likelihood 1.2.23__tar.gz → 1.2.25__tar.gz

{likelihood-1.2.23 → likelihood-1.2.25}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: likelihood
-Version: 1.2.23
+Version: 1.2.25
 Summary: A package that performs the maximum likelihood algorithm.
 Home-page: https://github.com/jzsmoreno/likelihood/
 Author: J. A. Moreno-Guerra
@@ -13,7 +13,7 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: black[jupyter]==24.1.1
+Requires-Dist: black[jupyter]>=24.3.0
 Requires-Dist: mypy-extensions==1.0.0
 Requires-Dist: types-openpyxl==3.1.0.15
 Requires-Dist: pydocstyle==6.3.0
@@ -31,6 +31,18 @@ Requires-Dist: pyvis; extra == "full"
 Requires-Dist: tensorflow==2.15.0; extra == "full"
 Requires-Dist: keras-tuner; extra == "full"
 Requires-Dist: scikit-learn; extra == "full"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: maintainer
+Dynamic: maintainer-email
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 ![likelihood](https://raw.githubusercontent.com/RodolfoFerro/likelihood/main/likelihood.png)
 

{likelihood-1.2.23 → likelihood-1.2.25}/likelihood/graph/nn.py

@@ -1,7 +1,9 @@
+import logging
 import os
 
-os.environ["TF_ENABLE_ONEDNN_OPTS"] = "0"
-import logging
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
+logging.getLogger("tensorflow").setLevel(logging.ERROR)
+
 import warnings
 from typing import List, Tuple
 
@@ -9,19 +11,16 @@ import numpy as np
 import pandas as pd
 import tensorflow as tf
 from IPython.display import clear_output
-from numpy import ndarray
 from pandas.core.frame import DataFrame
 from sklearn.metrics import f1_score
 from sklearn.model_selection import train_test_split
 
 from likelihood.tools import generate_feature_yaml
 
-logging.getLogger("tensorflow").setLevel(logging.ERROR)
-
-tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)
+tf.get_logger().setLevel("ERROR")
 
 
-def compare_similarity(arr1: ndarray, arr2: ndarray) -> int:
+def compare_similarity(arr1: np.ndarray, arr2: np.ndarray) -> int:
     """Compares the similarity between two arrays of categories.
 
     Parameters
@@ -44,9 +43,9 @@ def compare_similarity(arr1: ndarray, arr2: ndarray) -> int:
     return count
 
 
-def cal_adjency_matrix(
+def cal_adjacency_matrix(
     df: DataFrame, exclude_subset: List[str] = [], sparse: bool = True, **kwargs
-) -> Tuple[dict, ndarray]:
+) -> Tuple[dict, np.ndarray]:
     """Calculates the adjacency matrix for a given DataFrame.
     The adjacency matrix is a matrix that represents the similarity between each pair of categories.
     The similarity is calculated using the `compare_similarity` function.
@@ -133,7 +132,7 @@ class Data:
         target: str | None = None,
         exclude_subset: List[str] = [],
     ):
-        _, adjacency = cal_adjency_matrix(df, exclude_subset=exclude_subset, sparse=True)
+        _, adjacency = cal_adjacency_matrix(df, exclude_subset=exclude_subset, sparse=True)
         if target is not None:
             X = df.drop(columns=[target] + exclude_subset)
         else:

likelihood-1.2.25/likelihood/models/deep/autoencoders.py

@@ -0,0 +1,598 @@
+import logging
+import os
+import random
+from functools import partial
+from shutil import rmtree
+
+import matplotlib
+import matplotlib.colors as mcolors
+import matplotlib.pyplot as plt
+import numpy as np
+import pandas as pd
+from pandas.plotting import radviz
+
+os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
+logging.getLogger("tensorflow").setLevel(logging.ERROR)
+
+import warnings
+from functools import wraps
+
+import keras_tuner
+import tensorflow as tf
+from pandas.core.frame import DataFrame
+from sklearn.manifold import TSNE
+
+from likelihood.tools import OneHotEncoder
+
+tf.get_logger().setLevel("ERROR")
+
+
+def suppress_warnings(func):
+    @wraps(func)
+    def wrapper(*args, **kwargs):
+        with warnings.catch_warnings():
+            warnings.simplefilter("ignore")
+            return func(*args, **kwargs)
+
+    return wrapper
+
+
+@tf.keras.utils.register_keras_serializable(package="Custom", name="AutoClassifier")
+class AutoClassifier(tf.keras.Model):
+    """
+    An auto-classifier model that automatically determines the best classification strategy based on the input data.
+
+    Attributes:
+        - input_shape_parm: The shape of the input data.
+        - num_classes: The number of classes in the dataset.
+        - units: The number of neurons in each hidden layer.
+        - activation: The type of activation function to use for the neural network layers.
+
+    Methods:
+        __init__(self, input_shape_parm, num_classes, units, activation): Initializes an AutoClassifier instance with the given parameters.
+        build(self, input_shape_parm): Builds the model architecture based on input_shape_parm.
+        call(self, x): Defines the forward pass of the model.
+        get_config(self): Returns the configuration of the model.
+        from_config(cls, config): Recreates an instance of AutoClassifier from its configuration.
+    """
+
+    def __init__(self, input_shape_parm, num_classes, units, activation, **kwargs):
+        """
+        Initializes an AutoClassifier instance with the given parameters.
+
+        Parameters
+        ----------
+        input_shape_parm : `int`
+            The shape of the input data.
+        num_classes : `int`
+            The number of classes in the dataset.
+        units : `int`
+            The number of neurons in each hidden layer.
+        activation : `str`
+            The type of activation function to use for the neural network layers.
+
+        Keyword Arguments:
+        ----------
+        Additional keyword arguments to pass to the model.
+
+        classifier_activation : `str`
+            The activation function to use for the classifier layer. Default is "softmax". If the activation function is not a classification function, the model can be used in regression problems.
+        num_layers : `int`
+            The number of hidden layers in the classifier. Default is 1.
+        """
+        super(AutoClassifier, self).__init__()
+        self.input_shape_parm = input_shape_parm
+        self.num_classes = num_classes
+        self.units = units
+        self.activation = activation
+
+        self.encoder = None
+        self.decoder = None
+        self.classifier = None
+        self.classifier_activation = kwargs.get("classifier_activation", "softmax")
+        self.num_layers = kwargs.get("num_layers", 1)
+
+    def build(self, input_shape):
+        self.encoder = tf.keras.Sequential(
+            [
+                tf.keras.layers.Dense(units=self.units, activation=self.activation),
+                tf.keras.layers.Dense(units=int(self.units / 2), activation=self.activation),
+            ]
+        )
+
+        self.decoder = tf.keras.Sequential(
+            [
+                tf.keras.layers.Dense(units=self.units, activation=self.activation),
+                tf.keras.layers.Dense(units=self.input_shape_parm, activation=self.activation),
+            ]
+        )
+
+        self.classifier = tf.keras.Sequential()
+        if self.num_layers > 1:
+            for _ in range(self.num_layers - 1):
+                self.classifier.add(
+                    tf.keras.layers.Dense(units=self.units, activation=self.activation)
+                )
+        self.classifier.add(
+            tf.keras.layers.Dense(units=self.num_classes, activation=self.classifier_activation)
+        )
+
+    def call(self, x):
+        encoded = self.encoder(x)
+        decoded = self.decoder(encoded)
+        combined = tf.concat([decoded, encoded], axis=1)
+        classification = self.classifier(combined)
+        return classification
+
+    def get_config(self):
+        config = {
+            "input_shape_parm": self.input_shape_parm,
+            "num_classes": self.num_classes,
+            "units": self.units,
+            "activation": self.activation,
+            "classifier_activation": self.classifier_activation,
+            "num_layers": self.num_layers,
+        }
+        base_config = super(AutoClassifier, self).get_config()
+        return dict(list(base_config.items()) + list(config.items()))
+
+    @classmethod
+    def from_config(cls, config):
+        return cls(
+            input_shape_parm=config["input_shape_parm"],
+            num_classes=config["num_classes"],
+            units=config["units"],
+            activation=config["activation"],
+            classifier_activation=config["classifier_activation"],
+            num_layers=config["num_layers"],
+        )
+
+
+def call_existing_code(
+    units: int,
+    activation: str,
+    threshold: float,
+    optimizer: str,
+    input_shape_parm: None | int = None,
+    num_classes: None | int = None,
+    num_layers: int = 1,
+) -> AutoClassifier:
+    """
+    Calls an existing AutoClassifier instance.
+
+    Parameters
+    ----------
+    units : `int`
+        The number of neurons in each hidden layer.
+    activation : `str`
+        The type of activation function to use for the neural network layers.
+    threshold : `float`
+        The threshold for the classifier.
+    optimizer : `str`
+        The type of optimizer to use for the neural network layers.
+    input_shape_parm : `None` | `int`
+        The shape of the input data.
+    num_classes : `int`
+        The number of classes in the dataset.
+
+    Returns
+    -------
+    `AutoClassifier`
+        The AutoClassifier instance.
+    """
+    model = AutoClassifier(
+        input_shape_parm=input_shape_parm,
+        num_classes=num_classes,
+        units=units,
+        activation=activation,
+        num_layers=num_layers,
+    )
+    model.compile(
+        optimizer=optimizer,
+        loss=tf.keras.losses.CategoricalCrossentropy(),
+        metrics=[tf.keras.metrics.F1Score(threshold=threshold)],
+    )
+    return model
+
+
+def build_model(
+    hp, input_shape_parm: None | int, num_classes: None | int, **kwargs
+) -> AutoClassifier:
+    """Builds a neural network model using Keras Tuner's search algorithm.
+
+    Parameters
+    ----------
+    hp : `keras_tuner.HyperParameters`
+        The hyperparameters to tune.
+    input_shape_parm : `None` | `int`
+        The shape of the input data.
+    num_classes : `int`
+        The number of classes in the dataset.
+
+    Keyword Arguments:
+    ----------
+    Additional keyword arguments to pass to the model.
+
+    hyperparameters : `dict`
+        The hyperparameters to set.
+
+    Returns
+    -------
+    `keras.Model`
+        The neural network model.
+    """
+    hyperparameters = kwargs.get("hyperparameters", None)
+    hyperparameters_keys = hyperparameters.keys() if hyperparameters is not None else []
+
+    units = (
+        hp.Int(
+            "units",
+            min_value=int(input_shape_parm * 0.2),
+            max_value=int(input_shape_parm * 1.5),
+            step=2,
+        )
+        if "units" not in hyperparameters_keys
+        else hyperparameters["units"]
+    )
+    activation = (
+        hp.Choice("activation", ["sigmoid", "relu", "tanh", "selu", "softplus", "softsign"])
+        if "activation" not in hyperparameters_keys
+        else hyperparameters["activation"]
+    )
+    optimizer = (
+        hp.Choice("optimizer", ["sgd", "adam", "adadelta", "rmsprop", "adamax", "adagrad"])
+        if "optimizer" not in hyperparameters_keys
+        else hyperparameters["optimizer"]
+    )
+    threshold = (
+        hp.Float("threshold", min_value=0.1, max_value=0.9, sampling="log")
+        if "threshold" not in hyperparameters_keys
+        else hyperparameters["threshold"]
+    )
+    num_layers = (
+        hp.Int("num_layers", min_value=1, max_value=10, step=1)
+        if "num_layers" not in hyperparameters_keys
+        else hyperparameters["num_layers"]
+    )
+
+    model = call_existing_code(
+        units=units,
+        activation=activation,
+        threshold=threshold,
+        optimizer=optimizer,
+        input_shape_parm=input_shape_parm,
+        num_classes=num_classes,
+        num_layers=num_layers,
+    )
+    return model
+
+
+@suppress_warnings
+def setup_model(
+    data: DataFrame,
+    target: str,
+    epochs: int,
+    train_size: float = 0.7,
+    seed=None,
+    train_mode: bool = True,
+    filepath: str = "./my_dir/best_model",
+    method: str = "Hyperband",
+    **kwargs,
+) -> AutoClassifier:
+    """Setup model for training and tuning.
+
+    Parameters
+    ----------
+    data : `DataFrame`
+        The dataset to train the model on.
+    target : `str`
+        The name of the target column.
+    epochs : `int`
+        The number of epochs to train the model for.
+    train_size : `float`
+        The proportion of the dataset to use for training.
+    seed : `Any` | `int`
+        The random seed to use for reproducibility.
+    train_mode : `bool`
+        Whether to train the model or not.
+    filepath : `str`
+        The path to save the best model to.
+    method : `str`
+        The method to use for hyperparameter tuning. Options are "Hyperband" and "RandomSearch".
+
+    Keyword Arguments:
+    ----------
+    Additional keyword arguments to pass to the model.
+
+    max_trials : `int`
+        The maximum number of trials to perform.
+    directory : `str`
+        The directory to save the model to.
+    project_name : `str`
+        The name of the project.
+    objective : `str`
+        The objective to optimize.
+    verbose : `bool`
+        Whether to print verbose output.
+    hyperparameters : `dict`
+        The hyperparameters to set.
+
+    Returns
+    -------
+    model : `AutoClassifier`
+        The trained model.
+    """
+    max_trials = kwargs.get("max_trials", 10)
+    directory = kwargs.get("directory", "./my_dir")
+    project_name = kwargs.get("project_name", "get_best")
+    objective = kwargs.get("objective", "val_loss")
+    verbose = kwargs.get("verbose", True)
+    hyperparameters = kwargs.get("hyperparameters", None)
+
+    X = data.drop(columns=target)
+    input_sample = X.sample(1)
+    y = data[target]
+    assert (
+        X.select_dtypes(include=["object"]).empty == True
+    ), "Categorical variables within the DataFrame must be encoded, this is done by using the DataFrameEncoder from likelihood."
+    validation_split = 1.0 - train_size
+
+    if train_mode:
+        try:
+            if (not os.path.exists(directory)) and directory != "./":
+                os.makedirs(directory)
+            elif directory != "./":
+                print(f"Directory {directory} already exists, it will be deleted.")
+                rmtree(directory)
+                os.makedirs(directory)
+        except:
+            print("Warning: unable to create directory")
+
+        y_encoder = OneHotEncoder()
+        y = y_encoder.encode(y.to_list())
+        X = X.to_numpy()
+        input_sample.to_numpy()
+        X = np.asarray(X).astype(np.float32)
+        input_sample = np.asarray(input_sample).astype(np.float32)
+        y = np.asarray(y).astype(np.float32)
+
+        input_shape_parm = X.shape[1]
+        num_classes = y.shape[1]
+        global build_model
+        build_model = partial(
+            build_model,
+            input_shape_parm=input_shape_parm,
+            num_classes=num_classes,
+            hyperparameters=hyperparameters,
+        )
+
+        if method == "Hyperband":
+            tuner = keras_tuner.Hyperband(
+                hypermodel=build_model,
+                objective=objective,
+                max_epochs=epochs,
+                factor=3,
+                directory=directory,
+                project_name=project_name,
+                seed=seed,
+            )
+        elif method == "RandomSearch":
+            tuner = keras_tuner.RandomSearch(
+                hypermodel=build_model,
+                objective=objective,
+                max_trials=max_trials,
+                directory=directory,
+                project_name=project_name,
+                seed=seed,
+            )
+
+        tuner.search(X, y, epochs=epochs, validation_split=validation_split, verbose=verbose)
+        models = tuner.get_best_models(num_models=2)
+        best_model = models[0]
+        best_model(input_sample)
+
+        best_model.save(filepath, save_format="tf")
+
+        if verbose:
+            tuner.results_summary()
+    else:
+        best_model = tf.keras.models.load_model(filepath)
+
+    best_hps = tuner.get_best_hyperparameters(1)[0].values
+    return best_model, pd.DataFrame(best_hps, index=["Value"])
+
+
+class GetInsights:
+    def __init__(self, model: AutoClassifier, inputs: np.ndarray) -> None:
+        self.inputs = inputs
+        self.model = model
+        self.encoder_layer = self.model.encoder.layers[0]
+        self.decoder_layer = self.model.decoder.layers[0]
+        self.classifier_layer = self.model.classifier.layers[-2]
+        self.encoder_weights = self.encoder_layer.get_weights()[0]
+        self.decoder_weights = self.decoder_layer.get_weights()[0]
+        self.classifier_weights = self.classifier_layer.get_weights()[0]
+        colors = dict(mcolors.BASE_COLORS, **mcolors.CSS4_COLORS)
+
+        by_hsv = sorted(
+            (tuple(mcolors.rgb_to_hsv(mcolors.to_rgba(color)[:3])), name)
+            for name, color in colors.items()
+        )
+        self.sorted_names = [name for hsv, name in by_hsv if hsv[1] > 0.4 and hsv[2] >= 0.4]
+        random.shuffle(self.sorted_names)
+
+    def predictor_analyzer(
+        self,
+        frac=None,
+        cmap: str = "viridis",
+        aspect: str = "auto",
+        highlight: bool = True,
+        **kwargs,
+    ) -> None:
+        self._viz_weights(cmap=cmap, aspect=aspect, highlight=highlight, **kwargs)
+        inputs = self.inputs.copy()
+        y_labels = kwargs.get("y_labels", None)
+        if frac:
+            n = int(frac * self.inputs.shape[0])
+            indexes = np.random.choice(np.arange(inputs.shape[0]), n, replace=False)
+            inputs = inputs[indexes]
+        inputs[np.isnan(inputs)] = 0.0
+        encoded = self.model.encoder(inputs)
+        reconstructed = self.model.decoder(encoded)
+        combined = tf.concat([reconstructed, encoded], axis=1)
+        self.classification = self.model.classifier(combined).numpy().argmax(axis=1)
+        ax = plt.subplot(1, 2, 1)
+        plt.imshow(self.inputs, cmap=cmap, aspect=aspect)
+        plt.colorbar()
+        plt.title("Original Data")
+        plt.subplot(1, 2, 2, sharex=ax, sharey=ax)
+        plt.imshow(reconstructed, cmap=cmap, aspect=aspect)
+        plt.colorbar()
+        plt.title("Decoder Layer Reconstruction")
+        plt.show()
+
+        self._get_tsne_repr(inputs=inputs, frac=frac)
+        self._viz_tsne_repr(c=self.classification)
+
+        self.data = pd.DataFrame(encoded, columns=[f"Feature {i}" for i in range(encoded.shape[1])])
+        self.data_input = pd.DataFrame(
+            inputs,
+            columns=(
+                [f"Feature {i}" for i in range(inputs.shape[1])] if y_labels is None else y_labels
+            ),
+        )
+        self.data["class"] = self.classification
+        self.data_input["class"] = self.classification
+        radviz(self.data, "class", color=self.colors)
+        plt.title("Radviz Visualization of Latent Space")
+        plt.show()
+
+        radviz(self.data_input, "class", color=self.colors)
+        plt.title("Radviz Visualization of Input Data")
+        plt.show()
+        return self._statistics(self.data_input)
+
+    def _statistics(self, data_input: DataFrame, **kwargs) -> DataFrame:
+        data = data_input.copy(deep=True)
+
+        if not pd.api.types.is_string_dtype(data["class"]):
+            data["class"] = data["class"].astype(str)
+
+        data.ffill(inplace=True)
+        grouped_data = data.groupby("class")
+
+        numerical_stats = grouped_data.agg(["mean", "min", "max", "std", "median"])
+        numerical_stats.columns = ["_".join(col).strip() for col in numerical_stats.columns.values]
+
+        def get_mode(x):
+            mode_series = x.mode()
+            return mode_series.iloc[0] if not mode_series.empty else None
+
+        mode_stats = grouped_data.apply(get_mode, include_groups=False)
+        mode_stats.columns = [f"{col}_mode" for col in mode_stats.columns]
+        combined_stats = pd.concat([numerical_stats, mode_stats], axis=1)
+
+        return combined_stats.T
+
+    def _viz_weights(
+        self, cmap: str = "viridis", aspect: str = "auto", highlight: bool = True, **kwargs
+    ) -> None:
+        title = kwargs.get("title", "Encoder Layer Weights (Dense Layer)")
+        y_labels = kwargs.get("y_labels", None)
+        cmap_highlight = kwargs.get("cmap_highlight", "Pastel1")
+        highlight_mask = np.zeros_like(self.encoder_weights, dtype=bool)
+
+        plt.imshow(self.encoder_weights, cmap=cmap, aspect=aspect)
+        plt.colorbar()
+        plt.title(title)
+        if y_labels is not None:
+            plt.yticks(ticks=np.arange(self.encoder_weights.shape[0]), labels=y_labels)
+        if highlight:
+            for i, j in enumerate(self.encoder_weights.argmax(axis=1)):
+                highlight_mask[i, j] = True
+            plt.imshow(
+                np.ma.masked_where(~highlight_mask, self.encoder_weights),
+                cmap=cmap_highlight,
+                alpha=0.5,
+                aspect=aspect,
+            )
+        plt.show()
+
+    def _get_tsne_repr(self, inputs=None, frac=None) -> None:
+        if inputs is None:
+            inputs = self.inputs.copy()
+            if frac:
+                n = int(frac * self.inputs.shape[0])
+                indexes = np.random.choice(np.arange(inputs.shape[0]), n, replace=False)
+                inputs = inputs[indexes]
+            inputs[np.isnan(inputs)] = 0.0
+        self.latent_representations = inputs @ self.encoder_weights
+
+        tsne = TSNE(n_components=2)
+        self.reduced_data_tsne = tsne.fit_transform(self.latent_representations)
+
+    def _viz_tsne_repr(self, **kwargs) -> None:
+        c = kwargs.get("c", None)
+        self.colors = (
+            kwargs.get("colors", self.sorted_names[: len(np.unique(c))]) if c is not None else None
+        )
+        plt.scatter(
+            self.reduced_data_tsne[:, 0],
+            self.reduced_data_tsne[:, 1],
+            cmap=matplotlib.colors.ListedColormap(self.colors) if c is not None else None,
+            c=c,
+        )
+        if c is not None:
+            cb = plt.colorbar()
+            loc = np.arange(0, max(c), max(c) / float(len(self.colors)))
+            cb.set_ticks(loc)
+            cb.set_ticklabels(np.unique(c))
+        plt.title("t-SNE Visualization of Latent Space")
+        plt.xlabel("t-SNE 1")
+        plt.ylabel("t-SNE 2")
+        plt.show()
+
+
+########################################################################################
+
+if __name__ == "__main__":
+    # Example usage
+    import pandas as pd
+    from sklearn.datasets import load_iris
+    from sklearn.preprocessing import OneHotEncoder
+
+    # Load the dataset
+    iris = load_iris()
+
+    # Convert to a DataFrame for easy exploration
+    iris_df = pd.DataFrame(data=iris.data, columns=iris.feature_names)
+    iris_df["species"] = iris.target
+
+    X = iris_df.drop(columns="species")
+    y_labels = X.columns
+    X = X.values
+    y = iris_df["species"].values
+
+    X = np.asarray(X).astype(np.float32)
+
+    encoder = OneHotEncoder()
+    y = encoder.fit_transform(y.reshape(-1, 1)).toarray()
+    y = np.asarray(y).astype(np.float32)
+
+    model = AutoClassifier(
+        input_shape_parm=X.shape[1], num_classes=3, units=27, activation="selu", num_layers=2
+    )
+    model.compile(
+        optimizer="adam",
+        loss=tf.keras.losses.CategoricalCrossentropy(),
+        metrics=[tf.keras.metrics.F1Score(threshold=0.5)],
+    )
+    model.fit(X, y, epochs=50, validation_split=0.2)
+
+    insights = GetInsights(model, X)
+    summary = insights.predictor_analyzer(frac=1.0, y_labels=y_labels)
+    insights._get_tsne_repr()
+    insights._viz_tsne_repr()
+    insights._viz_tsne_repr(c=iris_df["species"])
+    insights._viz_weights()
+    print(summary)

likelihood-1.2.25/likelihood/models/hmm.py

@@ -0,0 +1,163 @@
+import logging
+import os
+import pickle
+from typing import List, Tuple
+
+import numpy as np
+from IPython.display import clear_output
+
+
+class HMM:
+    def __init__(self, n_states: int, n_observations: int):
+        self.n_states = n_states
+        self.n_observations = n_observations
+
+        # Initialize parameters with random values
+        self.pi = np.random.dirichlet(np.ones(n_states), size=1)[0]
+        self.A = np.random.dirichlet(np.ones(n_states), size=n_states)
+        self.B = np.random.dirichlet(np.ones(n_observations), size=n_states)
+
+    def save_model(self, filename: str = "./hmm") -> None:
+        filename = filename if filename.endswith(".pkl") else filename + ".pkl"
+        with open(filename, "wb") as f:
+            pickle.dump(self, f)
+
+    @staticmethod
+    def load_model(filename: str = "./hmm") -> "HMM":
+        filename = filename + ".pkl" if not filename.endswith(".pkl") else filename
+        with open(filename, "rb") as f:
+            return pickle.load(f)
+
+    def forward(self, sequence: List[int]) -> np.ndarray:
+        T = len(sequence)
+        alpha = np.zeros((T, self.n_states))
+
+        # Add a small constant (smoothing) to avoid log(0)
+        epsilon = 1e-10  # Small value to avoid taking log(0)
+
+        # Initialization (log-space)
+        alpha[0] = np.log(self.pi + epsilon) + np.log(self.B[:, sequence[0]] + epsilon)
+        alpha[0] -= np.log(np.sum(np.exp(alpha[0])))  # Normalization (log-space)
+
+        # Recursion (log-space)
+        for t in range(1, T):
+            for i in range(self.n_states):
+                alpha[t, i] = np.log(
+                    np.sum(np.exp(alpha[t - 1] + np.log(self.A[:, i] + epsilon)))
+                ) + np.log(self.B[i, sequence[t]] + epsilon)
+            alpha[t] -= np.log(np.sum(np.exp(alpha[t])))  # Normalization
+
+        return alpha
+
+    def backward(self, sequence: List[int]) -> np.ndarray:
+        T = len(sequence)
+        beta = np.ones((T, self.n_states))
+
+        # Backward recursion
+        for t in range(T - 2, -1, -1):
+            for i in range(self.n_states):
+                beta[t, i] = np.sum(self.A[i] * self.B[:, sequence[t + 1]] * beta[t + 1])
+
+        return beta
+
+    def viterbi(self, sequence: List[int]) -> np.ndarray:
+        T = len(sequence)
+        delta = np.zeros((T, self.n_states))
+        psi = np.zeros((T, self.n_states), dtype=int)
+
+        # Initialization
+        delta[0] = self.pi * self.B[:, sequence[0]]
+
+        # Recursion
+        for t in range(1, T):
+            for i in range(self.n_states):
+                delta[t, i] = np.max(delta[t - 1] * self.A[:, i]) * self.B[i, sequence[t]]
+                psi[t, i] = np.argmax(delta[t - 1] * self.A[:, i])
+
+        # Reconstruct the most probable path
+        state_sequence = np.zeros(T, dtype=int)
+        state_sequence[T - 1] = np.argmax(delta[T - 1])
+        for t in range(T - 2, -1, -1):
+            state_sequence[t] = psi[t + 1, state_sequence[t + 1]]
+
+        return state_sequence
+
+    def baum_welch(
+        self, sequences: List[List[int]], n_iterations: int, verbose: bool = False
+    ) -> None:
+        for iteration in range(n_iterations):
+            # Initialize accumulators
+            A_num = np.zeros((self.n_states, self.n_states))
+            B_num = np.zeros((self.n_states, self.n_observations))
+            pi_num = np.zeros(self.n_states)
+
+            for sequence in sequences:
+                T = len(sequence)
+                alpha = self.forward(sequence)
+                beta = self.backward(sequence)
+
+                # Update pi
+                gamma = (alpha * beta) / np.sum(alpha * beta, axis=1, keepdims=True)
+                pi_num += gamma[0]
+
+                # Update A and B
+                for t in range(T - 1):
+                    xi = np.zeros((self.n_states, self.n_states))
+                    denom = np.sum(alpha[t] * self.A * self.B[:, sequence[t + 1]] * beta[t + 1])
+
+                    for i in range(self.n_states):
+                        for j in range(self.n_states):
+                            xi[i, j] = (
+                                alpha[t, i]
+                                * self.A[i, j]
+                                * self.B[j, sequence[t + 1]]
+                                * beta[t + 1, j]
+                            ) / denom
+                        A_num[i] += xi[i]
+
+                    B_num[:, sequence[t]] += gamma[t]
+
+                # For the last step of the sequence
+                B_num[:, sequence[-1]] += gamma[-1]
+
+            # Normalize and update parameters
+            self.pi = pi_num / len(sequences)
+            self.A = A_num / np.sum(A_num, axis=1, keepdims=True)
+            self.B = B_num / np.sum(B_num, axis=1, keepdims=True)
+
+            # Logging parameters every 10 iterations
+            if iteration % 10 == 0 and verbose:
+                os.system("cls" if os.name == "nt" else "clear")
+                clear_output(wait=True)
+                logging.info(f"Iteration {iteration}:")
+                logging.info("Pi: %s", self.pi)
+                logging.info("A:\n%s", self.A)
+                logging.info("B:\n%s", self.B)
+
+    def decoding_accuracy(self, sequences: List[List[int]], true_states: List[List[int]]) -> float:
+        correct_predictions = 0
+        total_predictions = 0
+
+        for sequence, true_state in zip(sequences, true_states):
+            predicted_states = self.viterbi(sequence)
+            correct_predictions += np.sum(predicted_states == true_state)
+            total_predictions += len(sequence)
+
+        accuracy = (correct_predictions / total_predictions) * 100
+        return accuracy
+
+    def state_probabilities(self, sequence: List[int]) -> np.ndarray:
+        """
+        Returns the smoothed probabilities of the hidden states at each time step.
+        This is done by using both forward and backward probabilities.
+        """
+        alpha = self.forward(sequence)
+        beta = self.backward(sequence)
+
+        # Compute smoothed probabilities (gamma)
+        smoothed_probs = (alpha * beta) / np.sum(alpha * beta, axis=1, keepdims=True)
+
+        return smoothed_probs
+
+    def sequence_probability(self, sequence: List[int]) -> np.ndarray:
+        return self.state_probabilities(sequence)[-1]

{likelihood-1.2.23 → likelihood-1.2.25}/likelihood/models/simulation.py

@@ -5,7 +5,6 @@ from typing import List, Tuple, Union
 import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
-from numpy import ndarray
 from pandas.core.frame import DataFrame
 
 from likelihood.tools import DataScaler, FeatureSelection, OneHotEncoder, cdf, check_nan_inf
@@ -66,12 +65,12 @@ class SimulationEngine(FeatureSelection):
 
         super().__init__(**kwargs)
 
-    def predict(self, df: DataFrame, column: str) -> ndarray | list:
+    def predict(self, df: DataFrame, column: str) -> np.ndarray | list:
         # Let us assign the dictionary entries corresponding to the column
         w, quick_encoder, names_cols, dfe, numeric_dict = self.w_dict[column]
 
         df = df[names_cols].copy()
-        # Change the scale of the dataframe
+        # Change the scale of the DataFrame
         dataset = self.df.copy()
         dataset.drop(columns=column, inplace=True)
         numeric_df = dataset.select_dtypes(include="number")
@@ -85,7 +84,7 @@ class SimulationEngine(FeatureSelection):
             for col in numeric_df.columns:
                 df[col] = numeric_df[col].values
 
-        # Encoding the datadrame
+        # Encoding the DataFrame
         for num, colname in enumerate(dfe._encode_columns):
             if df[colname].dtype == "object":
                 encode_dict = dfe.encoding_list[num]
@@ -93,7 +92,7 @@ class SimulationEngine(FeatureSelection):
                     dfe._code_transformation_to, dictionary_list=encode_dict
                 )
 
-        # PREDICTION
+        # Prediction
         y = df.to_numpy() @ w
 
         # Categorical column
@@ -113,7 +112,7 @@ class SimulationEngine(FeatureSelection):
 
         return y[:]
 
-    def _encode(self, df: DataFrame) -> ndarray | list:
+    def _encode(self, df: DataFrame) -> np.ndarray | list:
         df = df.copy()
         column = df.columns[0]
         frec = df[column].value_counts() / len(df)

{likelihood-1.2.23 → likelihood-1.2.25}/likelihood.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.2
 Name: likelihood
-Version: 1.2.23
+Version: 1.2.25
 Summary: A package that performs the maximum likelihood algorithm.
 Home-page: https://github.com/jzsmoreno/likelihood/
 Author: J. A. Moreno-Guerra
@@ -13,7 +13,7 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.10
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: black[jupyter]==24.1.1
+Requires-Dist: black[jupyter]>=24.3.0
 Requires-Dist: mypy-extensions==1.0.0
 Requires-Dist: types-openpyxl==3.1.0.15
 Requires-Dist: pydocstyle==6.3.0
@@ -31,6 +31,18 @@ Requires-Dist: pyvis; extra == "full"
 Requires-Dist: tensorflow==2.15.0; extra == "full"
 Requires-Dist: keras-tuner; extra == "full"
 Requires-Dist: scikit-learn; extra == "full"
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: maintainer
+Dynamic: maintainer-email
+Dynamic: provides-extra
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 ![likelihood](https://raw.githubusercontent.com/RodolfoFerro/likelihood/main/likelihood.png)
 

{likelihood-1.2.23 → likelihood-1.2.25}/likelihood.egg-info/SOURCES.txt

@@ -12,6 +12,7 @@ likelihood/graph/__init__.py
 likelihood/graph/graph.py
 likelihood/graph/nn.py
 likelihood/models/__init__.py
+likelihood/models/hmm.py
 likelihood/models/regression.py
 likelihood/models/simulation.py
 likelihood/models/utils.py

{likelihood-1.2.23 → likelihood-1.2.25}/likelihood.egg-info/requires.txt

@@ -1,4 +1,4 @@
-black[jupyter]==24.1.1
+black[jupyter]>=24.3.0
 mypy-extensions==1.0.0
 types-openpyxl==3.1.0.15
 pydocstyle==6.3.0

likelihood-1.2.23/likelihood/models/deep/autoencoders.py

@@ -1,309 +0,0 @@
-import logging
-import os
-from functools import partial
-from shutil import rmtree
-
-import keras_tuner
-import numpy as np
-import pandas as pd
-import tensorflow as tf
-from pandas.core.frame import DataFrame
-
-from likelihood.tools import OneHotEncoder
-
-logging.getLogger("tensorflow").setLevel(logging.ERROR)
-
-tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)
-
-
-@tf.keras.utils.register_keras_serializable(package="Custom", name="AutoClassifier")
-class AutoClassifier(tf.keras.Model):
-    """
-    An auto-classifier model that automatically determines the best classification strategy based on the input data.
-
-    Attributes:
-        - input_shape_parm: The shape of the input data.
-        - num_classes: The number of classes in the dataset.
-        - units: The number of neurons in each hidden layer.
-        - activation: The type of activation function to use for the neural network layers.
-
-    Methods:
-        __init__(self, input_shape_parm, num_classes, units, activation): Initializes an AutoClassifier instance with the given parameters.
-        build(self, input_shape_parm): Builds the model architecture based on input_shape_parm.
-        call(self, x): Defines the forward pass of the model.
-        get_config(self): Returns the configuration of the model.
-        from_config(cls, config): Recreates an instance of AutoClassifier from its configuration.
-    """
-
-    def __init__(self, input_shape_parm, num_classes, units, activation):
-        """
-        Initializes an AutoClassifier instance with the given parameters.
-
-        Parameters
-        ----------
-        input_shape_parm : `int`
-            The shape of the input data.
-        num_classes : `int`
-            The number of classes in the dataset.
-        units : `int`
-            The number of neurons in each hidden layer.
-        activation : `str`
-            The type of activation function to use for the neural network layers.
-        """
-        super(AutoClassifier, self).__init__()
-        self.input_shape_parm = input_shape_parm
-        self.num_classes = num_classes
-        self.units = units
-        self.activation = activation
-
-        self.encoder = None
-        self.decoder = None
-        self.classifier = None
-
-    def build(self, input_shape):
-        self.encoder = tf.keras.Sequential(
-            [
-                tf.keras.layers.Dense(units=self.units, activation=self.activation),
-                tf.keras.layers.Dense(units=int(self.units / 2), activation=self.activation),
-            ]
-        )
-
-        self.decoder = tf.keras.Sequential(
-            [
-                tf.keras.layers.Dense(units=self.units, activation=self.activation),
-                tf.keras.layers.Dense(units=self.input_shape_parm, activation=self.activation),
-            ]
-        )
-
-        self.classifier = tf.keras.Sequential(
-            [tf.keras.layers.Dense(self.num_classes, activation="softmax")]
-        )
-
-    def call(self, x):
-        encoded = self.encoder(x)
-        decoded = self.decoder(encoded)
-        combined = tf.concat([decoded, encoded], axis=1)
-        classification = self.classifier(combined)
-        return classification
-
-    def get_config(self):
-        config = {
-            "input_shape_parm": self.input_shape_parm,
-            "num_classes": self.num_classes,
-            "units": self.units,
-            "activation": self.activation,
-        }
-        base_config = super(AutoClassifier, self).get_config()
-        return dict(list(base_config.items()) + list(config.items()))
-
-    @classmethod
-    def from_config(cls, config):
-        return cls(
-            input_shape_parm=config["input_shape_parm"],
-            num_classes=config["num_classes"],
-            units=config["units"],
-            activation=config["activation"],
-        )
-
-
-def call_existing_code(
-    units: int,
-    activation: str,
-    threshold: float,
-    optimizer: str,
-    input_shape_parm: None | int = None,
-    num_classes: None | int = None,
-) -> AutoClassifier:
-    """
-    Calls an existing AutoClassifier instance.
-
-    Parameters
-    ----------
-    units : `int`
-        The number of neurons in each hidden layer.
-    activation : `str`
-        The type of activation function to use for the neural network layers.
-    threshold : `float`
-        The threshold for the classifier.
-    optimizer : `str`
-        The type of optimizer to use for the neural network layers.
-    input_shape_parm : `None` | `int`
-        The shape of the input data.
-    num_classes : `int`
-        The number of classes in the dataset.
-
-    Returns
-    -------
-    `AutoClassifier`
-        The AutoClassifier instance.
-    """
-    model = AutoClassifier(
-        input_shape_parm=input_shape_parm,
-        num_classes=num_classes,
-        units=units,
-        activation=activation,
-    )
-    model.compile(
-        optimizer=optimizer,
-        loss=tf.keras.losses.CategoricalCrossentropy(),
-        metrics=[tf.keras.metrics.F1Score(threshold=threshold)],
-    )
-    return model
-
-
-def build_model(hp, input_shape_parm: None | int, num_classes: None | int) -> AutoClassifier:
-    """Builds a neural network model using Keras Tuner's search algorithm.
-
-    Parameters
-    ----------
-    hp : `keras_tuner.HyperParameters`
-        The hyperparameters to tune.
-    input_shape_parm : `None` | `int`
-        The shape of the input data.
-    num_classes : `int`
-        The number of classes in the dataset.
-
-    Returns
-    -------
-    `keras.Model`
-        The neural network model.
-    """
-    units = hp.Int(
-        "units", min_value=int(input_shape_parm * 0.2), max_value=input_shape_parm, step=2
-    )
-    activation = hp.Choice("activation", ["sigmoid", "relu", "tanh", "selu", "softplus"])
-    optimizer = hp.Choice("optimizer", ["sgd", "adam", "adadelta"])
-    threshold = hp.Float("threshold", min_value=0.1, max_value=0.9, sampling="log")
-
-    model = call_existing_code(
-        units=units,
-        activation=activation,
-        threshold=threshold,
-        optimizer=optimizer,
-        input_shape_parm=input_shape_parm,
-        num_classes=num_classes,
-    )
-    return model
-
-
-def setup_model(
-    data: DataFrame,
-    target: str,
-    epochs: int,
-    train_size: float = 0.7,
-    seed=None,
-    train_mode: bool = True,
-    filepath: str = "./my_dir/best_model",
-    **kwargs,
-) -> AutoClassifier:
-    """Setup model for training and tuning.
-
-    Parameters
-    ----------
-    data : `DataFrame`
-        The dataset to train the model on.
-    target : `str`
-        The name of the target column.
-    epochs : `int`
-        The number of epochs to train the model for.
-    train_size : `float`
-        The proportion of the dataset to use for training.
-    seed : `Any` | `int`
-        The random seed to use for reproducibility.
-    train_mode : `bool`
-        Whether to train the model or not.
-    filepath : `str`
-        The path to save the best model to.
-
-    Keyword Arguments:
-    ----------
-    Additional keyword arguments to pass to the model.
-
-    max_trials : `int`
-        The maximum number of trials to perform.
-    directory : `str`
-        The directory to save the model to.
-    project_name : `str`
-        The name of the project.
-    objective : `str`
-        The objective to optimize.
-    verbose : `bool`
-        Whether to print verbose output.
-
-    Returns
-    -------
-    model : `AutoClassifier`
-        The trained model.
-    """
-    max_trials = kwargs["max_trials"] if "max_trials" in kwargs else 10
-    directory = kwargs["directory"] if "directory" in kwargs else "./my_dir"
-    project_name = kwargs["project_name"] if "project_name" in kwargs else "get_best"
-    objective = kwargs["objective"] if "objective" in kwargs else "val_loss"
-    verbose = kwargs["verbose"] if "verbose" in kwargs else True
-
-    X = data.drop(columns=target)
-    input_sample = X.sample(1)
-    y = data[target]
-    # Verify if there are categorical columns in the dataframe
-    assert (
-        X.select_dtypes(include=["object"]).empty == True
-    ), "Categorical variables within the DataFrame must be encoded, this is done by using the DataFrameEncoder from likelihood."
-    validation_split = 1.0 - train_size
-    # Create my_dir path if it does not exist
-
-    if train_mode:
-        # Create a new directory if it does not exist
-        try:
-            if (not os.path.exists(directory)) and directory != "./":
-                os.makedirs(directory)
-            elif directory != "./":
-                print(f"Directory {directory} already exists, it will be deleted.")
-                rmtree(directory)
-                os.makedirs(directory)
-        except:
-            print("Warning: unable to create directory")
-
-        # Create a Classifier instance
-        y_encoder = OneHotEncoder()
-        y = y_encoder.encode(y.to_list())
-        X = X.to_numpy()
-        input_sample.to_numpy()
-        X = np.asarray(X).astype(np.float32)
-        input_sample = np.asarray(input_sample).astype(np.float32)
-        y = np.asarray(y).astype(np.float32)
-
-        input_shape_parm = X.shape[1]
-        num_classes = y.shape[1]
-        global build_model
-        build_model = partial(
-            build_model, input_shape_parm=input_shape_parm, num_classes=num_classes
-        )
-
-        # Create the AutoKeras model
-        tuner = keras_tuner.RandomSearch(
-            hypermodel=build_model,
-            objective=objective,
-            max_trials=max_trials,
-            directory=directory,
-            project_name=project_name,
-            seed=seed,
-        )
-
-        tuner.search(X, y, epochs=epochs, validation_split=validation_split)
-        models = tuner.get_best_models(num_models=2)
-        best_model = models[0]
-        best_model(input_sample)
-
-        # save model
-        best_model.save(filepath, save_format="tf")
-
-        if verbose:
-            tuner.results_summary()
-    else:
-        # Load the best model from the directory
-        best_model = tf.keras.models.load_model(filepath)
-
-    return best_model
-
-
-########################################################################################