TINTOlib 1.2.1__tar.gz → 1.3.0__tar.gz

{tintolib-1.2.1 → tintolib-1.3.0}/.gitignore

@@ -1,12 +1,14 @@
-**/__pycache__
-=======
-#Misc
+# Misc
+.DS_Store
 HyNNImages/
+SyntheticImages/
 logs/
 data/
 scripts_logs/
+objs.pkl
 
 # Byte-compiled / optimized / DLL files
+**/__pycache__/
 __pycache__/
 *.py[cod]
 *$py.class

{tintolib-1.2.1 → tintolib-1.3.0}/PKG-INFO

@@ -1,9 +1,13 @@
 Metadata-Version: 2.4
 Name: TINTOlib
-Version: 1.2.1
+Version: 1.3.0
 Summary: Converting tabular data into images
-Project-URL: Homepage, https://github.com/oeg-upm/TINTOlib-Documentation
-Author-email: BorjaRei <borjareinoso@gmail.com>, manwestc <jcastillo@fi.upm.es>, DavidGonzalezFernandez <david.g.f.2@gmail.com>, JiayunLiu <jiayun.liu@upm.es>, JavierLopez <javilecrin19@gmail.com>, SalvadorMartinez <salvador.martinez.moreno@gmail.com>
+Project-URL: Homepage, https://github.com/oeg-upm/TINTOlib
+Project-URL: Documentation, https://tintolib.readthedocs.io/
+Project-URL: Benchmarks, https://oeg-upm.github.io/TINTOlib/
+Project-URL: Repository, https://github.com/oeg-upm/TINTOlib
+Project-URL: Issues, https://github.com/oeg-upm/TINTOlib/issues
+Author-email: BorjaRei <borjareinoso@gmail.com>, manwestc <jcastillo@fi.upm.es>, DavidGonzalezFernandez <david.g.f.2@gmail.com>, JiayunLiu <jiayun.liu@upm.es>, JavierLopez <javilecrin19@gmail.com>
 License:                                  Apache License
                                    Version 2.0, January 2004
                                 http://www.apache.org/licenses/
@@ -209,24 +213,38 @@ License-File: LICENSE
 Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3
-Requires-Python: >=3.8
-Requires-Dist: asymmetric-greedy-search==0.1.0
-Requires-Dist: bitstring==4.3.0
-Requires-Dist: matplotlib==3.9.4
-Requires-Dist: numba==0.62.0
-Requires-Dist: numpy==1.26.4
-Requires-Dist: opencv-python==4.11.0.86
-Requires-Dist: pandas==2.2.3
-Requires-Dist: scikit-learn==1.6.1
-Requires-Dist: scipy==1.15.0
-Requires-Dist: tqdm==4.67.1
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Requires-Python: >=3.9
+Requires-Dist: bitstring<5,>=4.0
+Requires-Dist: matplotlib<4,>=3.7
+Requires-Dist: numpy<3,>=1.24
+Requires-Dist: pandas<3,>=2.0
+Requires-Dist: pillow<13,>=9.5
+Requires-Dist: scikit-learn<2,>=1.2
+Requires-Dist: scipy<2,>=1.10
+Requires-Dist: tqdm<5,>=4.65
+Provides-Extra: all
+Requires-Dist: asymmetric-greedy-search==0.1.0; extra == 'all'
+Requires-Dist: mpi4py<5,>=4; extra == 'all'
+Requires-Dist: numba<0.61,>=0.60; (python_version < '3.10') and extra == 'all'
+Requires-Dist: numba<0.63,>=0.62; (python_version >= '3.10') and extra == 'all'
+Provides-Extra: greedy
+Requires-Dist: asymmetric-greedy-search==0.1.0; extra == 'greedy'
+Requires-Dist: numba<0.61,>=0.60; (python_version < '3.10') and extra == 'greedy'
+Requires-Dist: numba<0.63,>=0.62; (python_version >= '3.10') and extra == 'greedy'
+Provides-Extra: refined
+Requires-Dist: mpi4py<5,>=4; extra == 'refined'
 Description-Content-Type: text/markdown
 
 # TINTOlib: Python Library to convert Tabular Data into Synthetic Images
 
 [![GitHub Pages](https://img.shields.io/badge/GitHub%20Pages-View%20Documentation-brightgreen?logo=github)](https://oeg-upm.github.io/TINTOlib/)
 [![License](https://img.shields.io/badge/license-Apache%202.0-blue)](https://github.com/oeg-upm/TINTOlib-Documentation/blob/main/LICENSE)
-[![Python Version](https://shields.io/badge/python-3.8+-blue)](https://pypi.python.org/pypi/)
+[![Python Version](https://shields.io/badge/python-3.9+-blue)](https://pypi.python.org/pypi/)
 [![Documentation Status](https://readthedocs.org/projects/morph-kgc/badge/?version=latest)](https://tintolib.readthedocs.io/en/latest/)
 [![Open In Colab-CNN](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/oeg-upm/TINTOlib-Crash_Course/blob/main/Notebooks/Challenge/Regression_CNN.ipynb)
 [![Open In Colab-CNN+MLP](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/oeg-upm/TINTOlib-Crash_Course/blob/main/Notebooks/Challenge/Regression_CNN%2BMLP.ipynb)
@@ -329,7 +347,7 @@ And use-case developed in **[INFFUS Paper](https://doi.org/10.1016/j.inffus.2022
 - Designed for tidy data (**target column last**)
 - Output: grayscale images from reduction and transformation methods
 - Compatible with **Linux, Windows, macOS**
-- Requires **Python 3.8+**
+- Requires **Python 3.9+**
 
 ---
 
@@ -374,12 +392,12 @@ This method relies on `mpi4py`, which enables parallel computation using MPI (Me
   Once MPI is installed:
 
     ```bash
-      pip install mpi4py
+      pip install "TINTOlib[refined]"
     ```
 - **MacOS / Windows:** Direct installation is usually supported:
 
     ```bash
-      pip install mpi4py
+      pip install "TINTOlib[refined]"
     ```
 
 ### SuperTML
@@ -408,10 +426,18 @@ On **Google Colab**, installing additional fonts is not permitted due to adminis
 
 TINTOlib already includes all necessary dependencies, so there’s no need to install them individually.
 
+Optional methods can be installed using dependency extras:
+
+```bash
+pip install "TINTOlib[greedy]"  # Greedy feature-to-pixel assignment
+pip install "TINTOlib[refined]" # REFINED and mpi4py
+pip install "TINTOlib[all]"     # All optional methods
+```
+
 However, if you prefer manual installation or want to explore the full environment:
 
-- The repository includes a `requirements.txt` file listing the **core dependencies** required to use TINTOlib. You can directly run the **TINTOlib-example.ipynb** notebook located in the examples/ folder using the dependencies listed in `requirements.txt`.
-- **mpi4py** is only required if you use the **REFINED** method; see the [platform-specific notes](#platform-specific-notes).
+- The repository includes a `requirements.txt` file that installs TINTOlib and its core dependencies in editable mode for local development.
+- **mpi4py** is installed by the `refined` extra and is only required for **REFINED**; an MPI runtime must also be installed as described in the [platform-specific notes](#platform-specific-notes).
 - **Other notebooks**, which include training deep learning models on the generated images, require additional libraries. To run them, install the extended dependencies from `requirements-example.txt`:
 
 ---

{tintolib-1.2.1 → tintolib-1.3.0}/README.md

@@ -2,7 +2,7 @@
 
 [![GitHub Pages](https://img.shields.io/badge/GitHub%20Pages-View%20Documentation-brightgreen?logo=github)](https://oeg-upm.github.io/TINTOlib/)
 [![License](https://img.shields.io/badge/license-Apache%202.0-blue)](https://github.com/oeg-upm/TINTOlib-Documentation/blob/main/LICENSE)
-[![Python Version](https://shields.io/badge/python-3.8+-blue)](https://pypi.python.org/pypi/)
+[![Python Version](https://shields.io/badge/python-3.9+-blue)](https://pypi.python.org/pypi/)
 [![Documentation Status](https://readthedocs.org/projects/morph-kgc/badge/?version=latest)](https://tintolib.readthedocs.io/en/latest/)
 [![Open In Colab-CNN](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/oeg-upm/TINTOlib-Crash_Course/blob/main/Notebooks/Challenge/Regression_CNN.ipynb)
 [![Open In Colab-CNN+MLP](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/oeg-upm/TINTOlib-Crash_Course/blob/main/Notebooks/Challenge/Regression_CNN%2BMLP.ipynb)
@@ -105,7 +105,7 @@ And use-case developed in **[INFFUS Paper](https://doi.org/10.1016/j.inffus.2022
 - Designed for tidy data (**target column last**)
 - Output: grayscale images from reduction and transformation methods
 - Compatible with **Linux, Windows, macOS**
-- Requires **Python 3.8+**
+- Requires **Python 3.9+**
 
 ---
 
@@ -150,12 +150,12 @@ This method relies on `mpi4py`, which enables parallel computation using MPI (Me
   Once MPI is installed:
 
     ```bash
-      pip install mpi4py
+      pip install "TINTOlib[refined]"
     ```
 - **MacOS / Windows:** Direct installation is usually supported:
 
     ```bash
-      pip install mpi4py
+      pip install "TINTOlib[refined]"
     ```
 
 ### SuperTML
@@ -184,10 +184,18 @@ On **Google Colab**, installing additional fonts is not permitted due to adminis
 
 TINTOlib already includes all necessary dependencies, so there’s no need to install them individually.
 
+Optional methods can be installed using dependency extras:
+
+```bash
+pip install "TINTOlib[greedy]"  # Greedy feature-to-pixel assignment
+pip install "TINTOlib[refined]" # REFINED and mpi4py
+pip install "TINTOlib[all]"     # All optional methods
+```
+
 However, if you prefer manual installation or want to explore the full environment:
 
-- The repository includes a `requirements.txt` file listing the **core dependencies** required to use TINTOlib. You can directly run the **TINTOlib-example.ipynb** notebook located in the examples/ folder using the dependencies listed in `requirements.txt`.
-- **mpi4py** is only required if you use the **REFINED** method; see the [platform-specific notes](#platform-specific-notes).
+- The repository includes a `requirements.txt` file that installs TINTOlib and its core dependencies in editable mode for local development.
+- **mpi4py** is installed by the `refined` extra and is only required for **REFINED**; an MPI runtime must also be installed as described in the [platform-specific notes](#platform-specific-notes).
 - **Other notebooks**, which include training deep learning models on the generated images, require additional libraries. To run them, install the extended dependencies from `requirements-example.txt`:
 
 ---

{tintolib-1.2.1 → tintolib-1.3.0}/TINTOlib/clusters.py

@@ -134,20 +134,205 @@ class Clusters(AbstractImageMethod):
         normalize=None,
         verbose = None,
         algorithm = default_algorithm,
-        n_clusters = default_n_clusters,
-        random_seed=default_random_seed,
-        n_init = default_n_init,
-        max_iter = default_max_iter,
-        algorithmMethod = default_algorithmMethod,
-        covariance_type = default_covariance_type,
-        ensamMethod = default_ensamMethod,
-        bandwidth = default_bandwidth,
-        kernel = default_kernel,
-        metric=default_metric,
-        RBFKmeans=default_RBFKmeans
+        n_clusters = None,
+        random_seed=None,
+        n_init = None,
+        max_iter = None,
+        algorithmMethod = None,
+        covariance_type = None,
+        ensamMethod = None,
+        bandwidth = None,
+        kernel = None,
+        metric = None,
+        RBFKmeans = None
         
     ):
         
+        # kmeans 
+        # --------------------------------------------------
+        if algorithm == 'kmeans' and (covariance_type or ensamMethod or bandwidth or kernel or metric):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: n_clusters, random_seed, n_init, max_iter, algorithmMethod and RBFKmeans."
+            )
+        else:
+            if algorithm == 'kmeans':
+                if n_clusters is None:
+                    n_clusters = self.default_n_clusters
+                if random_seed is None:
+                    random_seed = self.default_random_seed
+                if n_init is None:
+                    n_init = self.default_n_init
+                if max_iter is None:
+                    max_iter = self.default_max_iter
+                if algorithmMethod is None:
+                    algorithmMethod = self.default_algorithmMethod
+                if RBFKmeans is None:
+                    RBFKmeans = self.default_RBFKmeans
+        # --------------------------------------------------
+        
+        # gaussianMix
+        # --------------------------------------------------
+        if algorithm == 'gaussianMix' and (algorithmMethod or ensamMethod or bandwidth or kernel or metric or RBFKmeans):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: n_clusters, random_seed, n_init, max_iter and covariance_type."
+            )
+        else:
+            if algorithm == 'gaussianMix':
+                if n_clusters is None:
+                    n_clusters = self.default_n_clusters
+                if random_seed is None:
+                    random_seed = self.default_random_seed
+                if n_init is None:
+                    n_init = self.default_n_init
+                if max_iter is None:
+                    max_iter = self.default_max_iter
+                if covariance_type is None:
+                    covariance_type = self.default_covariance_type
+        # --------------------------------------------------
+        
+        # aggloKNN
+        # --------------------------------------------------
+        if algorithm == 'aggloKNN' and (covariance_type or ensamMethod or bandwidth or kernel or random_seed or n_init or max_iter or algorithmMethod or RBFKmeans):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: n_clusters and metric."
+            )
+        else:
+            if algorithm == 'aggloKNN':
+                if n_clusters is None:
+                    n_clusters = self.default_n_clusters
+                if metric is None:
+                    metric = self.default_metric
+        # --------------------------------------------------
+        
+        # kde
+        # --------------------------------------------------
+        if algorithm == 'kde' and (n_clusters or covariance_type or ensamMethod or random_seed or n_init or max_iter or algorithmMethod or RBFKmeans):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: metric, kernel and bandwidth."
+            )
+        else:
+            if algorithm == 'kde':
+                if kernel is None:
+                    kernel = self.default_kernel
+                if bandwidth is None:
+                    bandwidth = self.default_bandwidth
+                if metric is None:
+                    metric = self.default_metric
+        # --------------------------------------------------
+        
+        # kmedoids
+        # --------------------------------------------------
+        if algorithm == 'kmedoids' and (covariance_type or ensamMethod or n_init or algorithmMethod or RBFKmeans or kernel or bandwidth):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: metric, n_clusters, max_iter and random_seed."
+            )
+        else:
+            if algorithm == 'kmedoids':
+                if n_clusters is None:
+                    n_clusters = self.default_n_clusters
+                if max_iter is None:
+                    max_iter = self.default_max_iter
+                if random_seed is None:
+                    random_seed = self.default_random_seed
+                if metric is None:
+                    metric = self.default_metric
+        # --------------------------------------------------
+        
+        # factor
+        # --------------------------------------------------
+        if algorithm == 'factor' and (metric or max_iter or covariance_type or ensamMethod or n_init or algorithmMethod or RBFKmeans or kernel or bandwidth):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: n_clusters and random_seed."
+            )
+        else:
+            if algorithm == 'factor':
+                if n_clusters is None:
+                    n_clusters = self.default_n_clusters
+                if random_seed is None:
+                    random_seed = self.default_random_seed                
+        # --------------------------------------------------
+        
+        # mixMethod
+        # --------------------------------------------------
+        if algorithm == 'mixMethod':
+            if ensamMethod is None:
+                ensamMethod = self.default_ensamMethod
+            hiperAll={'n_clusters','random_seed','n_init','max_iter','algorithmMethod','RBFKmeans','covariance_type','metric','ensamMethod','bandwidth','kernel'}
+            hiperOK=set()
+            if 'kmeans' in ensamMethod:
+                hiperOK.update(['n_clusters','random_seed','n_init','max_iter','algorithmMethod','RBFKmeans'])
+            if 'gaussianMix' in ensamMethod:
+                hiperOK.update(['n_clusters','random_seed','n_init','max_iter','covariance_type'])
+            if 'aggloKNN' in ensamMethod:
+                hiperOK.update(['n_clusters','metric'])
+            if 'kmedoids' in ensamMethod:
+                hiperOK.update(['n_clusters','metric','random_seed','max_iter'])
+            if 'factor' in ensamMethod:
+                hiperOK.update(['n_clusters','random_seed'])
+            hiperOK.update(['ensamMethod'])
+            hiperKO=hiperAll-hiperOK
+        
+        vari = locals().copy()
+        vari.pop('self', None)
+    
+        if algorithm == 'mixMethod' and (any(vari[name] is not None for name in hiperKO)):
+            raise ValueError(
+                f"Incorrect hyperparameters for the algorithm: '{algorithm}'. "
+                f"The accepted hyperparameters for this algorithm are: {hiperOK}."
+            )
+        else:
+            if algorithm == 'mixMethod':
+                if 'kmeans' in ensamMethod:
+                    if n_clusters is None:
+                        n_clusters = self.default_n_clusters
+                    if random_seed is None:
+                        random_seed = self.default_random_seed
+                    if n_init is None:
+                        n_init = self.default_n_init
+                    if max_iter is None:
+                        max_iter = self.default_max_iter
+                    if algorithmMethod is None:
+                        algorithmMethod = self.default_algorithmMethod
+                    if RBFKmeans is None:
+                        RBFKmeans = self.default_RBFKmeans
+                if 'gaussianMix' in ensamMethod:
+                    if n_clusters is None:
+                        n_clusters = self.default_n_clusters
+                    if random_seed is None:
+                        random_seed = self.default_random_seed
+                    if n_init is None:
+                        n_init = self.default_n_init
+                    if max_iter is None:
+                        max_iter = self.default_max_iter
+                    if covariance_type is None:
+                        covariance_type = self.default_covariance_type
+                if 'aggloKNN' in ensamMethod:
+                    if n_clusters is None:
+                        n_clusters = self.default_n_clusters
+                    if metric is None:
+                        metric = self.default_metric
+                if 'kmedoids' in ensamMethod:
+                    if n_clusters is None:
+                        n_clusters = self.default_n_clusters
+                    if max_iter is None:
+                        max_iter = self.default_max_iter
+                    if random_seed is None:
+                        random_seed = self.default_random_seed
+                    if metric is None:
+                        metric = self.default_metric
+                if 'factor' in ensamMethod:
+                    if n_clusters is None:
+                        n_clusters = self.default_n_clusters
+                    if random_seed is None:
+                        random_seed = self.default_random_seed                    
+        # --------------------------------------------------
+        
         if algorithm=="aggloKNN":
             self.METRICS_TYPES = {'euclidean', 'manhattan', 'cosine'}
         
@@ -160,19 +345,31 @@ class Clusters(AbstractImageMethod):
                 f"Allowed values are: {self.ALGORITHMS}"
             )
         
-        if covariance_type not in self.COV_TYPES:
+        if algorithm == 'gaussianMix' and covariance_type not in self.COV_TYPES:
+            raise ValueError(
+                f"Invalid covariance_type '{covariance_type}'. "
+                f"Allowed values are: {self.COV_TYPES}"
+            )
+        
+        if algorithm == 'mixMethod' and 'gaussianMix' in ensamMethod and covariance_type not in self.COV_TYPES:
             raise ValueError(
                 f"Invalid covariance_type '{covariance_type}'. "
                 f"Allowed values are: {self.COV_TYPES}"
             )
         
-        if kernel not in self.KERNEL_TYPES:
+        if algorithm == 'kde' and kernel not in self.KERNEL_TYPES:
             raise ValueError(
                 f"Invalid kernel '{kernel}'. "
                 f"Allowed values are: {self.KERNEL_TYPES}"
             )
         
-        if metric not in self.METRICS_TYPES:
+        if algorithm in ['kde', 'aggloKNN', 'kmedoids'] and metric not in self.METRICS_TYPES:
+            raise ValueError(
+                f"Invalid metric '{metric}'. "
+                f"Allowed values are: {self.METRICS_TYPES}"
+            )
+        
+        if algorithm == 'mixMethod' and any(method in ensamMethod for method in ['aggloKNN', 'kmedoids']) and metric not in self.METRICS_TYPES:
             raise ValueError(
                 f"Invalid metric '{metric}'. "
                 f"Allowed values are: {self.METRICS_TYPES}"
@@ -195,7 +392,7 @@ class Clusters(AbstractImageMethod):
                 "Algorithms cannot be repeated."
             )
             
-        if (algorithm=="kde" or algorithm=="aggloKNN") and (n_clusters=="auto" or isinstance(n_clusters, list)):
+        if algorithm in ["kde","aggloKNN"] and (n_clusters=="auto" or isinstance(n_clusters, list)):
             raise ValueError(
                 "For these algorithms ('kde', 'aggloKNN'), the 'auto' option to automatically select the optimal number of clusters is not enabled."
             )
@@ -321,7 +518,7 @@ class Clusters(AbstractImageMethod):
                 listDist=listDist[(listDist[:,0]!=ind) & (listDist[:,1]!=ind)]
                 ind=newInd
     
-    # pip install pyclustering
+    # YA NO SE USA: pip install pyclustering
     def __kmedoids_pam(self,X, K, max_iter=100, random_state=0):
         
         """

{tintolib-1.2.1 → tintolib-1.3.0}/TINTOlib/deepInsight.py

@@ -78,8 +78,10 @@ class DeepInsight(ParamImageMethod):
     ):
         if (algorithm_rd not in [constants.pca_algorithm, constants.tsne_algorithm,constants.kpca_algorithm]):
             raise TypeError(f"Algorithm_rd parameter must be in {[constants.pca_algorithm, constants.tsne_algorithm,constants.kpca_algorithm]}")
+        
+        if(algorithm_rd != constants.pca_algorithm and (relocate == True or group_method == constants.relevance_option or assignment_method == constants.relevance_assigner)): 
+            raise TypeError(f"PCA extract method is required to optimize or group using features relevance based on loadings")
 
-        if(algorithm_rd != constants.pca_algorithm): relocate = False
 
         super().__init__( image_dim,problem,transformer,verbose,assignment_method,relocate,algorithm_opt,group_method,zoom,format,cmap,random_seed)
 
@@ -96,17 +98,25 @@ class DeepInsight(ParamImageMethod):
         Returns:
             features coordinate matrix
         """
-        match self.__algorithm_rd:
-            case constants.pca_algorithm:
-                return PCA(n_components=2,random_state=self._random_seed).fit_transform(x)
-            case constants.tsne_algorithm:
-                return TSNE(n_components=2,metric='cosine',random_state=self._random_seed,perplexity=self._image_dim).fit_transform(x)
-            case constants.kpca_algorithm:
-                return KernelPCA(n_components=2,kernel='rbf',random_state=self._random_seed).fit_transform(x)
+        if self.__algorithm_rd == constants.pca_algorithm:
+            return PCA(n_components=2, random_state=self._random_seed).fit_transform(x)
+        elif self.__algorithm_rd == constants.tsne_algorithm:
+            return TSNE(
+                n_components=2,
+                metric='cosine',
+                random_state=self._random_seed,
+                perplexity=self._image_dim
+            ).fit_transform(x)
+        elif self.__algorithm_rd == constants.kpca_algorithm:
+            return KernelPCA(
+                n_components=2,
+                kernel='rbf',
+                random_state=self._random_seed
+            ).fit_transform(x)
 
     def _compute_relevance(self, x=None,features_coord=None):
         """
-
+               Compute features relevance considering the values of the eigenvector associated with the largest eigenvalue in a PCA transformation.
                Args:
                    x: features dataset not transposed
                    features_coord: features coordinates retrieved using a features extraction method
@@ -114,7 +124,7 @@ class DeepInsight(ParamImageMethod):
                    Array that contains features relevance
         """
         pca_fitted=PCA(n_components=2,random_state=self._random_seed).fit(x)
-        features_relevance = MinMaxScaler().fit_transform(pca_fitted.components_[0,:].reshape(-1,1)) + 1
+        features_relevance = MinMaxScaler().fit_transform(np.absolute(pca_fitted.components_[0,:].reshape(-1,1))) + 1
         return features_relevance
 
     def _get_features_coords(self, x):
@@ -129,4 +139,4 @@ class DeepInsight(ParamImageMethod):
         features_coord = self.__apply_dimensionality_reduction(x)
         # Apply Convex Hull algorithm to get limit points, calculate mimimun rectangle area and rotate
         rotmat, rect_coords, limit_points = get_minimum_rectangle(features_coord)
-        return np.dot(rotmat, features_coord.T).T
+        return np.dot(rotmat, features_coord.T).T

{tintolib-1.2.1 → tintolib-1.3.0}/TINTOlib/fotomics.py

@@ -140,19 +140,18 @@ class Fotomics(ParamImageMethod):
         lower_limits = (percentiles[:, 0] - iqrs).reshape(-1, 1)
         upper_limits = (percentiles[:, 1] + iqrs).reshape(-1, 1)
         x_t = x.T
-        match self.__outliers_treatment:
-            case constants.avg_option:
-                subs = np.tile(np.mean(x, axis=0), (x.shape[0], 1)).T
-                x_t = np.where(x_t >= lower_limits, x_t, subs)
-                x_t = np.where(x_t <= upper_limits, x_t, subs)
-            case constants.max_min_option:
-                max = np.tile(lower_limits.T, (x.shape[0], 1)).T
-                min = np.tile(upper_limits.T, (x.shape[0], 1)).T
-                x_t = np.where(x_t >= lower_limits, x_t, max)
-                x_t = np.where(x_t <= upper_limits, x_t, min)
-            case constants.zero_option:
-                x_t = np.where(x_t >= lower_limits, x_t, 0)
-                x_t = np.where(x_t <= upper_limits, x_t, 0)
+        if self.__outliers_treatment == constants.avg_option:
+            subs = np.tile(np.mean(x, axis=0), (x.shape[0], 1)).T
+            x_t = np.where(x_t >= lower_limits, x_t, subs)
+            x_t = np.where(x_t <= upper_limits, x_t, subs)
+        elif self.__outliers_treatment == constants.max_min_option:
+            max = np.tile(lower_limits.T, (x.shape[0], 1)).T
+            min = np.tile(upper_limits.T, (x.shape[0], 1)).T
+            x_t = np.where(x_t >= lower_limits, x_t, max)
+            x_t = np.where(x_t <= upper_limits, x_t, min)
+        elif self.__outliers_treatment == constants.zero_option:
+            x_t = np.where(x_t >= lower_limits, x_t, 0)
+            x_t = np.where(x_t <= upper_limits, x_t, 0)
         return x_t.T
 
     def _get_features_coords(self,x):

{tintolib-1.2.1 → tintolib-1.3.0}/TINTOlib/igtd.py

@@ -8,7 +8,9 @@ import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
 from scipy.spatial.distance import pdist, squareform
-from scipy.stats import rankdata, spearmanr
+from scipy.stats import rankdata, spearmanr, rankdata, wasserstein_distance
+from sklearn.manifold import Isomap
+from sklearn.metrics.pairwise import pairwise_distances
 
 # Typing imports
 from typing import List, Optional, Union
@@ -17,6 +19,7 @@ from sklearn.preprocessing import MinMaxScaler
 
 # Local application/library imports
 from TINTOlib.mappingMethod import MappingMethod
+from TINTOlib.utils.geometry import tropical_distance
 
 
 ###########################################################
@@ -187,6 +190,24 @@ class IGTD(MappingMethod):
             corr1 = np.dot(np.transpose(data), data)
             corr2 = data.shape[0] - np.dot(np.transpose(1 - data), 1 - data)
             corr = corr1 / corr2
+        elif self.fea_dist_method == 'Wasserstein':
+            corr = squareform(pdist(np.transpose(data), lambda u, v: wasserstein_distance(u, v)))
+            corr = np.max(corr) - corr
+            corr = corr / np.max(corr)
+        elif self.fea_dist_method == 'Jensen':
+            corr = squareform(pdist(np.transpose(data), metric='jensenshannon'))
+            corr = np.max(corr) - corr
+            corr = corr / np.max(corr)
+        elif self.fea_dist_method == 'Geodesic':
+            isomap = Isomap(n_components = 3)
+            embedding = isomap.fit_transform(np.transpose(data))
+            corr = pairwise_distances(embedding, metric = 'euclidean')
+            corr = np.max(corr) - corr
+            corr = corr / np.max(corr)
+        elif self.fea_dist_method == 'Tropical':
+            corr = squareform(pdist(np.transpose(data), metric= tropical_distance))
+            corr = np.max(corr) - corr
+            corr = corr / np.max(corr)
 
         corr = 1 - corr
         corr = np.around(a=corr, decimals=10)

{tintolib-1.2.1 → tintolib-1.3.0}/TINTOlib/paramImageMethod.py

@@ -156,13 +156,12 @@ class ParamImageMethod(MappingMethod):
         Returns:
             imgs_coord: Matrix with pixel values by averaging features values
         """
-        match self._group_method:
-            case constants.avg_option:
-                return self._avg_features_by_pixels(x)
-            case constants.relevance_option:
-                return self._rev_features_by_pixels(x)
-            case _:
-                return self._avg_features_by_pixels(x)
+        if self._group_method == constants.avg_option:
+            return self._avg_features_by_pixels(x)
+        elif self._group_method == constants.relevance_option:
+            return self._rev_features_by_pixels(x)
+        else:
+            return self._avg_features_by_pixels(x)
 
 
     def _avg_features_by_pixels(self,x):
@@ -180,23 +179,19 @@ class ParamImageMethod(MappingMethod):
 
     def _rev_features_by_pixels(self,x):
         """
-        Calculate the pixel values by weighted sum of the value of each feature and its relevance
+        Calculate the pixel values by relevance weighted average of the features values
         Args:
             x:Dataset
 
         Returns:
-
+            pixel_values: Dataframe with pixel values calculated and its positions
         """
-
         features_info=np.hstack((self._features_positions, self._features_relevance))
         features_dot_rev=(x.T)*(features_info[:,2].reshape(-1,1))
         rev_sum = (pd.DataFrame(features_info).groupby([0, 1], as_index=False).transform('sum')).to_numpy()
         features_rev=features_dot_rev/rev_sum
-        pixels_sum=(pd.DataFrame(np.hstack((self._features_positions, features_rev))).groupby([0, 1], as_index=False).sum()).to_numpy()
-        rev_count = (pd.DataFrame(features_info).groupby([0, 1], as_index=False).count()).to_numpy()
-        pixels_values=pixels_sum[:,2:]/rev_count[:,2].reshape(-1,1)
-        imgs_coord=pd.DataFrame(np.hstack((pixels_sum[:,:2], pixels_values)))
-        return imgs_coord
+        pixels_values=(pd.DataFrame(np.hstack((self._features_positions, features_rev))).groupby([0, 1], as_index=False).sum())
+        return pixels_values
 
 
     def _create_images(self,imgs_coord,y):