Merge pull request #14 from fjprichard/dev

Dev

Author: fjprichard

PyAFBF.egg-info/PKG-INFO

@@ -1,74 +1,117 @@
-Metadata-Version: 1.2
-Name: PyAFBF
-Version: 0.0.3
-Summary: Sample image textures from anisotropic fractional Brownian fields
-Home-page: https://github.com/fjprichard/PyAFBF/
-Author: Frederic Richard
-Author-email: frederic.richard@univ-amu.fr
-License: GPLv3
-Project-URL: Documentation, https://fjprichard.github.io/PyAFBF/
-Project-URL: Source, https://github.com/fjprichard/PyAFBF/
-Project-URL: Tracker, https://github.com/fjprichard/PyAFBF/issues
-Description: The Package PyAFBF is intended for the simulation of rough anisotropic image textures. Textures are sampled from a mathematical model called the anisotropic fractional Brownian field. More details can be found on the `documentation <https://fjprichard.github.io/PyAFBF/>`_.
-        
-        Package features
-        ================
-        
-        - Simulation of rough anisotropic textures,
-        
-        - Computation of field features (semi-variogram, regularity, anisotropy indices) that can serve as texture attributes,
-        
-        - Random definition of simulated fields,
-        
-        - Extensions to related fields (deformed fields, intrinsic fields, heterogeneous fields, binary patterns).
-        
-        
-        Installation from sources
-        =========================
-        
-        The package source can be downloaded from the `repository <https://github.com/fjprichard/PyAFBF>`_. 
-        
-        To install the package, write the following commands in the root directory where the package was downloaded:
-        
-            python setup.py install
-            or
-            pip install -e .
-        
-        After installation, the package can be tested by running 
-        
-            python tests/TestAll.py
-        
-        If no output appears, then the package is correctly installed.
-        
-        Communication to the author
-        ===========================
-        
-        PyAFBF is developed and maintained by Frédéric Richard. For feed-back, contributions, bug reports, contact directly the `author <https://github.com/fjprichard>`_, or use the `discussion <https://github.com/fjprichard/PyAFBF/discussions>`_ facility.
-        
-        
-        Licence
-        =======
-        
-        PyAFBF is under licence GNU GPL, version 3.
-        
-        
-        Credits
-        =======
-        
-        PyAFBF is written and maintained by Frederic Richard, Professor at Aix-Marseille University, France.
-        
-        
-        
-Keywords: simulation,anisotropic fractional Brownian field,image texture
-Platform: Linux
-Platform: MacOSX
-Platform: Windows
-Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
-Classifier: Development Status :: 5 - Production/Stable
-Classifier: Intended Audience :: Science/Research
-Classifier: Intended Audience :: End Users/Desktop
-Classifier: Intended Audience :: Developers
-Classifier: Natural Language :: English
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3.8
-Classifier: Topic :: Scientific/Engineering :: Image Processing
+Metadata-Version: 2.1
+Name: PyAFBF
+Version: 0.2.0
+Summary: Sample image textures from anisotropic fractional Brownian fields
+Home-page: https://github.com/fjprichard/PyAFBF/
+Author: Frederic Richard
+Author-email: frederic.richard@univ-amu.fr
+License: GPLv3
+Project-URL: Documentation, https://fjprichard.github.io/PyAFBF/
+Project-URL: Source, https://github.com/fjprichard/PyAFBF/
+Project-URL: Tracker, https://github.com/fjprichard/PyAFBF/issues
+Description: .. image:: https://zenodo.org/badge/368267301.svg
+           :target: https://zenodo.org/badge/latestdoi/368267301
+        
+        The Package PyAFBF is intended for the simulation of rough anisotropic image textures. Textures are sampled from a mathematical model called the anisotropic fractional Brownian field. More details can be found on the `documentation <https://fjprichard.github.io/PyAFBF/>`_.
+        
+        Package features
+        ================
+        
+        - Simulation of rough anisotropic textures,
+        
+        - Computation of field features (semi-variogram, regularity, anisotropy indices) that can serve as texture attributes,
+        
+        - Random definition of simulated fields,
+        
+        - Extensions to related fields (deformed fields, intrinsic fields, heterogeneous fields, binary patterns).
+        
+        
+        Installation from sources
+        =========================
+        
+        The package source can be downloaded from the `repository <https://github.com/fjprichard/PyAFBF>`_. 
+        
+        The package can be installed through PYPI with
+         
+         pip install PyAFBF
+         
+        To install the package in a Google Collab environment, please type
+        
+         !pip install imgaug==0.2.6
+         
+         !pip install PyAFBF
+        
+        Communication to the author
+        ===========================
+        
+        PyAFBF is developed and maintained by Frédéric Richard. For feed-back, contributions, bug reports, contact directly the `author <https://github.com/fjprichard>`_, or use the `discussion <https://github.com/fjprichard/PyAFBF/discussions>`_ facility.
+        
+        
+        Licence
+        =======
+        
+        PyAFBF is under licence GNU GPL, version 3.
+        
+        
+        Citation
+        ========
+        
+        When using PyAFBF, please cite the original paper
+        
+        H. Biermé, M. Moisan, and F. Richard. A turning-band method for the simulation of anisotropic fractional Brownian field. J. Comput. Graph. Statist., 24(3):885–904, 2015.
+        
+        and the JOSS paper:
+        
+        
+        .. image:: https://joss.theoj.org/papers/10.21105/joss.03821/status.svg
+           :target: https://doi.org/10.21105/joss.03821
+        
+        
+        Contents
+        ========
+        
+            - Quick start guide
+               - Getting started
+               - Customed models
+               - Tuning model parameters
+               - Model features
+               - Simulating with turning-band fields
+            - Example gallery
+               - Basic examples
+               - Extended anisotropic fields
+               - Heterogeneous fields
+               - Related anisotropic fields
+            - API: main classes
+               - AFBF (field)
+               - Turning band field (tbfield)
+            - API: auxiliary classes
+               - Periodic functions (perfunction)
+               - Coordinates (coordinates)
+               - Spatial data (sdata)
+               - Process (process)
+               - Turning bands (tbparameters)
+               - ndarray
+        
+        
+        
+        Credits
+        =======
+        
+        PyAFBF is written and maintained by Frederic Richard, Professor at Aix-Marseille University, France.
+        
+        
+        
+Keywords: anisotropic fractional Brownian field,image texture,simulation
+Platform: Linux
+Platform: MacOSX
+Platform: Windows
+Classifier: License :: OSI Approved :: GNU General Public License v3 (GPLv3)
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Science/Research
+Classifier: Intended Audience :: End Users/Desktop
+Classifier: Intended Audience :: Developers
+Classifier: Natural Language :: English
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Topic :: Scientific/Engineering :: Image Processing
+Provides-Extra: dev

PyAFBF.egg-info/requires.txt

@@ -1,3 +1,6 @@
 numpy>=1.19.2
 matplotlib>=3.3.2
 scipy>=1.5.2
+
+[dev]
+doctest

afbf/Classes/Field.py

@@ -43,7 +43,7 @@
 
 """
 from afbf.utilities import pi, linspace, unique, concatenate, power
-from afbf.utilities import ceil, amin, amax, mean, reshape, nonzero
+from afbf.utilities import ceil, amin, amax, mean, reshape, nonzero, argmin
 from afbf.utilities import sum, zeros, array
 from afbf.Classes.PeriodicFunction import DiscreteFunctionDescription
 from afbf import sdata, coordinates, perfunction
@@ -233,7 +233,8 @@ def NormalizeModel(self):
             topo = perfunction(self.hurst.ftype, self.hurst.fparam.size)
             self.topo = topo
             topo.fname = 'Topothesy function'
-            topo.finter = self.hurst.finter
+            topo.finter = zeros(self.hurst.finter.shape)
+            topo.finter[:] = self.hurst.finter[:]
             topo.fparam = zeros(self.hurst.fparam.shape)
             if self.hurst.steptrans:
                 topo.trans = self.hurst.trans
@@ -391,7 +392,7 @@ def ChangeOrder(self, neworder):
     def ComputeFeatures(self):
         """Compute several features of the field.
 
-        :returns: Attributes H, hurst_argmin_length, hurst_index_aniso,
+        :returns: Attributes H, Hmax, hurst_argmin_length, hurst_index_aniso,
             aniso_indices_topo, aniso_indices_hurst, aniso_sharpness_topo,
             aniso_sharpness_hurst, aniso_indices_mixed1, aniso_indices_mixed2,
             aniso_sharpness_mixed1, aniso_sharpness_mixed2.
@@ -423,13 +424,6 @@ def ComputeFeatures(self):
         self.aniso_sharpness_mixed_1 = sh1
         self.aniso_sharpness_mixed_2 = sh2
 
-        # Analysis of the Hurst function.
-        # Find the support of the topothesy function on (-pi/2, pi/2).
-        ind = nonzero(self.topo.values != 0)
-        self.topo.values = self.topo.values[ind]
-        self.hurst.values = self.hurst.values[ind]
-        self.hurst.t = self.hurst.t[ind]
-
         # Minimum and maximum of the Hurst function on this support.
         self.H = amin(self.hurst.values)
         self.Hmax = amax(self.hurst.values)
@@ -457,10 +451,44 @@ def ComputeFeatures(self):
             hmean = hmean - pi
         self.hurst_argmin_mean = hmean
 
+    def ComputeFeatures_Hurst(self):
+        """Compute several features of the field related to the Hurst index.
+        :returns: Attributes H, Hmax, hurst_argmin_lenght, hurst_argmin_center.
+        """
+
+        if self.CheckValidity() is False:
+            return(0)
+
+        if "step" not in self.hurst.ftype or "step" not in self.topo.ftype:
+            print("ComputeFeature_Hurst: only apply to step functions.")
+            return(0)
+
+        finter = self.hurst.finter
+        inter = concatenate((finter[0, -1].reshape((1, 1)) - pi, finter),
+                            axis=1)
+        centers = inter[:, 0:-1] + (inter[:, 1:] - inter[:, 0:-1]) / 2
+        self.hurst.Evaluate(centers)
+        self.H = amin(self.hurst.fparam)
+        self.Hmax = amax(self.hurst.fparam)
+        i = argmin(self.hurst.values)
+
+        if i > 0:
+            s = self.hurst.finter[0, i] - self.hurst.finter[0, i - 1]
+            c = self.hurst.finter[0, i - 1] + s / 2
+        else:
+            s = self.hurst.finter[0, i] - self.hurst.finter[0, -1] + pi
+            c = self.hurst.finter[0, -1] - pi + s / 2
+
+        self.hurst_argmin_lenght = inter[0, i + 1] - inter[0, i]
+        c = centers[0, i]
+        if c < - pi / 2:
+            c = c + pi
+        self.hurst_argmin_center = c
+
 
 def BETA_H(coord, alp1, alp2, H):
     r"""Approximation of an integral useful for the computation of
-    semi-variogram.
+        semi-variogram.
 
     The approximated integral is defined as:
 

afbf/Classes/PeriodicFunction.py

@@ -207,6 +207,8 @@ def __init__(self, ftype='step-smooth', param=0, fname='noname'):
         self.ftype = 'undefined'
         self.fparam = None
         self.finter = None
+        self.steptrans = None
+        self.trans = None
         self.translate = 0
         self.rescale = 1
 
@@ -217,7 +219,6 @@ def __init__(self, ftype='step-smooth', param=0, fname='noname'):
         else:
             print("PeriodicFunction.__init__: invalid representation.")
             print("PeriodicFunction.__init__: function not defined.")
-
         self.t = None
         self.values = None
         self.basis = None
@@ -532,6 +533,7 @@ def SampleStepIntervals(self):
                 k = mod(k + 1, ninter)
             else:
                 k = 1
+
             # Set interval bounds between -pi/2 and pi/2.
             finter = finter - pi / 2
             if self.steptrans:

afbf/Classes/RandomProcess.py

@@ -45,7 +45,7 @@
 """
 
 from afbf.utilities import pi, randn, rand, absolute, concatenate, fft, arange
-from afbf.utilities import amin, reshape, mean, zeros, diff
+from afbf.utilities import amin, reshape, mean, zeros
 from afbf.utilities import real, cumsum, power, sqrt, plt, log, nonzero
 from afbf import perfunction
 
@@ -141,8 +141,7 @@ def ComputeAutocovarianceSpectrum(self):
     def ComputeSemiVariogram(self, lags, logvario=0):
         """Compute the semi-variogram of the process at lags given in lags.
 
-        :param :ref:`ndarray` lags: Lags where to compute the variogram.
-
+        :param `ndarray`_ lags: Lags where to compute the variogram.
         :param logvario:
             if logvario>0, a log semi-variogram is computed.
             The default is 0.
@@ -175,7 +174,7 @@ def ComputeFBMAutocovariance(self, T):
     def ComputeFBMSemiVariogram(self, lags, logvario=0):
         """Compute the semi-variogram of the fbm at lags given in lags.
 
-        :param :ref:`ndarray` lags: Lags where to compute the variogram.
+        :param `ndarray`_ lags: Lags where to compute the variogram.
         :param logvario: if logvario>0, a log semi-variogram is computed.
             The default is 0.
         :type logvario: int, optional

afbf/Classes/SpatialData.py

@@ -43,7 +43,7 @@
 
 from afbf.utilities import reshape, arange, repmat, concatenate, array, zeros
 from afbf.utilities import amin, amax, mean, matmul, floor, linspace, sign
-from afbf.utilities import power, plt, sqrt, unique
+from afbf.utilities import power, plt, sqrt
 from afbf.utilities import make_axes_locatable, ndarray
 from afbf.utilities import imread
 
@@ -285,6 +285,7 @@ def Display(self, nfig=1):
         plt.figure(nfig)
         plt.plot(self.xy[:, 0] / self.N, self.xy[:, 1] / self.N, "rx")
         plt.axis("equal")
+        plt.show()
 
 
 class sdata: