Source code for meepmeep.backends.numba.point3dd.velocity

#  MeepMeep: fast orbit calculations for exoplanet modelling
#  Copyright (C) 2022-2026 Hannu Parviainen
#
#  This program is free software: you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation, either version 3 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#  along with this program.  If not, see <https://www.gnu.org/licenses/>.

"""Single-expansion-point 3D velocity evaluators with orbital-parameter derivatives."""

from numba import njit, prange, types
from numba.extending import overload
from numpy import floor, zeros, ndarray
from numpy.typing import NDArray

from ._common import _is_1d_array


@njit(fastmath=True, inline='always')
def _vel_cd_w(time, c, dc, dvx, dvy, dvz):
    """Write-into kernel shared by the scalar and vector evaluators.

    Writes the seven-parameter gradients into the caller-provided ``(7,)``
    buffers ``dvx``, ``dvy``, and ``dvz`` and returns the velocity values,
    so the hot vector loops reuse preallocated rows instead of allocating
    per sample.
    """
    vx = c[0, 1] + time * (2.0 * c[0, 2] + time * (3.0 * c[0, 3] + time * 4.0 * c[0, 4]))
    vy = c[1, 1] + time * (2.0 * c[1, 2] + time * (3.0 * c[1, 3] + time * 4.0 * c[1, 4]))
    vz = c[2, 1] + time * (2.0 * c[2, 2] + time * (3.0 * c[2, 3] + time * 4.0 * c[2, 4]))
    for k in range(7):
        dvx[k] = dc[k, 0, 1] + time * (2.0 * dc[k, 0, 2] + time * (3.0 * dc[k, 0, 3] + time * 4.0 * dc[k, 0, 4]))
        dvy[k] = dc[k, 1, 1] + time * (2.0 * dc[k, 1, 2] + time * (3.0 * dc[k, 1, 3] + time * 4.0 * dc[k, 1, 4]))
        dvz[k] = dc[k, 2, 1] + time * (2.0 * dc[k, 2, 2] + time * (3.0 * dc[k, 2, 3] + time * 4.0 * dc[k, 2, 4]))
    return vx, vy, vz


@njit(fastmath=True)
def _vel_cd_s(time, c, dc):
    """Scalar kernel for :func:`vel_cd`. See that function for documentation."""
    dvx = zeros(7)
    dvy = zeros(7)
    dvz = zeros(7)
    vx, vy, vz = _vel_cd_w(time, c, dc, dvx, dvy, dvz)
    return vx, vy, vz, dvx, dvy, dvz


def _vel_cd_v_body(time, c, dc):
    """Vector-kernel body for :func:`vel_cd`; see that function for documentation.

    Compiled twice: ``vel_cd_v`` is the serial kernel (``prange`` compiles
    as a plain ``range`` without ``parallel=True``) and ``vel_cd_vp`` the
    parallel twin. The loop writes only into per-sample output elements,
    so no per-thread scratch is needed.
    """
    n = time.size
    vx = zeros(n)
    vy = zeros(n)
    vz = zeros(n)
    dvx = zeros((n, 7))
    dvy = zeros((n, 7))
    dvz = zeros((n, 7))
    for j in prange(n):
        vx[j], vy[j], vz[j] = _vel_cd_w(time[j], c, dc, dvx[j], dvy[j], dvz[j])
    return vx, vy, vz, dvx, dvy, dvz


vel_cd_v = njit(fastmath=True)(_vel_cd_v_body)
vel_cd_vp = njit(fastmath=True, parallel=True)(_vel_cd_v_body)


[docs] def vel_cd(time: float | NDArray, c: NDArray, dc: NDArray): """ Evaluate the (vx, vy, vz) velocity and its orbital-parameter derivatives at an expansion-point-centered time. Centered velocity companion to `position.pos_cd`. The velocity components are obtained by analytically differentiating the 5th-order position polynomials, yielding 4th-order polynomials in `time` that are evaluated using Horner's scheme. The same differentiation is applied to the parameter-derivative coefficients so the result is the velocity together with its seven partial derivatives with respect to `(tc, p, a, i, e, w, lan)`. Accepts a scalar time or a 1-D array of times and dispatches to the appropriate kernel at compile time (inside ``@njit``) or at call time (pure Python), mirroring the value-only `velocity.vel_c`. Parameters ---------- time : float or ndarray Time(s) relative to the Taylor series expansion point. c : NDArray A (3, 5) coefficient matrix produced by `solve3d`. Rows index the spatial dimensions (x, y, z) and columns the Taylor order from position through snap (pre-scaled by the factorial of the order). dc : NDArray A (7, 3, 5) tensor of parameter-derivative coefficients produced by `solve3d_d`. The leading axis enumerates the seven Keplerian parameters in the canonical order `(tc, p, a, i, e, w, lan)`; the remaining axes mirror the layout of `c`. Returns ------- vx : float or ndarray Sky-plane x velocity in stellar radii per unit time. Shape (N,) for an array `time`. vy : float or ndarray Sky-plane y velocity in stellar radii per unit time. Shape (N,) for an array `time`. vz : float or ndarray Line-of-sight z velocity in stellar radii per unit time. Positive values indicate motion toward the observer. Shape (N,) for an array `time`. dvx : NDArray Partial derivatives of `vx` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. dvy : NDArray Partial derivatives of `vy` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. dvz : NDArray Partial derivatives of `vz` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. Notes ----- The pre-factors `1, 2, 3, 4` in front of `c[d, 1..4]` and `dc[k, d, 1..4]` are the chain-rule factors from differentiating `c[d, n] * time^n` (and likewise the derivative coefficients) with respect to `time`. Differentiation drops the truncation order by one, so the velocity polynomials are 4th order even though the underlying position expansion is 5th order. """ if isinstance(time, ndarray): return vel_cd_v(time, c, dc) return _vel_cd_s(time, c, dc)
@overload(vel_cd, jit_options={'fastmath': True}) def _vel_cd_overload(time, c, dc): if _is_1d_array(time): def impl(time, c, dc): return vel_cd_v(time, c, dc) return impl if isinstance(time, types.Float): def impl(time, c, dc): return _vel_cd_s(time, c, dc) return impl return None @njit(fastmath=True) def _vel_d_s(time, tc, p, c, dc, te): """Scalar kernel for :func:`vel_d`. See that function for documentation.""" epoch = floor((time - tc - te + 0.5 * p) / p) return _vel_cd_s(time - (tc + te + epoch * p), c, dc) def _vel_d_v_body(time, tc, p, c, dc, te): """Vector-kernel body for :func:`vel_d`; see that function for documentation. Compiled twice: ``vel_d_v`` is the serial kernel (``prange`` compiles as a plain ``range`` without ``parallel=True``) and ``vel_d_vp`` the parallel twin. The loop writes only into per-sample output elements, so no per-thread scratch is needed. """ n = time.size vx = zeros(n) vy = zeros(n) vz = zeros(n) dvx = zeros((n, 7)) dvy = zeros((n, 7)) dvz = zeros((n, 7)) for j in prange(n): epoch = floor((time[j] - tc - te + 0.5 * p) / p) vx[j], vy[j], vz[j] = _vel_cd_w(time[j] - (tc + te + epoch * p), c, dc, dvx[j], dvy[j], dvz[j]) return vx, vy, vz, dvx, dvy, dvz vel_d_v = njit(fastmath=True)(_vel_d_v_body) vel_d_vp = njit(fastmath=True, parallel=True)(_vel_d_v_body) def vel_d(time: float | NDArray, tc: float, p: float, c: NDArray, dc: NDArray, te: float = 0.0): """ Evaluate the (vx, vy, vz) velocity and its parameter derivatives at an absolute time. Direct counterpart of the centered `vel_cd`: it accepts an absolute observation time `time`, folds it back into a single orbital epoch around the expansion point `te`, and then evaluates the 4th-order velocity polynomials (and their parameter-derivative companions) using Horner's scheme via the centered kernel. The epoch bin is selected, not differentiated, so the gradient matches the centered `vel_cd` evaluated at the folded time. Accepts a scalar time or a 1-D array of times and dispatches to the appropriate kernel at compile time (inside ``@njit``) or at call time (pure Python). Parameters ---------- time : float or NDArray Absolute observation time(s) in the same units as `tc` and `p` (typically days). Scalar or array inputs are both accepted; the return type matches. tc : float Transit-centre time (time of inferior conjunction), on the same time axis as `time`. p : float Orbital period, used to fold `time` into a single epoch around the expansion point. c : NDArray A (3, 5) coefficient matrix produced by `solve3d`. Rows index the spatial dimensions (x, y, z) and columns the Taylor order from position through snap (pre-scaled by the factorial of the order). dc : NDArray A (7, 3, 5) tensor of parameter-derivative coefficients produced by `solve3d_d`. The leading axis enumerates the seven Keplerian parameters in the canonical order `(tc, p, a, i, e, w, lan)`; the remaining axes mirror the layout of `c`. te : float, optional Expansion-point offset from the transit centre [days] - the same value that was passed to `solve3d`. Defaults to 0.0, the expansion point at the transit centre. Returns ------- vx : float or ndarray Sky-plane x velocity in stellar radii per unit time. Shape (N,) for an array `time`. vy : float or ndarray Sky-plane y velocity in stellar radii per unit time. Shape (N,) for an array `time`. vz : float or ndarray Line-of-sight z velocity in stellar radii per unit time. Positive values indicate motion toward the observer. Shape (N,) for an array `time`. dvx : NDArray Partial derivatives of `vx` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. dvy : NDArray Partial derivatives of `vy` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. dvz : NDArray Partial derivatives of `vz` w.r.t. `(tc, p, a, i, e, w, lan)`. Shape (7,) for a scalar `time`, (N, 7) for an array `time`. """ if isinstance(time, ndarray): return vel_d_v(time, tc, p, c, dc, te) return _vel_d_s(time, tc, p, c, dc, te) @overload(vel_d, jit_options={'fastmath': True}) def _vel_d_overload(time, tc, p, c, dc, te=0.0): if _is_1d_array(time): def impl(time, tc, p, c, dc, te=0.0): return vel_d_v(time, tc, p, c, dc, te) return impl if isinstance(time, types.Float): def impl(time, tc, p, c, dc, te=0.0): return _vel_d_s(time, tc, p, c, dc, te) return impl return None