PySDKit 0.4.28__tar.gz → 0.4.29__tar.gz

{pysdkit-0.4.28 → pysdkit-0.4.29}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PySDKit
-Version: 0.4.28
+Version: 0.4.29
 Summary: A Python library for signal decomposition algorithms with a unified interface.
 Author: josefinez, Deeksha Manjunath, Yuan Feng, JacktheFowler
 Author-email: Whenxuan Wang <wwhenxuan@gmail.com>, RuiZhe Wang <3133986068@qq.com>, WenTong Zhao <23049200290@stu.xidian.edu.cn>
@@ -182,7 +182,7 @@ plot_IMFs_amplitude_spectra(IMFs, smooth="exp")   # use exp smooth
 |               [`ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/acmd.py) (Adaptive Chirp Mode Decomposition)               |                         [[paper]]()                          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/121373-data-driven-adaptive-chirp-mode-decomposition?s_tid=srchtitle) |    ✔️    |
 |    [`BA-ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/ba_acmd.py) (Bandwidth-aware adaptive chirp mode decomposition)    | [[paper]](https://journals.sagepub.com/doi/abs/10.1177/14759217231174699) | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/132792-bandwidth-aware-adaptive-chirp-mode-decomposition-ba-acmd?s_tid=srchtitle) |    ✖️    |
 |               [`JMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/jmd.py) (Jump Plus AM-FM Mode Decomposition)                |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169388-jump-plus-am-fm-mode-decomposition-jmd?s_tid=prof_contriblnk) |   ️✔️    |
-|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✖️    |
+|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✔️    |
 |                                            [`ESMD`]() (Extreme-Point Symmetric Mode Decomposition)                                            |          [[paper]](https://arxiv.org/abs/1303.6540)          |         [[code]](https://github.com/WuShichao/esmd)          |    ✖️    |
 |                                           [`STNBMD`]() (Short-Time Narrow-Band Mode Decomposition)                                            | [[paper]](https://www.sciencedirect.com/science/article/pii/S0022460X16002443?via%3Dihub) | [[code]](https://ww2.mathworks.cn/matlabcentral/fileexchange/56226-short-time-narrow-band-mode-decomposition-stnbmd-toolbox) |    ✖️    |
 | [`STL`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/tsa/_stl.py) (Seasonal-Trend decomposition using LOESS) | [[paper]](https://www.nniiem.ru/file/news/2016/stl-statistical-model.pdf) | [[code]](https://www.statsmodels.org/stable/examples/notebooks/generated/stl_decomposition.html) |    ✔️    |

{pysdkit-0.4.28 → pysdkit-0.4.29}/PySDKit.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PySDKit
-Version: 0.4.28
+Version: 0.4.29
 Summary: A Python library for signal decomposition algorithms with a unified interface.
 Author: josefinez, Deeksha Manjunath, Yuan Feng, JacktheFowler
 Author-email: Whenxuan Wang <wwhenxuan@gmail.com>, RuiZhe Wang <3133986068@qq.com>, WenTong Zhao <23049200290@stu.xidian.edu.cn>
@@ -182,7 +182,7 @@ plot_IMFs_amplitude_spectra(IMFs, smooth="exp")   # use exp smooth
 |               [`ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/acmd.py) (Adaptive Chirp Mode Decomposition)               |                         [[paper]]()                          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/121373-data-driven-adaptive-chirp-mode-decomposition?s_tid=srchtitle) |    ✔️    |
 |    [`BA-ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/ba_acmd.py) (Bandwidth-aware adaptive chirp mode decomposition)    | [[paper]](https://journals.sagepub.com/doi/abs/10.1177/14759217231174699) | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/132792-bandwidth-aware-adaptive-chirp-mode-decomposition-ba-acmd?s_tid=srchtitle) |    ✖️    |
 |               [`JMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/jmd.py) (Jump Plus AM-FM Mode Decomposition)                |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169388-jump-plus-am-fm-mode-decomposition-jmd?s_tid=prof_contriblnk) |   ️✔️    |
-|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✖️    |
+|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✔️    |
 |                                            [`ESMD`]() (Extreme-Point Symmetric Mode Decomposition)                                            |          [[paper]](https://arxiv.org/abs/1303.6540)          |         [[code]](https://github.com/WuShichao/esmd)          |    ✖️    |
 |                                           [`STNBMD`]() (Short-Time Narrow-Band Mode Decomposition)                                            | [[paper]](https://www.sciencedirect.com/science/article/pii/S0022460X16002443?via%3Dihub) | [[code]](https://ww2.mathworks.cn/matlabcentral/fileexchange/56226-short-time-narrow-band-mode-decomposition-stnbmd-toolbox) |    ✖️    |
 | [`STL`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/tsa/_stl.py) (Seasonal-Trend decomposition using LOESS) | [[paper]](https://www.nniiem.ru/file/news/2016/stl-statistical-model.pdf) | [[code]](https://www.statsmodels.org/stable/examples/notebooks/generated/stl_decomposition.html) |    ✔️    |

{pysdkit-0.4.28 → pysdkit-0.4.29}/PySDKit.egg-info/SOURCES.txt

@@ -48,7 +48,6 @@ pysdkit/_fmd/fmd.py
 pysdkit/_fmd/nfmd.py
 pysdkit/_hvd/__init__.py
 pysdkit/_hvd/hvd.py
-pysdkit/_ifd/__init__.py
 pysdkit/_itd/__init__.py
 pysdkit/_itd/itd.py
 pysdkit/_jmd/__init__.py
@@ -115,6 +114,7 @@ pysdkit/tests/test_faemd.py
 pysdkit/tests/test_faemd2d.py
 pysdkit/tests/test_faemd3d.py
 pysdkit/tests/test_itd.py
+pysdkit/tests/test_jmd.py
 pysdkit/tests/test_lmd.py
 pysdkit/tests/test_moving_decomp.py
 pysdkit/tests/test_ovmd.py

{pysdkit-0.4.28 → pysdkit-0.4.29}/README.md

@@ -128,7 +128,7 @@ plot_IMFs_amplitude_spectra(IMFs, smooth="exp")   # use exp smooth
 |               [`ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/acmd.py) (Adaptive Chirp Mode Decomposition)               |                         [[paper]]()                          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/121373-data-driven-adaptive-chirp-mode-decomposition?s_tid=srchtitle) |    ✔️    |
 |    [`BA-ACMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_vmd/ba_acmd.py) (Bandwidth-aware adaptive chirp mode decomposition)    | [[paper]](https://journals.sagepub.com/doi/abs/10.1177/14759217231174699) | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/132792-bandwidth-aware-adaptive-chirp-mode-decomposition-ba-acmd?s_tid=srchtitle) |    ✖️    |
 |               [`JMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/jmd.py) (Jump Plus AM-FM Mode Decomposition)                |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169388-jump-plus-am-fm-mode-decomposition-jmd?s_tid=prof_contriblnk) |   ️✔️    |
-|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✖️    |
+|        [`MJMD`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/_jmd/mjmd.py) (Multivariate Jump Plus AM-FM Mode Decomposition)        |         [[paper]](https://arxiv.org/abs/2407.07800)          | [[code]](https://www.mathworks.com/matlabcentral/fileexchange/169393-multivariate-jump-plus-am-fm-mode-decomposition-mjmd?s_tid=prof_contriblnk) |    ✔️    |
 |                                            [`ESMD`]() (Extreme-Point Symmetric Mode Decomposition)                                            |          [[paper]](https://arxiv.org/abs/1303.6540)          |         [[code]](https://github.com/WuShichao/esmd)          |    ✖️    |
 |                                           [`STNBMD`]() (Short-Time Narrow-Band Mode Decomposition)                                            | [[paper]](https://www.sciencedirect.com/science/article/pii/S0022460X16002443?via%3Dihub) | [[code]](https://ww2.mathworks.cn/matlabcentral/fileexchange/56226-short-time-narrow-band-mode-decomposition-stnbmd-toolbox) |    ✖️    |
 | [`STL`](https://github.com/wwhenxuan/PySDKit/blob/main/pysdkit/tsa/_stl.py) (Seasonal-Trend decomposition using LOESS) | [[paper]](https://www.nniiem.ru/file/news/2016/stl-statistical-model.pdf) | [[code]](https://www.statsmodels.org/stable/examples/notebooks/generated/stl_decomposition.html) |    ✔️    |

{pysdkit-0.4.28 → pysdkit-0.4.29}/pysdkit/__init__.py

@@ -2,7 +2,7 @@
 A Python library for signal decomposition algorithms.
 """
 
-__version__ = "0.4.28"
+__version__ = "0.4.29"
 
 # Empirical Mode Decomposition
 from ._emd import EMD
@@ -88,6 +88,9 @@ from ._ewt import ewt, EWT
 # Jump Plus AM-FM Mode Decomposition
 from ._jmd import JMD
 
+# Multivariate Jump Plus AM-FM Mode Decomposition
+from ._jmd import MJMD
+
 # Feature Mode Decomposition
 from ._fmd import FMD
 
@@ -151,6 +154,7 @@ Adaptive Local Iterative Filtering              |    ALIF
 Adaptive Polymorphic Mode Decomposition         |    APMD
 Empirical Wavelet Transform                     |    EWT
 Jump Plus AM-FM Mode Decomposition              |    JMD
+Multivariate Jump Plus AM-FM Mode Decomposition |    MJMD
 Feature Mode Decomposition                      |    FMD
 Moving Average Decomposition                    |    Moving
 Seasonal-Trend decomposition using LOESS        |    STL
@@ -191,6 +195,7 @@ __all__ = [
     "ewt",
     "EWT",
     "JMD",
+    "MJMD",
     "FMD",
     "Moving_Decomp",
     "STL",

{pysdkit-0.4.28 → pysdkit-0.4.29}/pysdkit/_hvd/hvd.py

@@ -173,7 +173,7 @@ class HVD(object):
             return imfs
 
 
-def make_window_real_valued(H, N):
+def make_window_real_valued(H: np.ndarray, N: int) -> np.ndarray:
     """
     Make window real valued
 

{pysdkit-0.4.28 → pysdkit-0.4.29}/pysdkit/_jmd/__init__.py

@@ -5,3 +5,6 @@ Created on 2025/02/05 13:32:00
 @email: wwhenxuan@gmail.com
 """
 from .jmd import JMD
+from .mjmd import MJMD
+
+__all__ = ["JMD", "MJMD"]

pysdkit-0.4.29/pysdkit/_jmd/jmd.py

@@ -0,0 +1,288 @@
+# -*- coding: utf-8 -*-
+"""
+Created on 2025/02/05 13:31:52
+@author: Whenxuan Wang
+@email: wwhenxuan@gmail.com
+
+Jump Plus AM-FM Mode Decomposition (JMD).
+
+Faithful port of the authors' MATLAB ``JMD.m``.
+"""
+from __future__ import annotations
+
+from typing import Optional, Tuple, Union
+
+import numpy as np
+from numpy import linalg
+import scipy.sparse as sp
+
+from pysdkit.utils import fft, fftshift, ifft, ifftshift
+
+
+class JMD(object):
+    """
+    Jump Plus AM-FM Mode Decomposition.
+
+    Decomposes a nonstationary signal into AM–FM oscillatory modes and a
+    discontinuous jump component by jointly solving a VMD-like bandwidth
+    minimization and a jump prior (JOT-style ADMM).
+
+    Mojtaba Nazari, Anders Rosendal Korshøj and Naveed ur Rehman,
+    "Jump Plus AM-FM Mode Decomposition," IEEE TSP,
+    https://doi.org/10.48550/arXiv.2407.07800
+    """
+
+    def __init__(
+        self,
+        K: int,
+        alpha: float = 5000.0,
+        init: str = "zero",
+        tol: float = 1e-6,
+        beta: float = 0.03,
+        b_bar: float = 0.45,
+        tau: float = 5.0,
+        max_iter: int = 2000,
+    ) -> None:
+        """
+        :param K: number of AM–FM modes to recover
+        :param alpha: balancing parameter of the mode bandwidth
+        :param init: ``'zero'`` / ``'uniform'`` / ``'random'`` omega initialization
+        :param tol: convergence tolerance
+        :param beta: jump-constraint weight (≈ 1 / expected number of jumps)
+        :param b_bar: parameter related to the jump prior
+        :param tau: jump ADMM step (used to form ``gamma``; set carefully for noise)
+        :param max_iter: maximum number of iterations
+        """
+        self.K = int(K)
+        self.alpha = float(alpha)
+        self.init = str(init).lower()
+        self.tol = float(tol)
+        self.beta = float(beta)
+        self.b_bar = float(b_bar)
+        self.tau = float(tau)
+        self.max_iter = int(max_iter)
+
+        self.u: Optional[np.ndarray] = None
+        self.v: Optional[np.ndarray] = None
+        self.omega: Optional[np.ndarray] = None
+
+    def __call__(
+        self, signal: np.ndarray, return_all: bool = False
+    ) -> Union[Tuple[np.ndarray, np.ndarray, np.ndarray], np.ndarray]:
+        return self.fit_transform(signal=signal, return_all=return_all)
+
+    def __str__(self) -> str:
+        return "Jump Plus AM-FM Mode Decomposition (JMD)"
+
+    def jump_step(self, freqs: np.ndarray, T: int) -> Tuple[
+        float,
+        np.ndarray,
+        np.ndarray,
+        np.ndarray,
+        np.ndarray,
+        np.ndarray,
+        np.ndarray,
+        np.ndarray,
+        float,
+        float,
+        float,
+    ]:
+        """Initialize jump-ADMM variables (matches MATLAB jump block)."""
+        b = 2.0 / (self.b_bar**2)
+        gamma = self.tau * (0.5 * b * self.beta)
+        v = np.zeros(T, dtype=float)
+        d = np.ones(T)
+        D = sp.diags([-d, d], [0, 1], shape=(T, T), format="csr").tolil()
+        D[-1, :] = 0
+        D = D.tocsr()
+        DTD = (D.T @ D).toarray()
+        x = np.zeros(T, dtype=float)
+        rho = np.zeros(T, dtype=float)
+        coef1 = 1.0 / gamma
+        mu = 2.0 * self.beta / gamma
+        SPDiag = sp.eye(T, format="csr")
+        j_hat_plus = np.zeros((self.max_iter, len(freqs)), dtype=complex)
+        return (
+            b,
+            v,
+            x,
+            D.toarray(),
+            DTD,
+            SPDiag.toarray(),
+            j_hat_plus,
+            rho,
+            coef1,
+            mu,
+            gamma,
+        )
+
+    @staticmethod
+    def enc_fmirror(signal: np.ndarray) -> np.ndarray:
+        """Mirror extension matching ``JMD.m`` / VMD style."""
+        signal = np.asarray(signal, dtype=float).ravel()
+        t = signal.size
+        half = t // 2
+        return np.concatenate([signal[half - 1 :: -1], signal, signal[: half - 1 : -1]])
+
+    @staticmethod
+    def dec_fmirror(u: np.ndarray, T: int) -> np.ndarray:
+        """Remove mirrored flanks."""
+        return u[:, T // 4 : 3 * T // 4].copy()
+
+    @staticmethod
+    def _soft_jump(h: np.ndarray, mu: float, b: float) -> np.ndarray:
+        """Element-wise jump proximal map from MATLAB ``min(max(...),1).*h``."""
+        abs_h = np.abs(h) + np.finfo(float).eps
+        scale = (1.0 / (1.0 - mu * b)) - (
+            mu * np.sqrt(2.0 * b) / (1.0 - mu * b)
+        ) / abs_h
+        scale = np.clip(scale, 0.0, 1.0)
+        return scale * h
+
+    def _init_omega(self, fs: float) -> np.ndarray:
+        """Initialization of omega_k (MATLAB ``init`` cases 0/1/2)."""
+        omega_plus = np.zeros((self.max_iter, self.K), dtype=float)
+        if self.init == "zero":
+            return omega_plus
+        if self.init == "uniform":
+            for i in range(self.K):
+                omega_plus[0, i] = (0.5 / self.K) * i
+            return omega_plus
+        if self.init == "random":
+            omega_plus[0, :] = np.sort(
+                np.exp(np.log(fs) + (np.log(0.5) - np.log(fs)) * np.random.rand(self.K))
+            )
+            return omega_plus
+        raise ValueError("init must be 'zero', 'uniform', or 'random'")
+
+    def fit_transform(
+        self, signal: np.ndarray, return_all: bool = False
+    ) -> Union[Tuple[np.ndarray, np.ndarray, np.ndarray], np.ndarray]:
+        """
+        Run JMD on a univariate signal.
+
+        :param signal: 1-D time-domain signal
+        :param return_all: if True, return ``(modes, jump, omega)``
+        :return: modes ``(K, N)`` or ``(modes, jump, omega)``
+        """
+        signal = np.asarray(signal, dtype=float).ravel()
+        if signal.size % 2 == 1:
+            signal = signal[:-1]
+        if signal.size < 4:
+            raise ValueError("signal length must be at least 4 samples")
+        if self.K < 1:
+            raise ValueError("K must be >= 1")
+
+        # MATLAB: demean, restore mean on the jump at the end
+        shift = float(np.mean(signal))
+        signal = signal - shift
+
+        f = self.enc_fmirror(signal)
+        T = f.size
+        half_T = T // 2
+        t = np.arange(1, T + 1, dtype=float) / T
+        freqs = t - 0.5 - 1.0 / T
+        fs = 1.0 / signal.size
+
+        a2 = 50.0
+        t2 = np.arange(0.01, np.sqrt(2.0 / a2) + 1e-12, 0.001)
+        phi1 = (-a2 / 2.0) * (t2**2) + (np.sqrt(2.0 * a2) * t2)
+        if self.max_iter > phi1.size:
+            phi = np.concatenate([phi1, np.ones(self.max_iter - phi1.size)])
+        else:
+            phi = phi1[: self.max_iter].copy()
+        Alpha = self.alpha * phi
+
+        f_hat = fftshift(fft(f))
+        f_hat_plus = f_hat.copy()
+        f_hat_plus[:half_T] = 0.0
+
+        u_hat_plus = np.zeros((self.max_iter, freqs.size, self.K), dtype=complex)
+        omega_plus = self._init_omega(fs=fs)
+
+        u_diff = self.tol + np.spacing(1)
+        n = 0
+        sum_uk = 0.0
+
+        b, v, x, D, DTD, SPDiag, j_hat_plus, rho, coef1, mu, gamma = self.jump_step(
+            freqs=freqs, T=T
+        )
+        A_jump = SPDiag + gamma * DTD
+
+        while u_diff > self.tol and n < self.max_iter - 1:
+            sum_uk = u_hat_plus[n, :, self.K - 1] + sum_uk - u_hat_plus[n, :, 0]
+
+            for k in range(self.K):
+                if k > 0:
+                    sum_uk = u_hat_plus[n + 1, :, k - 1] + sum_uk - u_hat_plus[n, :, k]
+                u_hat_plus[n + 1, :, k] = (f_hat_plus - sum_uk - j_hat_plus[n, :]) / (
+                    1.0 + Alpha[n] * (freqs - omega_plus[n, k]) ** 2
+                )
+
+                num = np.dot(
+                    freqs[half_T:T], np.abs(u_hat_plus[n + 1, half_T:T, k]) ** 2
+                )
+                den = np.sum(np.abs(u_hat_plus[n + 1, half_T:T, k]) ** 2)
+                omega_plus[n + 1, k] = num / (den + np.finfo(float).eps)
+
+            # Back to time domain for jump update
+            u_hat = np.zeros((T, self.K), dtype=complex)
+            for k in range(self.K):
+                u_hat[half_T:T, k] = u_hat_plus[n + 1, half_T:T, k]
+                u_hat[half_T:0:-1, k] = np.conj(u_hat_plus[n + 1, half_T:T, k])
+                u_hat[0, k] = np.conj(u_hat[-1, k])
+
+            u = np.zeros((self.K, T), dtype=float)
+            for k in range(self.K):
+                u[k, :] = np.real(ifft(ifftshift(u_hat[:, k])))
+
+            rhs = gamma * (D.T @ x) - (D.T @ rho) + f - np.sum(u, axis=0)
+            v = linalg.solve(A_jump, rhs)
+
+            dv = D @ v
+            h = dv + coef1 * rho
+            x = self._soft_jump(h, mu=mu, b=b)
+            rho = rho - gamma * (x - dv)
+
+            v = v - (np.mean(v) - np.mean(f))
+
+            j_hat_plus[n + 1, :] = fftshift(fft(v))
+            j_hat_plus[n + 1, :half_T] = 0.0
+
+            n += 1
+            u_diff = np.spacing(1)
+            for i in range(self.K):
+                diff = u_hat_plus[n, :, i] - u_hat_plus[n - 1, :, i]
+                u_diff += (1.0 / T) * np.vdot(diff, diff).real
+            jdiff = j_hat_plus[n, :] - j_hat_plus[n - 1, :]
+            u_diff += (1.0 / T) * np.vdot(jdiff, jdiff).real
+            u_diff = abs(u_diff)
+
+        n_keep = min(n, self.max_iter - 1)
+        omega_hist = omega_plus[: n_keep + 1, :]
+
+        u_hat = np.zeros((T, self.K), dtype=complex)
+        for k in range(self.K):
+            u_hat[half_T:T, k] = u_hat_plus[n_keep, half_T:T, k]
+            u_hat[half_T:0:-1, k] = np.conj(u_hat_plus[n_keep, half_T:T, k])
+            u_hat[0, k] = np.conj(u_hat[-1, k])
+
+        u = np.zeros((self.K, T), dtype=float)
+        for k in range(self.K):
+            u[k, :] = np.real(ifft(ifftshift(u_hat[:, k])))
+
+        u = self.dec_fmirror(u, T=T)
+        v = v[T // 4 : 3 * T // 4] + shift
+
+        # Sort modes by ascending center frequency (MATLAB cleanup)
+        order = np.argsort(omega_hist[-1, :])
+        u = u[order, :]
+        omega = omega_hist[:, order]
+
+        self.u = u
+        self.v = v
+        self.omega = omega
+
+        if return_all:
+            return u, v, omega[-1]
+        return u