Add _backend to project structure

aaanalysis/__init__.py

@@ -1,14 +1,27 @@
-from aaanalysis.data_handling import load_dataset, load_scales
-from aaanalysis.feature_engineering import AAclust, AAclustPlot, CPP, CPPPlot, SequenceFeature, SplitRange
-from aaanalysis.pu_learning import dPULearn
-from aaanalysis.plotting import (plot_get_clist, plot_get_cmap, plot_get_cdict,
+from .data_handling import load_dataset, load_scales
+from .feature_engineering import AAclust, AAclustPlot, CPP, CPPPlot, SequenceFeature, SplitRange
+from .pu_learning import dPULearn
+from .plotting import (plot_get_clist, plot_get_cmap, plot_get_cdict,
                                  plot_settings, plot_legend, plot_gcfs)
-from aaanalysis.config import options
+from .config import options
 
-__all__ = ["load_dataset", "load_scales",
-           "AAclust", "AAclustPlot",
-           "CPP", "CPPPlot", "SequenceFeature", "SplitRange",
-           "dPULearn", "plot_get_clist", "plot_get_cmap", "plot_get_cdict",
-           "plot_settings", "plot_legend", "plot_gcfs", "options"]
+__all__ = [
+    "load_dataset",
+    "load_scales",
+    "AAclust",
+    "AAclustPlot",
+    "CPP",
+    "CPPPlot",
+    "SequenceFeature",
+    "SplitRange",
+    "dPULearn",
+    "plot_get_clist",
+    "plot_get_cmap",
+    "plot_get_cdict",
+    "plot_settings",
+    "plot_legend",
+    "plot_gcfs",
+    "options"
+]
 
 

aaanalysis/_utils/utils_cpp.py

@@ -29,7 +29,7 @@
                     "\nwhere all numbers should be non-negative integers, and N/C means N or C."\
     .format(STR_SEGMENT, STR_PATTERN, STR_PERIODIC_PATTERN)
 
-
+# TODO to CPP backend
 # II Main Functions
 # General check functions
 def check_color(name=None, val=None, accept_none=False):

aaanalysis/data_handling/__init__.py

@@ -1,4 +1,7 @@
-from aaanalysis.data_handling.load_dataset_ import load_dataset
-from aaanalysis.data_handling.load_scales_ import load_scales
+from ._load_dataset import load_dataset
+from ._load_scales import load_scales
 
-__all__ = ["load_dataset", "load_scales"]
+__all__ = [
+    "load_dataset",
+    "load_scales"
+]

aaanalysis/explainers/tree_based.py → aaanalysis/explainable_ai/tree_based.py

@@ -17,7 +17,7 @@
 
 # II Main Functions
 class TreeModel:
-    """A wrapper for Tree based prediction models and Tree explainers from SHAP package to
+    """A wrapper for Tree based prediction models and Tree explainable_ai from SHAP package to
     explain prediction (typically binary classification) results at global and individual level"""
     def __init__(self, model=None):
         """"""
@@ -33,7 +33,7 @@ def add_feat_import(self, df_feat=None):
 
 
 class ShapModel:
-    """A wrapper for Tree explainers from SHAP package"""
+    """A wrapper for Tree explainable_ai from SHAP package"""
     def __init__(self, model=None):
         """"""
 

aaanalysis/feature_engineering/__init__.py

@@ -1,9 +1,14 @@
-from aaanalysis.feature_engineering.aaclust_ import AAclust
-from aaanalysis.feature_engineering.aaclust_plot_ import AAclustPlot
-from aaanalysis.feature_engineering.cpp.feature import SequenceFeature
-from aaanalysis.feature_engineering.cpp.feature import SplitRange
-from aaanalysis.feature_engineering.cpp_plot_ import CPPPlot
-from aaanalysis.feature_engineering.cpp_ import CPP
+from ._aaclust import AAclust
+from ._aaclust_plot import AAclustPlot
+from ._cpp_plot import CPPPlot
+from ._cpp import CPP
+from ._backend.cpp.feature import SequenceFeature, SplitRange
 
-
-__all__ = ["AAclust", "AAclustPlot", "CPP", "CPPPlot", "SequenceFeature", "SplitRange"]
+__all__ = [
+    "AAclust",
+    "AAclustPlot",
+    "SequenceFeature",
+    "SplitRange",
+    "CPP",
+    "CPPPlot",
+]

aaanalysis/feature_engineering/aaclust_.py → aaanalysis/feature_engineering/_aaclust.py

@@ -14,12 +14,11 @@
 from aaanalysis.template_classes import Wrapper
 import aaanalysis.utils as ut
 
-from aaanalysis.feature_engineering.aaclust._aaclust import (estimate_lower_bound_n_clusters, optimize_n_clusters, merge_clusters,
-                                                             compute_centers, compute_medoids)
-from aaanalysis.feature_engineering.aaclust._aaclust_bic import bic_score
-from aaanalysis.feature_engineering.aaclust._aaclust_comp_corr import compute_correlation
-from aaanalysis.feature_engineering.aaclust._aaclust_name_clusters import name_clusters
-
+from ._backend.aaclust.aaclust_fit import estimate_lower_bound_n_clusters, optimize_n_clusters, merge_clusters
+from ._backend.aaclust.aaclust_eval import bic_score
+from ._backend.aaclust.aaclust_methods import (compute_centers, compute_medoids,
+                                               name_clusters,
+                                               compute_correlation)
 
 # I Helper Functions
 # Check parameter functions
@@ -159,7 +158,7 @@ def fit(self,
             X: ut.ArrayLike2D,
             n_clusters: Optional[int] = None,
             on_center: bool = True,
-            min_th: float = 0,
+            min_th: float = 0.3,
             merge_metric: Union[str, None] = "euclidean",
             names: Optional[List[str]] = None) -> "AAclust":
         """
@@ -229,8 +228,7 @@ def fit(self,
         check_match_X_n_clusters(X=X, n_clusters=n_clusters, accept_none=True)
         check_match_X_names(X=X, names=names, accept_none=True)
 
-        args = dict(model=self.model_class, model_kwargs=self._model_kwargs, min_th=min_th, on_center=on_center,
-                    verbose=self._verbose)
+        args = dict(model=self.model_class, model_kwargs=self._model_kwargs, min_th=min_th, on_center=on_center)
 
         # Clustering using given clustering models
         if n_clusters is not None:
@@ -240,15 +238,21 @@ def fit(self,
         # Clustering using AAclust algorithm
         else:
             # Step 1.: Estimation of lower bound of k (number of clusters)
+            if self._verbose:
+                ut.print_out("1. Estimation of lower bound of k (number of clusters)", end="")
             n_clusters_lb = estimate_lower_bound_n_clusters(X, **args)
             # Step 2. Optimization of k by recursive clustering
+            if self._verbose:
+                objective_fct = "min_cor_center" if on_center else "min_cor_all"
+                ut.print_out(f"2. Optimization of k by recursive clustering ({objective_fct}, min_th={min_th}, k={n_clusters_lb})", end="")
             n_clusters = optimize_n_clusters(X, n_clusters=n_clusters_lb, **args)
             self.model = self.model_class(n_clusters=n_clusters, **self._model_kwargs)
             labels = self.model.fit(X).labels_.tolist()
             # Step 3. Cluster merging (optional)
             if merge_metric is not None:
-                labels = merge_clusters(X, labels=labels, min_th=min_th, on_center=on_center,
-                                        metric=merge_metric, verbose=self._verbose)
+                if self._verbose:
+                    ut.print_out(f"3. Cluster merging (k={len(labels)})", end="")
+                labels = merge_clusters(X, labels=labels, min_th=min_th, on_center=on_center, metric=merge_metric)
                 n_clusters = len(set(labels))
 
         # Obtain cluster centers and medoids

aaanalysis/feature_engineering/aaclust_plot_.py → aaanalysis/feature_engineering/_aaclust_plot.py

@@ -1,18 +1,15 @@
 """
 This is a script for the plotting class of AAclust.
 """
-import time
-import pandas as pd
-import numpy as np
 from sklearn.decomposition import PCA
 from typing import Optional, Dict, Union, List, Tuple, Type
 from sklearn.base import  TransformerMixin
-import matplotlib.pyplot as plt
-import seaborn as sns
 
 import aaanalysis as aa
 import aaanalysis.utils as ut
-from aaanalysis.feature_engineering.aaclust_plot._aaclust_plot_eval import plot_eval
+
+from ._backend.aaclust_plot.aaclust_plot_eval import plot_eval
+
 
 # I Helper Functions
 def _get_components(data=None, model_class=None):

aaanalysis/feature_engineering/_backend/aaclust/_utils_aaclust.py

@@ -0,0 +1,23 @@
+"""
+This is a script for utility functions for aaclust object.
+"""
+import numpy as np
+
+
+# II Main Functions
+def _cluster_center(X):
+    """Compute cluster center (i.e., arithmetical mean over all data points/observations of a cluster)"""
+    return X.mean(axis=0)[np.newaxis, :]
+
+
+def _cluster_medoid(X):
+    """Obtain cluster medoids (i.e., scale closest to cluster center used as representative scale for a cluster)"""
+    # Create new array with cluster center and given array
+    center_X = np.concatenate([_cluster_center(X), X], axis=0)
+    # Get index for scale with the highest correlation with cluster center
+    ind_max = np.corrcoef(center_X)[0, 1:].argmax()
+    return ind_max
+
+
+
+

aaanalysis/feature_engineering/aaclust/_aaclust_bic.py → aaanalysis/feature_engineering/_backend/aaclust/aaclust_eval.py

@@ -1,12 +1,14 @@
 """
-This is a script for computing the Bayesian Information Criterion (BIC).
+This is a script for computing the Bayesian Information Criterion (BIC) used in the AAclust.eval() method.
 """
 import numpy as np
 from scipy.spatial import distance
-from aaanalysis.feature_engineering.aaclust._aaclust import compute_centers
+
+from .aaclust_methods import compute_centers
 
 # I Helper Functions
 
+
 # II Main function
 def bic_score(X, labels=None):
     """Computes the BIC metric for given clusters.
@@ -53,5 +55,4 @@ def bic_score(X, labels=None):
 
     bic_components = size_clusters * (log_size_clusters - log_n_samples) - 0.5 * size_clusters * n_features * log_bcv - 0.5 * (size_clusters - 1) * n_features
     bic = np.sum(bic_components) - const_term
-
     return bic