Release v0.3.1 #232 from WEgeophysics/develop

v0.3.1 (March 7, 2024)
-------------------------

Changes and enhancements have been made to the API since version ``v0.3.0``, leading to the introduction of new features and the resolution of various bugs.

- |API change| From now on, only the selected utilities listed below are available as the public API, which has led to shorter loading times for the package. The available 
  public API functions include: 

    - :func:`watex.utils.bi_selector` available publicly as :func:`watex.bi_selector`
    - :func:`watex.utils.cleaner` available publicly as :func:`watex.cleaner`
    - :func:`watex.utils.erpSelector` available publicly as :func:`watex.erpSelector`
    - :func:`watex.utils.erpSmartDetector` available publicly as :func:`watex.erpSmartDetector`
    - :func:`watex.utils.fetch_data` available publicly as :func:`watex.fetch_data`
    - :func:`watex.utils.fittensor` available publicly as :func:`watex.fittensor`
    - :func:`watex.utils.get2dtensor` available publicly as :func:`watex.get2dtensor`
    - :func:`watex.utils.make_erp` available publicly as :func:`watex.make_erp`
    - :func:`watex.utils.make_naive_pipe` available publicly as :func:`watex.make_naive_pipe`
    - :func:`watex.utils.make_ves` available publicly as :func:`watex.make_ves`
    - :func:`watex.utils.magnitude` available publicly as :func:`watex.magnitude`
    - :func:`watex.utils.naive_imputer` available publicly as :func:`watex.naive_imputer`
    - :func:`watex.utils.naive_scaler` available publicly as :func:`watex.naive_scaler`
    - :func:`watex.utils.ohmicArea` available publicly as :func:`watex.ohmicArea`
    - :func:`watex.utils.plotAnomaly` available publicly as :func:`watex.plotAnomaly`
    - :func:`watex.utils.plot_confidence_in` available publicly as :func:`watex.plot_confidence_in`
    - :func:`watex.utils.plotOhmicArea` available publicly as :func:`watex.plotOhmicArea`
    - :func:`watex.utils.plot_sfi` available publicly as :func:`watex.plot_sfi`
    - :func:`watex.utils.power` available publicly as :func:`watex.power`
    - :func:`watex.utils.qc` available publicly as :func:`watex.qc`
    - :func:`watex.utils.read_data` available publicly as :func:`watex.read_data`
    - :func:`watex.utils.selectfeatures` available publicly as :func:`watex.selectfeatures`
    - :func:`watex.utils.sfi` available publicly as :func:`watex.sfi`
    - :func:`watex.utils.shape` available publicly as :func:`watex.shape`
    - :func:`watex.utils.smart_label_classifier` available publicly as :func:`watex.smart_label_classifier`
    - :func:`watex.utils.to_numeric_dtypes` available publicly as :func:`watex.to_numeric_dtypes`
    - :func:`watex.utils.type_` available publicly as :func:`watex.type_`
    - :func:`watex.utils.vesSelector` available publicly as :func:`watex.vesSelector`

- |API change| The parameter ``edi_obj`` in :func:`watex.utils.plot_skew` has been deprecated and replaced by ``edis_list``, which refers to either a collection of :term:`EDI` files or a full path to EDI files. Two-dimensional skewness, represented as :math:`eta`, can now be visualized by setting the parameter ``view='2D'``.

- |API change| The ``xgboost`` library is no longer automatically installed as a dependency. Users must install it separately for complete model functionality or use the ``dev`` option as shown below:

  .. code-block:: bash 

    pip install watex[dev]

- |Fix| The :class:`watex.em.EM` module now throws a :class:`watex.exceptions.EDIError` instead of an ``AttributeError`` when an EDI file is expected but an object is passed.

- |Fix| The `base_estimator` parameter is no longer available in scikit-learn versions greater than 1.2; it has been renamed to `estimator`. Consequently, :mod:`watex.models` and :mod:`watex.models.premodels` have been updated to reflect this change.

- |Feature| The :func:`watex.utils.plotutils.plot_l_curve` function has been introduced to plot the Hansen L-curve, with an option to highlight the Hansen point. This feature uses the L-curve criterion to determine the most suitable model after performing multiple inversions with different :math:`\tau` values.

docs/whatsnew/index.rst

@@ -13,6 +13,7 @@ Version 0.3
 .. toctree::
    :maxdepth: 2
 
+   v0.3.1
    v0.3.0
 
 
@@ -22,6 +23,7 @@ Version 0.2
 .. toctree::
    :maxdepth: 2
 
+   v0.2.8
    v0.2.7
    v0.2.6
    v0.2.5

watex/methods/em.py

@@ -3947,7 +3947,7 @@ def filter_noises (
       kind of filtering technique to smooth data. Can be: 
           
           - 'base': for simple moving-average using convolution strategy 
-          - 'ama': for naive adapttive moving average 
+          - 'ama': for naive adaptative moving average 
           - 'butter': for Butterworth filter using bandpass strategy. (lowcut 
             highcut) can be set using the `frange` parameters.
           - 'tv': for Torres-Verdin filter [1]_
@@ -4510,7 +4510,7 @@ def _ss_auto_regulator (
 
         init_freq = len(edi_obj_init.Z.freq) 
         half_two = (init_freq -2) //2
-        nfreq =  half_two + (half_two//2 )
+        nfreq =  round (  + (half_two//2 )) 
         skipfreq = half_two * 2  -nfreq 
 
     else: 

watex/utils/plotutils.py

@@ -3942,6 +3942,9 @@ def plot_voronoi(
     line_width=1. ,
     line_alpha=1.,   
     fig_size = (7, 7), 
+    cmap='set1', 
+    show_grid=True, 
+    alpha=0.2, 
     fig_title = ''
     ):
     """Plots the Voronoi diagram of the k-means clusters overlaid with 
@@ -3991,18 +3994,27 @@ def plot_voronoi(
 
     from scipy.spatial import Voronoi, voronoi_plot_2d
 
-    ax.scatter(X[:, 0], X[:, 1], c=y, cmap='Set1', alpha=0.2, 
+    ax.scatter(X[:, 0], X[:, 1], c=y, cmap=cmap, alpha=0.2, 
                label = 'Voronoi plot')
     vor = Voronoi(cluster_centers)
     voronoi_plot_2d(vor, ax=ax, show_vertices=show_vertices, 
-                    alpha=0.5, 
+                    alpha=alpha, 
                     line_colors=line_colors,
                     line_width=line_width ,
                     line_alpha=line_alpha,  
                     )
     #ax.legend() 
     ax.set_title (fig_title , fontsize=20)
     #fig.suptitle(fig_title, fontsize=20) 
+    # Make the right and bottom spines thicker and black
+    ax.spines['left'].set_linewidth(2)
+    ax.spines['left'].set_color('black')
+    ax.spines['bottom'].set_linewidth(2)
+    ax.spines['bottom'].set_color('black')
+
+    if show_grid: plt.grid() 
+    else: ax.grid(False)
+
     return ax 
 
 

watex/view/mlplot.py

@@ -2887,12 +2887,18 @@ def plotLearningInspections (
 
     if not is_iterable( axes) :
         axes =[axes ] 
+    cscores =[]; vscores =[]
     for kk, model in enumerate ( models ) : 
         title = titles[kk] or  get_estimator_name (model )
-        plotLearningInspection(model, X=X , y=y, axes = axes [:, kk], 
+        _, csc, vsc = plotLearningInspection(model, X=X , y=y, axes = axes [:, kk], 
                                title =title, 
                                **kws)
-
+        cscores.append(csc); vscores.append (vsc)
+    print("*"*77)
+    print("Convergence scores =", np.around(np.mean(cscores), 4))
+    print("validation scores =", np.around (np.nanmean(vscores), 4))
+    print("*"*77)
+
     if savefig : 
         fig.savefig (savefig , dpi = 300 )
     plt.show () if savefig is None else plt.close () 
@@ -3098,7 +3104,7 @@ def plotLearningInspection(
     axes[2].set_ylabel("Score")
     axes[2].set_title(f"Performance of {title_name}")
 
-    return axes
+    return axes , np.mean([train_scores[-1], test_scores[-1]]),test_scores_mean 
 #XXX
 def plot_matshow(
     arr, / , labelx:List[str] =None, labely:List[str]=None,