given
order (npoints>order)
Parameters:
ring: Lattice description
method: ``'linopt'`` (dfault) returns the tunes from the
:py:func:`linopt6` function,
``'fft'`` tracks a single particle and computes the tunes with fft,
``'laskar'`` tracks a single particle and computes the tunes with
NAFF.
dpm: Maximum momentum deviation
npoints: Number of momentum deviations
order: Order of the fit
dp: Central momentum deviation
Returns:
fit (ndarray): Horizontal Vertical polynomial coefficients
from ``numpy.polyfit`` of shape (2, order+1)
dpa: dp array of shape (npoints,)
qz: tune array of shape (npoints, 2)
"""
if order == 0:
return get_chrom(ring, dp=dp, **kwargs)
elif order > npoints - 1:
raise ValueError('order should be smaller than npoints-1')
else:
dpa = numpy.linspace(-dpm, dpm, npoints)
qz = []
for dpi in dpa:
qz.append(get_tune(ring, method=method, dp=dp+dpi, remove_dc=True,
**kwargs))
fit = numpy.polyfit(dpa, qz, order)[::-1]
fitx = fit[:, 0]/factorial(numpy.arange(order + 1))
fity = fit[:, 1]/factorial(numpy.arange(order + 1))
return numpy.array([fitx, fity]), dpa, numpy.array(qz)
[docs]def gen_detuning_elem(ring: Lattice, orbit: Optional[Orbit] = None) -> Element:
"""Generates a detuning element
Parameters:
ring: Lattice description
orbit: Avoids looking for initial the closed orbit if it is
already known ((6,) array).
Returns:
detuneElem: Element reproducing the detuning of the ring with
amplitude and momentum
"""
if orbit is None:
orbit, _ = find_orbit(ring)
lindata0, _, _ = ring.linopt6(get_chrom=False, orbit=orbit)
xsi = get_chrom(ring.radiation_off(copy=True))
r0, r1, x, q_dx, y, q_dy = detuning(ring.radiation_off(copy=True),
xm=1.0e-4, ym=1.0e-4, npoints=3)
nonlin_elem = Element('NonLinear', PassMethod='DeltaQPass',
Betax=lindata0.beta[0], Betay=lindata0.beta[1],
Alphax=lindata0.alpha[0], Alphay=lindata0.alpha[1],
Qpx=xsi[0], Qpy=xsi[1],
A1=r1[0][0], A2=r1[0][1], A3=r1[1][1],
T1=-orbit, T2=orbit)
return nonlin_elem