/*
 * Copyright (c) 2002, 2017 Jens Keiner, Stefan Kunis, Daniel Potts
 *
 * 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 2 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, write to the Free Software Foundation, Inc., 51
 * Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
 */

/**
 * @defgroup nfsft NFSFT - Nonequispaced fast spherical Fourier transform
 * @{
 *
 * This module implements nonuniform fast spherical Fourier transforms. In the
 * following, we abbreviate the term "nonuniform fast spherical Fourier
 * transform" by NFSFT.
 *
 * \section Preliminaries
 * This section summarises basic definitions and properties related to spherical
 * Fourier transforms.
 *
 * \subsection sc Spherical Coordinates
 * Every point in \f$\mathbb{R}^3\f$ can be described in \e spherical \e
 * coordinates by a vector \f$(r,\vartheta,\varphi)^{\mathrm{T}}\f$ with the
 * radius \f$r \in \mathbb{R}^{+}\f$ and two angles \f$\vartheta \in [0,\pi]\f$,
 * \f$\varphi \in [-\pi,\pi)\f$.
 * We denote by \f$\mathbb{S}^2\f$ the two-dimensional unit sphere embedded
 * into \f$\mathbb{R}^3\f$, i.e.
 * \f[
 *   \mathbb{S}^2 := \left\{\mathbf{x} \in \mathbb{R}^{3}:\;
 *   \|\mathbf{x}\|_2=1\right\}
 * \f]
 * and identify a point from \f$\mathbb{S}^2\f$ with the corresponding vector
 * \f$(\vartheta,\varphi)^{\mathrm{T}}\f$. The
 * spherical coordinate system is illustrated in the following figure:
 *
 * <center>
 * \image html sphere.png ""
 * \image latex sphere.pdf "" width=0.45\textwidth
 * </center>
 *
 * For consistency with the other modules and the conventions used there, we
 * also use \e swapped \e scaled \e spherical \e coordinates \f$x_1 :=
 * \frac{\varphi}{2\pi}\f$, \f$x_2 := \frac{\vartheta}{2\pi}\f$ and identify a
 * point from \f$\mathbb{S}^2\f$ with the vector
 * \f$\mathbf{x} := \left(x_1,x_2\right) \in
 *  [-\frac{1}{2}, \frac{1}{2}) \times [0,\frac{1}{2}]\f$.
 *
 * \subsection lp Legendre Polynomials
 * The \e Legendre \e polynomials \f$P_k : [-1,1]
 * \rightarrow \mathbb{R}$, $k \in \mathbb{N}_{0}\f$ as \e classical \e
 * orthogonal \e polynomials are given by their corresponding \e Rodrigues \e
 * formula
 * \f[
 *   P_k(t) := \frac{1}{2^k k!} \frac{\text{d}^k}{\text{d} t^k}
 *   \left(t^2-1\right)^k.
 * \f]
 * The corresponding three-term recurrence relation is
 * \f[
 *   (k+1)P_{k+1}(t) = (2k+1) x P_{k}(t) - k P_{k-1}(t) \quad (k \in
 *   \mathbb{N}_0).
 * \f]
 * With
 * \f[
 *   \left< f,g \right>_{\text{L}^2\left([-1,1]\right)} :=
 *   \int_{-1}^{1} f(t) g(t) \text{d} t
 * \f]
 * being the usual \f$\text{L}^2\left([-1,1]\right)\f$ inner product,
 * the Legendre polynomials obey the orthogonality condition
 * \f[
 *   \left< P_k,P_l \right>_{\text{L}^2\left([-1,1]\right)} = \frac{2}{2k+1}
 *   \delta_{k,l}.
 * \f]
 *
 * \remark The normalisation constant \f$ c_k := \sqrt{\frac{2k+1}{2}}\f$
 * renders the scaled Legendre polynomials \f$c_k P_k\f$ orthonormal with
 * respect to the induced \f$\text{L}^2\left([-1,1]\right)\f$ norm
 * \f[
 *   \|f\|_{\text{L}^2\left([-1,1]\right)} :=
 *   \left(<f,f>_{\text{L}^2\left([-1,1]\right)}\right)^{1/2} =
 *   \left(\int_{-1}^{1} |f(t)|^2 \; \text{d} t\right)^{1/2}.
 * \f]
 *
 * \subsection alf Associated Legendre Functions
 * The \a associated \a Legendre \a functions \f$P_k^n : [-1,1] \rightarrow
 * \mathbb{R} \f$, \f$n \in \mathbb{N}_0\f$, \f$k \ge n\f$ are defined by
 * \f[
 *   P_k^n(t) := \left(\frac{(k-n)!}{(k+n)!}\right)^{1/2}
 *   \left(1-t^2\right)^{n/2} \frac{\text{d}^n}{\text{d} t^n} P_k(t).
 * \f]
 * For \f$n = 0\f$, they coincide with the Legendre polynomials, i.e.
 * \f$P_k^0 = P_k\f$.
 * The associated Legendre functions obey the three-term recurrence relation
 * \f[
 *   P_{k+1}^n(t) = v_{k}^n t P_k^n(t) + w_{k}^n P_{k-1}^n(t) \quad (k \ge n),
 * \f]
 * with \f$P_{n-1}^n(t) := 0\f$, \f$P_{n}^n(t) := \frac{\sqrt{(2n)!}}{2^n n!}
 * \left(1-t^2\right)^{n/2}\f$, and
 * \f[
 *   v_{k}^n := \frac{2k+1}{((k-n+1)(k+n+1))^{1/2}}\; ,\qquad
 *   w_{k}^n := - \frac{((k-n)(k+n))^{1/2}}{((k-n+1)(k+n+1))^{1/2}}.
 * \f]
 * For fixed \f$n\f$, the set \f$\left\{P_k^n:\: k
 * \ge n\right\}\f$ forms a complete set of orthogonal functions in
 * \f$\text{L}^2\left([-1,1]\right)\f$
 * with
 * \f[
 *   \left< P_k^n,P_l^n \right>_{\text{L}^2\left([-1,1]\right)} = \frac{2}{2k+1}
 *   \delta_{k,l} \quad (0 \le n \le k,l).
 * \f]
 *
 * \remark The normalisation constant \f$ c_k = \sqrt{\frac{2k+1}{2}}\f$
 * renders the scaled associated Legendre functions \f$c_k P_k^n\f$ orthonormal
 * with respect to the induced \f$\text{L}^2\left([-1,1]\right)\f$ norm
 * \f[
 *   \|f\|_{\text{L}^2\left([-1,1]\right)} :=
 *   \left(<f,f>_{\text{L}^2\left([-1,1]\right)}\right)^{1/2} =
 *   \left(\int_{-1}^{1} |f(t)|^2 \; \text{d} t\right)^{1/2}.
 * \f]
 *
 * \subsection sh Spherical Harmonics
 * The standard orthogonal basis of spherical harmonics for \f$\text{L}^2
 * \left(\mathbb{S}^2\right)\f$ with yet unnormalised basis functions
 * \f$\tilde{Y}_k^n : \mathbb{S}^2 \rightarrow \mathbb{C}\f$ is given by
 * \f[
 *   \tilde{Y}_k^n(\vartheta,\varphi) := P_k^{|n|}(\cos\vartheta)
 *   \mathrm{e}^{\mathrm{i} n \varphi}
 * \f]
 * with the usual \f$\text{L}^2\left(\mathbb{S}^2\right)\f$ inner product
 * \f[
 *   \left< f,g \right>_{\mathrm{L}^2\left(\mathbb{S}^2\right)} :=
 *   \int_{\mathbb{S}^2} f(\vartheta,\varphi) \overline{g(\vartheta,\varphi)}
 *   \: \mathrm{d} \mathbf{\xi} := \int_{-\pi}^{\pi} \int_{0}^{\pi}
 *   f(\vartheta,\varphi) \overline{g(\vartheta,\varphi)} \sin \vartheta
 *   \; \mathrm{d} \vartheta \; \mathrm{d} \varphi.
 * \f]
 * The normalisation constant \f$c_k^n := \sqrt{\frac{2k+1}{4\pi}}\f$ renders
 * the  scaled basis functions
 * \f[
 *   Y_k^n(\vartheta,\varphi) := c_k^n P_k^{|n|}(\cos\vartheta)
 *   \mathrm{e}^{\mathrm{i} n \varphi}
 * \f]
 * orthonormal with respect to the induced \f$\text{L}^2\left(\mathbb{S}^2
 * \right)\f$ norm
 * \f[
 *   \|f\|_{\text{L}^2\left(\mathbb{S}^2\right)} =
 *   \left(<f,f>_{\text{L}^2\left(\mathbb{S}^2\right)}\right)^{1/2} =
 *   \left(\int_{-\pi}^{\pi} \int_{0}^{\pi} |f(\vartheta,\varphi)|^2 \sin
 *   \vartheta \; \mathrm{d} \vartheta \; \mathrm{d} \varphi\right)^{1/2}.
 * \f]
 * A function \f$f \in \mathrm{L}^2\left(\mathbb{S}^2\right)\f$ has the
 * orthogonal expansion
 * \f[
 *   f = \sum_{k=0}^{\infty} \sum_{n=-k}^{k} \hat{f}(k,n) Y_k^n,
 * \f]
 * where the coefficients \f$\hat{f}(k,n) := \left< f, Y_k^{n}
 * \right>_{\mathrm{L}^2\left(\mathbb{S}^2\right)}\f$ are the \e spherical
 * \e Fourier \e coefficients and the equivalence is understood in the
 * \f$\mathrm{L}^2\f$-sense.
 *
 *
 * \section nfsfts Nonuniform Fast Spherical Fourier Transforms
 *
 * This section describes the input and output relation of the spherical
 * Fourier transform algorithms and the layout of the corresponding plan
 * structure.
 *
 * \subsection ndsft Nonuniform Discrete Spherical Fourier Transform
 * The \e nonuniform \e discrete \e spherical \e Fourier \e transform (\e NDSFT)
 * is defined as follows:
 * \f[
 *     \begin{array}{rcl}
 *       \text{\textbf{Input}} & : & \text{coefficients }
 *         \hat{f}(k,n) \in \mathbb{C} \text{ for } k=0,\ldots,N,\;n=-k,
 *         \ldots,k,\; N \in \mathbb{N}_0,\\[1ex]
 *                             &   & \text{arbitrary nodes } \mathbf{x}(m) \in
 *         [-\frac{1}{2},\frac{1}{2}] \times [0,\frac{1}{2}]
 *         \text{ for } m=0,\ldots,M-1, M \in \mathbb{N}. \\[1ex]
 *       \text{\textbf{Task}}  & : & \text{evaluate } f(m) := f\left(
 *       \mathbf{x}(m)\right) = \sum_{k=0}^N \sum_{n=-k}^k \hat{f}_k^n
 *         Y_k^n\left(\mathbf{x}(m)\right) \text{ for } m=0,\ldots,M-1.
 *         \\[1ex]
 *       \text{\textbf{Output}} & : & \text{coefficients } f(m) \in
 *         \mathbb{C} \text{ for } m=0,\ldots,M-1.\\
 *     \end{array}
 * \f]
 *
 * \subsection andsft Adjoint Nonuniform Discrete Spherical Fourier Transform
 * The \e adjoint \e nonuniform \e discrete \e spherical \e Fourier \e transform
 * (\e adjoint \e NDSFT)
 * is defined as follows:
 * \f[
 *     \begin{array}{rcl}
 *       \text{\textbf{Input}} & : & \text{coefficients } f(m) \in
 *         \mathbb{C} \text{ for } m=0,\ldots,M-1, M \in \mathbb{N},\\
 *                             &   & \text{arbitrary nodes } \mathbf{x}(m) \in
 *         [-\frac{1}{2},\frac{1}{2}] \times [0,\frac{1}{2}] \text{ for }
 *         m=0,\ldots,M-1, N \in \mathbb{N}_0.\\[1ex]
 *       \text{\textbf{Task}}  & : & \text{evaluate } \hat{f}(k,n)
 *         := \sum_{m=0}^{M-1} f(m) \overline{Y_k^n\left(\mathbf{x}(m)\right)}cd Do
 *         \text{ for } k=0,\ldots,N,\;n=-k,\ldots,k.\\[1ex]
 *       \text{\textbf{Output}} & : & \text{coefficients }
 *         \hat{f}(k,n) \in \mathbb{C} \text{ for }
 *         k=0,\ldots,N,\;n=-k,\ldots,k.\\[1ex]
 *     \end{array}
 * \f]
 *
 * \subsection dl Data Layout
 * This section describes the public  layout of the \ref nfsft_plan structure
 * which
 * contains all data for the computation of the aforementioned spherical Fourier
 * transforms. The structure contains private (no read or write allowed), public
 * read-only (only
 * read access permitted), and public read-write (read and write access allowed)
 * members. In the following, we indicate read and write access by \c read and
 * \c write. The public members are structured as follows:
 * \li \c N_total (\c read)
 *        The total number of components in \c f_hat. If the bandwidth is
 *        \f$N \in \mathbb{N}_0\f$, the total number of components in \c f_hat
 *        is \c N_total \f$ = (2N+2)^2\f$.
 * \li \c M_total (\c read)
 *        the total number of samples \f$M\f$
 * \li \c f_hat (\c read-write)
 *        The flattened array of spherical Fourier coefficents. The array
 *        has length \f$(2N+2)^2\f$ such that valid indices \f$i \in
 *        \mathbb{N}_0\f$ for array access \c f_hat \c[ \f$i\f$ \c] are
 *        \f$i=0,1,\ldots,(2N+2)^2-1\f$.
 *        However, only read and write access to indices corresponding to
 *        spherical Fourier coefficients \f$\hat{f}(k,n)\f$ is defined. The index
 *        \f$i\f$ corresponding to the spherical Fourier coefficient
 *        \f$\hat{f}(k,n)\f$ with \f$0 \le k \le M\f$, \f$-k \le n \le k\f$ is
 *        \f$i = (N+2)(N-n+1)+N+k+1\f$. For convenience, the helper macro
 *        \ref NFSFT_INDEX(k,n) provides the necessary index calculations such
 *        that
 *        one can write \c f_hat[ \c NFSFT_INDEX(\f$k,n\f$\c)] \c =
 *        \c ... to access
 *        the component corresponding to \f$\hat{f}(k,n)\f$.
 *        The data layout is due to implementation details.
 * \li \c f (\c read-write)
 *        the array of coefficients \f$f(m)\f$ for \f$m=0,\ldots,M-1\f$ such
 *        that \c f[\f$m\f$\c] = \f$f(m)\f$
 * \li \c N (\c read)
 *        the bandwidth \f$N \in \mathbb{N}_0\f$
 * \li \c x
 *        the array of nodes \f$\mathbf{x}(m) \in
 *        [-\frac{1}{2},\frac{1}{2}] \times [0,\frac{1}{2}]\f$ for \f$m = 0,
 *        \ldots,M-1\f$ such that \c x[\f$2m\f$\c] = \f$x_1\f$ and
 *        \c x[\f$2m+1\f$\c] = \f$x_2\f$
 *
 * \subsection gtn Good to know...
 * When using the routines of this module you should bear in mind the following:
 * \li The bandwidth \f$N_{\text{max}}\f$ up to which precomputation is
 *   performed is always chosen as the next power of two with respect to the
 *   specified maximum bandwidth.
 * \li By default, the NDSFT transforms (see \ref nfsft_direct_trafo, \ref nfsft_trafo)
 *   are allowed to destroy the input \c f_hat while the input \c x is
 *   preserved. On the contrary, the adjoint NDSFT transforms
 *   (see \ref nfsft_direct_adjoint, \ref nfsft_adjoint) do not destroy the input
 *   \c f and \c x by default. The desired behaviour can be assured by using the
 *   \ref NFSFT_PRESERVE_F_HAT, \ref NFSFT_PRESERVE_X, \ref NFSFT_PRESERVE_F and
 *   \ref NFSFT_DESTROY_F_HAT, \ref NFSFT_DESTROY_X, \ref NFSFT_DESTROY_F
 *   flags.
 */

/*! \struct nfsft_plan
 * NFSFT transform plan
 */

/*! \fn void nfsft_init(nfsft_plan *plan, int N, int M)
 * Creates a transform plan.
 *
 * \arg plan a pointer to a \ref nfsft_plan structure
 * \arg N the bandwidth \f$N \in \mathbb{N}_0\f$
 * \arg M the number of nodes \f$M \in \mathbb{N}\f$
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_init_advanced(nfsft_plan* plan, int N, int M, unsigned int nfsft_flags)
 * Creates a transform plan.
 *
 * \arg plan a pointer to a \verbatim nfsft_plan \endverbatim structure
 * \arg N the bandwidth \f$N \in \mathbb{N}_0\f$
 * \arg M the number of nodes \f$M \in \mathbb{N}\f$
 * \arg nfsft_flags the NFSFT flags
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_init_guru(nfsft_plan *plan, int N, int M, unsigned int nfsft_flags, nsigned int nfft_flags, int nfft_cutoff)
 * Creates a transform plan.
 *
 * \arg plan a pointer to a \verbatim nfsft_plan \endverbatim structure
 * \arg N the bandwidth \f$N \in \mathbb{N}_0\f$
 * \arg M the number of nodes \f$M \in \mathbb{N}\f$
 * \arg nfsft_flags the NFSFT flags
 * \arg nfft_cutoff the NFFT cutoff parameter
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_precompute(int N, R kappa, unsigned int nfsft_flags, unsigned int fpt_flags)
 * Performes precomputation up to the next power of two with respect to a given
 * bandwidth \f$N \in \mathbb{N}_2\f$. The threshold parameter \f$\kappa \in
 * \mathbb{R}^{+}\f$ determines the number of stabilization steps computed in
 * the discrete polynomial transform and thereby its accuracy.
 *
 * \arg N the bandwidth \f$N \in \mathbb{N}_0\f$
 * \arg threshold the threshold \f$\kappa \in \mathbb{R}^{+}\f$
 * \arg nfsft_precomputation_flags the NFSFT precomputation flags
 * \arg fpt_precomputation_flags the FPT precomputation flags
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_forget(void)
 * Forgets all precomputed data.
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_direct_trafo(nfsft_plan* plan)
 * Executes a direct NDSFT, i.e. computes for \f$m = 0,\ldots,M-1\f$
 * \f[
 *   f(m) = \sum_{k=0}^N \sum_{n=-k}^k \hat{f}(k,n) Y_k^n\left(2\pi x_1(m),
 *   2\pi x_2(m)\right).
 * \f]
 *
 * \arg plan the plan
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_direct_adjoint(nfsft_plan* plan)
 * Executes a direct adjoint NDSFT, i.e. computes for \f$k=0,\ldots,N;
 * n=-k,\ldots,k\f$
 * \f[
 *   \hat{f}(k,n) = \sum_{m = 0}^{M-1} f(m) Y_k^n\left(2\pi x_1(m),
 *   2\pi x_2(m)\right).
 * \f]
 *
 * \arg plan the plan
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_trafo(nfsft_plan* plan)
 * Executes a NFSFT, i.e. computes for \f$m = 0,\ldots,M-1\f$
 * \f[
 *   f(m) = \sum_{k=0}^N \sum_{n=-k}^k \hat{f}(k,n) Y_k^n\left(2\pi x_1(m),
 *   2\pi x_2(m)\right).
 * \f]
 *
 * \arg plan the plan
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_adjoint(nfsft_plan* plan)
 * Executes an adjoint NFSFT, i.e. computes for \f$k=0,\ldots,N;
 * n=-k,\ldots,k\f$
 * \f[
 *   \hat{f}(k,n) = \sum_{m = 0}^{M-1} f(m) Y_k^n\left(2\pi x_1(m),
 *   2\pi x_2(m)\right).
 * \f]
 *
 * \arg plan the plan
 *
 * \author Jens Keiner
 */

/*! \fn void nfsft_finalize(nfsft_plan *plan)
 * Destroys a plan.
 *
 * \arg plan the plan to be destroyed
 *
 * \author Jens Keiner
 */

/*! \def NFSFT_NORMALIZED
 * By default, all computations are performed with respect to the
 * unnormalized basis functions
 * \f[
 *   \tilde{Y}_k^n(\vartheta,\varphi) = P_k^{|n|}(\cos\vartheta)
 *   \mathrm{e}^{\mathrm{i} n \varphi}.
 * \f]
 * If this flag is set, all computations are carried out using the \f$L_2\f$-
 * normalized basis functions
 * \f[
 *   Y_k^n(\vartheta,\varphi) = \sqrt{\frac{2k+1}{4\pi}} P_k^{|n|}(\cos\vartheta)
 *   \mathrm{e}^{\mathrm{i} n \varphi}.
 * \f]
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_USE_NDFT
 * If this flag is set, the fast NFSFT algorithms (see \ref nfsft_trafo,
 * \ref nfsft_adjoint) will use internally the exact but usually slower direct
 * NDFT algorithm in favor of fast but approximative NFFT algorithm.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_USE_DPT
 * If this flag is set, the fast NFSFT algorithms (see \ref nfsft_trafo,
 * \ref nfsft_adjoint) will use internally the usually slower direct
 * DPT algorithm in favor of the fast FPT algorithm.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_MALLOC_X
 * If this flag is set, the init methods (see \ref nfsft_init , \ref
 * nfsft_init_advanced , and \ref nfsft_init_guru) will allocate memory and the
 * method \ref nfsft_finalize will free the array \c x for you. Otherwise,
 * you have to assure by yourself that \c x points to an array of
 * proper size before excuting a transform and you are responsible for freeing
 * the corresponding memory before program termination.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_MALLOC_F_HAT
 * If this flag is set, the init methods (see \ref nfsft_init , \ref
 * nfsft_init_advanced , and \ref nfsft_init_guru) will allocate memory and the
 * method \ref nfsft_finalize will free the array \c f_hat for you. Otherwise,
 * you have to assure by yourself that \c f_hat points to an array of
 * proper size before excuting a transform and you are responsible for freeing
 * the corresponding memory before program termination.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_MALLOC_F
 * If this flag is set, the init methods (see \ref nfsft_init , \ref
 * nfsft_init_advanced , and \ref nfsft_init_guru) will allocate memory and the
 * method \ref nfsft_finalize will free the array \c f for you. Otherwise,
 * you have to assure by yourself that \c f points to an array of
 * proper size before excuting a transform and you are responsible for freeing
 * the corresponding memory before program termination.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_PRESERVE_F_HAT
 * If this flag is set, it is guaranteed that during an execution of
 * \ref nfsft_direct_trafo or \ref nfsft_trafo the content of \c f_hat remains
 * unchanged.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_PRESERVE_X
 * If this flag is set, it is guaranteed that during an execution of
 * \ref nfsft_direct_trafo, \ref nfsft_trafo or \ref nfsft_direct_adjoint, \ref nfsft_adjoint
 * the content of \c x remains
 * unchanged.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_PRESERVE_F
 * If this flag is set, it is guaranteed that during an execution of
 * \ref nfsft_direct_adjoint or \ref nfsft_adjoint the content of \c f remains
 * unchanged.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_DESTROY_F_HAT
 * If this flag is set, it is explicitely allowed that during an execution of
 * \ref nfsft_direct_trafo or \ref nfsft_trafo the content of \c f_hat may be changed.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_DESTROY_X
 * If this flag is set, it is explicitely allowed that during an execution of
 * \ref nfsft_direct_trafo, \ref nfsft_trafo or \ref nfsft_direct_adjoint, \ref nfsft_adjoint
 * the content of \c x may be changed.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_DESTROY_F
 * If this flag is set, it is explicitely allowed that during an execution of
 * \ref nfsft_direct_adjoint or \ref nfsft_adjoint the content of \c f may be changed.
 *
 * \see nfsft_init
 * \see nfsft_init_advanced
 * \see nfsft_init_guru
 * \author Jens Keiner
 */

/*! \def NFSFT_NO_DIRECT_ALGORITHM
 * If this flag is set, the transforms \ref nfsft_direct_trafo and \ref nfsft_direct_adjoint
 * do not work. Setting this flag saves some memory for precomputed data.
 *
 * \see nfsft_precompute
 * \see nfsft_direct_trafo
 * \see nfsft_direct_adjoint
 * \author Jens Keiner
 */

/*! \def NFSFT_NO_FAST_ALGORITHM
 * If this flag is set, the transforms \ref nfsft_trafo and \ref nfsft_adjoint
 * do not work. Setting this flag saves memory for precomputed data.
 *
 * \see nfsft_precompute
 * \see nfsft_trafo
 * \see nfsft_adjoint
 * \author Jens Keiner
 */

/*! \def NFSFT_ZERO_F_HAT
 * If this flag is set, the transforms \ref nfsft_adjoint and
 * \ref nfsft_direct_adjoint set all unused entries in \c f_hat not corresponding to
 * spherical Fourier coefficients to zero.
 *
 * \author Jens Keiner
 */

/*! \def NFSFT_EQUISPACED
 * If this flag is set, we use the equispaced FFT instead of the NFFT.
 * This implies that the nodes are fixed to 
 * \f[ \varphi_i = 2\pi \frac{i}{2N+2}, \qquad i=-N-1,\dots,N, \f]
 * \f[ \vartheta_j = 2\pi \frac{j}{2N+2}, \qquad j=0,\dots,N+1. \f]
 *
 * \author Michael Quellmalz
 */

/*! \def NFSFT_INDEX(k,n,plan)
 * This helper macro expands to the index \f$i\f$
 * corresponding to the spherical Fourier coefficient
 * \f$\mathtt{f\_hat}(k,n)\f$ for \f$0 \le k \le N\f$, \f$-k \le n \le k\f$ with
 * \f[
 *   (N+2)(N-n+1)+N+k+1
 * \f]
 */

/*! \def NFSFT_F_HAT_SIZE(N)
 * This helper macro expands to the logical size of a spherical Fourier coefficients
 * array for a bandwidth N.
 */

/** @}
 */