freealg 0.0.2__py3-none-any.whl → 0.1.0__py3-none-any.whl

freealg/freeform.py

@@ -12,15 +12,17 @@
 # =======
 
 import numpy
+from scipy.stats import gaussian_kde
 from functools import partial
 from ._util import compute_eig, force_density
 from ._jacobi import jacobi_proj, jacobi_approx, jacobi_stieltjes
 from ._chebyshev import chebyshev_proj, chebyshev_approx, chebyshev_stieltjes
 from ._damp import jackson_damping, lanczos_damping, fejer_damping, \
     exponential_damping, parzen_damping
-from ._plot_util import plot_coeff, plot_density, plot_hilbert, \
-    plot_stieltjes, plot_glue_fit
+from ._plot_util import plot_fit, plot_density, plot_hilbert, plot_stieltjes
 from ._pade import fit_pade, eval_pade
+from ._decompress import decompress
+from ._sample import qmc_sample
 
 __all__ = ['FreeForm']
 
@@ -45,6 +47,13 @@ class FreeForm(object):
         The support of the density of :math:`\\mathbf{A}`. If `None`, it is
         estimated from the minimum and maximum of the eigenvalues.
 
+    p : float, default=0.001
+        The edges of the support of the distribution is detected by the
+        :math:`p`-quantile on the left and :math:`(1-p)`-quantile on the right.
+        If the argument ``support`` is directly provided, this option is
+        ignored. This value should be between 0 and 1, ideally a small
+        number close to zero.
+
     Notes
     -----
 
@@ -64,6 +73,10 @@ class FreeForm(object):
     psi : numpy.array
         Jacobi coefficients.
 
+    n   : int
+        Initial array size (assuming a square matrix when :math:`\\mathbf{A}`
+        is 2D)
+
     Methods
     -------
 
@@ -94,7 +107,7 @@ class FreeForm(object):
     # init
     # ====
 
-    def __init__(self, A, support=None):
+    def __init__(self, A, support=None, p=0.001):
         """
         Initialization.
         """
@@ -106,16 +119,20 @@ class FreeForm(object):
         if A.ndim == 1:
             # When A is a 1D array, it is assumed A is the eigenvalue array.
             self.eig = A
+            self.n = len(A)
         elif A.ndim == 2:
             # When A is a 2D array, it is assumed A is the actual array,
             # and its eigenvalues will be computed.
             self.A = A
+            self.n = A.shape[0]
+            assert A.shape[0] == A.shape[1], \
+                'Only square matrices are permitted.'
             self.eig = compute_eig(A)
 
         # Support
         if support is None:
-            self.lam_m = self.eig.min()
-            self.lam_p = self.eig.max()
+            self.lam_m, self.lam_p = self._detect_support(self.eig, p,
+                                                          smoothen=True)
         else:
             self.lam_m = support[0]
             self.lam_p = support[1]
@@ -126,20 +143,46 @@ class FreeForm(object):
         self.psi = None
         self.alpha = None
         self.beta = None
+        self._pade_sol = None
+
+    # ==============
+    # detect support
+    # ==============
+
+    def _detect_support(self, eig, p, smoothen=True):
+        """
+        """
+
+        # Using quantile directly.
+        if smoothen:
+            kde = gaussian_kde(eig)
+            xs = numpy.linspace(eig.min(), eig.max(), 1000)
+            fs = kde(xs)
+
+            cdf = numpy.cumsum(fs)
+            cdf /= cdf[-1]
+
+            lam_m = numpy.interp(p, cdf, xs)
+            lam_p = numpy.interp(1-p, cdf, xs)
+        else:
+            lam_m, lam_p = numpy.quantile(eig, [p, 1-p])
+
+        return lam_m, lam_p
 
     # ===
     # fit
     # ===
 
     def fit(self, method='jacobi', K=10, alpha=0.0, beta=0.0, reg=0.0,
-            damp=None, force=False, plot=False, latex=False, save=False):
+            damp=None, force=False, pade_p=0, pade_q=1, odd_side='left',
+            optimizer='ls', plot=False, latex=False, save=False):
         """
         Fit model to eigenvalues.
 
         Parameters
         ----------
 
-        method : {``'jacobi'``, ``'chebyshev'``}, default=``'jacobi'``
+        method : {``'jacobi'``, ``'chebyshev'``}, default= ``'jacobi'``
             Method of approximation, either by Jacobi polynomials or Chebyshev
             polynomials of the second kind.
 
@@ -167,8 +210,30 @@ class FreeForm(object):
             If `True`, it forces the density to have unit mass and to be
             strictly positive.
 
+        pade_p : int, default=0
+            Degree of polynomial :math:`P(z)` is :math:`q+p` where :math:`p`
+            can only be ``-1``, ``0``, or ``1``. See notes below.
+
+        pade_q : int, default=1
+            Degree of polynomial :math:`Q(z)` is :math:`q`. See notes below.
+
+        odd_side : {``'left'``, ``'right'``}, default= ``'left'``
+            In case of odd number of poles (when :math:`q` is odd), the extra
+            pole is set to the left or right side of the support interval,
+            while all other poles are split in half to the left and right. Note
+            that this is only for the initialization of the poles. The
+            optimizer will decide best location by moving them to the left or
+            right of the support.
+
+        optimizer : {``'ls'``, ``'de'``}, default= ``'ls'``
+            Optimizer for Pade approximation, including:
+
+            * ``'ls'``: least square (local, fast)
+            * ``'de'``: differential evolution (global, slow)
+
         plot : bool, default=False
-            If `True`, density is plotted.
+            If `True`, the approximation coefficients and Pade approximation to
+            the Hilbert transform are plotted.
 
         latex : bool, default=False
             If `True`, the plot is rendered using LaTeX. This option is
@@ -185,6 +250,20 @@ class FreeForm(object):
         psi : (K+1, ) numpy.ndarray
             Coefficients of fitting Jacobi polynomials
 
+        Notes
+        -----
+
+        The Pade approximation for the glue function :math:`G(z)` is
+
+        .. math::
+
+            G(z) = \\frac{P(z)}{Q(z)},
+
+        where :math:`P(z)` and :math:`Q(z)` are polynomials of order
+        :math:`p+q` and :math:`q` respectively. Note that :math:`p` can only
+        be -1, 0, or 1, effectively making Pade approximation of order
+        :math:`q-1:q`, :math:`q:q`, or :math:`q-1:q`.
+
         Examples
         --------
 
@@ -245,8 +324,23 @@ class FreeForm(object):
         self.alpha = alpha
         self.beta = beta
 
+        # For holomorphic continuation for the lower half-plane
+        x_supp = numpy.linspace(self.lam_m, self.lam_p, 1000)
+        g_supp = 2.0 * numpy.pi * self.hilbert(x_supp)
+
+        # Fit a pade approximation
+        # self._pade_sol = fit_pade(x_supp, g_supp, self.lam_m,
+        #                           self.lam_p, pade_p, pade_q, delta=1e-8,
+        #                           B=numpy.inf, S=numpy.inf)
+        self._pade_sol = fit_pade(x_supp, g_supp, self.lam_m, self.lam_p,
+                                  p=pade_p, q=pade_q, odd_side=odd_side,
+                                  safety=1.0, max_outer=40, xtol=1e-12,
+                                  ftol=1e-12, optimizer=optimizer, verbose=0)
+
         if plot:
-            plot_coeff(psi, latex=latex, save=save)
+            g_supp_approx = eval_pade(x_supp[None, :], self._pade_sol)[0, :]
+            plot_fit(psi, x_supp, g_supp, g_supp_approx, support=self.support,
+                     latex=latex, save=save)
 
         return self.psi
 
@@ -444,30 +538,12 @@ class FreeForm(object):
     # glue
     # ====
 
-    def _glue(self, z, p, q, plot_glue=False, latex=False, save=False):
+    def _glue(self, z):
         """
         """
 
-        # Holomorphic continuation for the lower half-plane
-        x_supp = numpy.linspace(self.lam_m, self.lam_p, 1000)
-        g_supp = 2.0 * numpy.pi * self.hilbert(x_supp)
-
-        # Fit a pade approximation
-        sol = fit_pade(x_supp, g_supp, self.lam_m, self.lam_p, p, q,
-                       delta=1e-8, B=numpy.inf, S=numpy.inf)
-
-        # Unpack optimized parameters
-        s = sol['s']
-        a = sol['a']
-        b = sol['b']
-
         # Glue function
-        g = eval_pade(z, s, a, b)
-
-        if plot_glue:
-            g_supp_approx = eval_pade(x_supp[None, :], s, a, b)[0, :]
-            plot_glue_fit(x_supp, g_supp, g_supp_approx, support=self.support,
-                          latex=latex, save=save)
+        g = eval_pade(z, self._pade_sol)
 
         return g
 
@@ -475,10 +551,10 @@ class FreeForm(object):
     # stieltjes
     # =========
 
-    def stieltjes(self, x, y, p=1, q=2, plot=False, plot_glue=False,
-                  latex=False, save=False):
+    def stieltjes(self, x, y, plot=False, latex=False, save=False):
         """
-        Compute Stieltjes transform of the spectral density.
+        Compute Stieltjes transform of the spectral density over a 2D Cartesian
+        grid on the complex plane.
 
         Parameters
         ----------
@@ -492,19 +568,9 @@ class FreeForm(object):
             The y axis of the grid where the Stieltjes transform is evaluated.
             If `None`, a grid on the interval ``[-1, 1]`` is used.
 
-        p : int, default=1
-            Degree of polynomial :math:`P(z)`. See notes below.
-
-        q : int, default=1
-            Degree of polynomial :math:`Q(z)`. See notes below.
-
         plot : bool, default=False
             If `True`, density is plotted.
 
-        plot_glue : bool, default=False
-            If `True`, the fit of glue function to Hilbert transform is
-            plotted.
-
         latex : bool, default=False
             If `True`, the plot is rendered using LaTeX. This option is
             relevant only if ``plot=True``.
@@ -536,7 +602,7 @@ class FreeForm(object):
         References
         ----------
 
-        .. [1] rbd
+        .. [1] tbd
 
         Examples
         --------
@@ -591,35 +657,158 @@ class FreeForm(object):
         m1[mask_m, :] = numpy.conjugate(
             stieltjes(numpy.conjugate(z[mask_m, :])))
 
+        # Second Riemann sheet
         m2[mask_p, :] = m1[mask_p, :]
-        m2[mask_m, :] = -m1[mask_m, :] + self._glue(
-                z[mask_m, :], p, q, plot_glue=plot_glue, latex=latex,
-                save=save)
+        m2[mask_m, :] = -m1[mask_m, :] + self._glue(z[mask_m, :])
 
         if plot:
             plot_stieltjes(x, y, m1, m2, self.support, latex=latex, save=save)
 
         return m1, m2
 
+    # ==============
+    # eval stieltjes
+    # ==============
+
+    def _eval_stieltjes(self, z):
+        """
+        Compute Stieltjes transform of the spectral density.
+
+        Parameters
+        ----------
+
+        z : numpy.array
+            The z values in the complex plan where the Stieltjes transform is
+            evaluated.
+
+
+        Returns
+        -------
+
+        m_p : numpy.ndarray
+            The Stieltjes transform on the principal branch.
+
+        m_m : numpy.ndarray
+            The Stieltjes transform continued to the secondary branch.
+
+        See Also
+        --------
+        density
+        hilbert
+
+        Notes
+        -----
+
+        Notes.
+
+        References
+        ----------
+
+        .. [1] tbd
+
+        Examples
+        --------
+
+        .. code-block:: python
+
+            >>> from freealg import FreeForm
+        """
+
+        if self.psi is None:
+            raise RuntimeError('"fit" the model first.')
+
+        z = numpy.asarray(z)
+        shape = z.shape
+        if len(shape) == 0:
+            shape = (1,)
+        z = z.reshape(-1, 1)
+
+        # Stieltjes function
+        if self.method == 'jacobi':
+            stieltjes = partial(jacobi_stieltjes, psi=self.psi,
+                                support=self.support, alpha=self.alpha,
+                                beta=self.beta)
+        elif self.method == 'chebyshev':
+            stieltjes = partial(chebyshev_stieltjes, psi=self.psi,
+                                support=self.support)
+
+        mask_p = z.imag >= 0.0
+        mask_m = z.imag < 0.0
+
+        m1 = numpy.zeros_like(z)
+        m2 = numpy.zeros_like(z)
+
+        # Upper half-plane
+        m1[mask_p] = stieltjes(z[mask_p].reshape(-1, 1)).reshape(-1)
+
+        # Lower half-plane, use Schwarz reflection
+        m1[mask_m] = numpy.conjugate(
+            stieltjes(numpy.conjugate(z[mask_m].reshape(-1, 1)))).reshape(-1)
+
+        # Second Riemann sheet
+        m2[mask_p] = m1[mask_p]
+        m2[mask_m] = -m1[mask_m] + self._glue(
+            z[mask_m].reshape(-1, 1)).reshape(-1)
+
+        m1, m2 = m1.reshape(*shape), m2.reshape(*shape)
+
+        return m1, m2
+
     # ==========
     # decompress
     # ==========
 
-    def decompress(self, n):
+    def decompress(self, size, x=None, delta=1e-6, iterations=500,
+                   step_size=0.1, tolerance=1e-4, plot=False, latex=False,
+                   save=False):
         """
         Free decompression of spectral density.
 
         Parameters
         ----------
 
-        n : int
-            Size of the matrix.
+        size : int
+            Size of the decompressed matrix.
+
+        x : numpy.array, default=None
+            Positions where density to be evaluated at. If `None`, an interval
+            slightly larger than the support interval will be used.
+
+        delta: float, default=1e-4
+            Size of the perturbation into the upper half plane for Plemelj's
+            formula.
+
+        iterations: int, default=500
+            Maximum number of Newton iterations.
+
+        step_size: float, default=0.1
+            Step size for Newton iterations.
+
+        tolerance: float, default=1e-4
+            Tolerance for the solution obtained by the Newton solver. Also
+            used for the finite difference approximation to the derivative.
+
+        plot : bool, default=False
+            If `True`, density is plotted.
+
+        latex : bool, default=False
+            If `True`, the plot is rendered using LaTeX. This option is
+            relevant only if ``plot=True``.
+
+        save : bool, default=False
+            If not `False`, the plot is saved. If a string is given, it is
+            assumed to the save filename (with the file extension). This option
+            is relevant only if ``plot=True``.
 
         Returns
         -------
 
         rho : numpy.array
-            Spenctral density
+            Spectral density
+
+        eigs : numpy.array
+            Estimated eigenvalues as low-discrepancy samples of the estimated
+            spectral density.
 
         See Also
         --------
@@ -630,7 +819,7 @@ class FreeForm(object):
         Notes
         -----
 
-        Not implemented.
+        Work in progress.
 
         References
         ----------
@@ -645,4 +834,15 @@ class FreeForm(object):
             >>> from freealg import FreeForm
         """
 
-        pass
+        rho, x, (lb, ub) = decompress(self, size, x=x, delta=delta,
+                                      iterations=iterations,
+                                      step_size=step_size, tolerance=tolerance)
+        x, rho = x.ravel(), rho.ravel()
+
+        if plot:
+            plot_density(x, rho, support=(lb, ub),
+                         label='Decompression', latex=latex, save=save)
+
+        eigs = numpy.sort(qmc_sample(x, rho, size))
+
+        return rho, eigs

{freealg-0.0.2.dist-info → freealg-0.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: freealg
-Version: 0.0.2
+Version: 0.1.0
 Summary: Free probability for large matrices
 Keywords: leaderboard bot chat
 Platform: Linux
@@ -29,6 +29,7 @@ Requires-Dist: scipy
 Requires-Dist: texplot
 Requires-Dist: matplotlib
 Requires-Dist: colorcet
+Requires-Dist: networkx
 Provides-Extra: test
 Provides-Extra: docs
 Dynamic: classifier

freealg-0.1.0.dist-info/RECORD

@@ -0,0 +1,21 @@
+freealg/__init__.py,sha256=K92neXJZ9VE1U_j_pj28Qyq1MzlMXhOuYK2ZihgwCaU,463
+freealg/__version__.py,sha256=kUR5RAFc7HCeiqdlX36dZOHkUI5wI6V_43RpEcD8b-0,22
+freealg/_chebyshev.py,sha256=Cw48gXF6kd3IAQuLTEWadySeKGnY9TynzX9MNmycMUU,4697
+freealg/_damp.py,sha256=k2vtBtWOxQBf4qXaWu_En81lQBXbEO4QbxxWpvuVhdE,1802
+freealg/_decompress.py,sha256=H7ocq09gQnCY-q_8xHi6qyYj3qp239MCgj406hn0yeE,3344
+freealg/_jacobi.py,sha256=HVnaujwAcaILVFGEgxk26UyIaLdP2FZMY44C8TG3Qcc,4763
+freealg/_pade.py,sha256=wjFH1lfhuaQANNNMLVa8eAlKbP3sz8l8ZMw-P2X3T4A,12867
+freealg/_plot_util.py,sha256=aL0u7FHdCwLTj4dZ4SSGA00wab0Voem2nBAsBVvo6XY,18400
+freealg/_sample.py,sha256=K1ZxKoiuPbEKyh-swL5X7gz1kYcQno6Mof0o1xF38tg,2323
+freealg/_util.py,sha256=wJ-t8LMZZFEr2PsZEVqTJP_jQTQ3rHUf0dO27F6L7YQ,2310
+freealg/freeform.py,sha256=nZqGZBEHVT7Cx1UbOe2-cupWBANNdmc44vm9Eyy9I0Q,25390
+freealg/distributions/__init__.py,sha256=Hnk9bJi4Wy8I_1uuskRyrT2DUpPN1YmBY5uK7XI3U_o,644
+freealg/distributions/kesten_mckay.py,sha256=SFYg2_6_MEX9NGfOIW_rRfM9kgF-bdjQiVS0ENJBemQ,15379
+freealg/distributions/marchenko_pastur.py,sha256=0XHhjw1ZDkigGfjU9lmPRgmwecM1-n6VBkGjRJbxpeY,15869
+freealg/distributions/wachter.py,sha256=cG-3XTP5CHSK5o6SQfa3lqQR__ZibeWZ9AoeMxr3Nnk,15602
+freealg/distributions/wigner.py,sha256=gj3XWK4X7wLsB-inoVpn4pHUUM7NOQdgMGw01LEqkcw,14914
+freealg-0.1.0.dist-info/licenses/LICENSE.txt,sha256=J-EEYEtxb3VVf_Bn1TYfWnpY5lMFIM15iLDDcnaDTPA,1443
+freealg-0.1.0.dist-info/METADATA,sha256=hsJ1JS0j2HpMLsplIIdHlQeqOzMZamBkAVRL5xc1pYQ,2912
+freealg-0.1.0.dist-info/WHEEL,sha256=DnLRTWE75wApRYVsjgc6wsVswC54sMSJhAEd4xhDpBk,91
+freealg-0.1.0.dist-info/top_level.txt,sha256=eR2wrgYwDdnnJ9Zf5PruPqe4kQav0GMvRsqct6y00Q8,8
+freealg-0.1.0.dist-info/RECORD,,

freealg-0.0.2.dist-info/RECORD

@@ -1,16 +0,0 @@
-freealg/__init__.py,sha256=K92neXJZ9VE1U_j_pj28Qyq1MzlMXhOuYK2ZihgwCaU,463
-freealg/__version__.py,sha256=QvlVh4JTl3JL7jQAja76yKtT-IvF4631ASjWY1wS6AQ,22
-freealg/_chebyshev.py,sha256=Cw48gXF6kd3IAQuLTEWadySeKGnY9TynzX9MNmycMUU,4697
-freealg/_damp.py,sha256=k2vtBtWOxQBf4qXaWu_En81lQBXbEO4QbxxWpvuVhdE,1802
-freealg/_jacobi.py,sha256=F9K0IzbExpxGQY2OJQb-XT4Ee1zWTZvMXRUdU8cbwSE,4916
-freealg/_pade.py,sha256=rWILLpEL910YdixC2c5Cw77HqogAYoxi2YZgRVXEELM,4022
-freealg/_plot_util.py,sha256=OILqOqKdLGEabk41KnmxNftItzMMFHOL-LZ35yNvdyk,18653
-freealg/_util.py,sha256=wJ-t8LMZZFEr2PsZEVqTJP_jQTQ3rHUf0dO27F6L7YQ,2310
-freealg/freeform.py,sha256=LLkGoPLxvn8VMHuVQ1LGNkcAYZUi3GduFB8qlsr69qg,18595
-freealg/distributions/__init__.py,sha256=7t4HbP_EofiFDYLH6jbD94AIumOdcHn1y_Qo54mpLFM,614
-freealg/distributions/marchenko_pastur.py,sha256=k8SoEgB2DFXLqGF-Hyqm8Sfh1DlF3khMssuxak5Uaqw,15827
-freealg-0.0.2.dist-info/licenses/LICENSE.txt,sha256=J-EEYEtxb3VVf_Bn1TYfWnpY5lMFIM15iLDDcnaDTPA,1443
-freealg-0.0.2.dist-info/METADATA,sha256=lMjLvkUpPUAr_QPthjnrIfg9KTs8DLEpTG68w8Ak0rU,2888
-freealg-0.0.2.dist-info/WHEEL,sha256=DnLRTWE75wApRYVsjgc6wsVswC54sMSJhAEd4xhDpBk,91
-freealg-0.0.2.dist-info/top_level.txt,sha256=eR2wrgYwDdnnJ9Zf5PruPqe4kQav0GMvRsqct6y00Q8,8
-freealg-0.0.2.dist-info/RECORD,,