likelihood 2.0.0__py3-none-any.whl → 2.0.1__py3-none-any.whl

likelihood/models/deep/rl.py

@@ -27,12 +27,12 @@ class Env:
 
         Parameters
         ----------
-            model : Any
-                Model with `.predict()` method (e.g., Keras model).
-            maxlen : int
-                Maximum length of deque. By default it is set to `100`.
-            name : str
-                The name of the environment. By default it is set to `likenasium`.
+        model : Any
+            Model with `.predict()` method (e.g., Keras model).
+        maxlen : int
+            Maximum length of deque. By default it is set to `100`.
+        name : str
+            The name of the environment. By default it is set to `likenasium`.
         """
         self.model = model
         self.maxlen = maxlen
@@ -49,14 +49,14 @@ class Env:
 
         Parameters
         ----------
-            state : `np.ndarray`
-                Current state to process (input to the model).
-            action : int
-                Expected action to process.
+        state : `np.ndarray`
+            Current state to process (input to the model).
+        action : `int`
+            Expected action to process.
 
         Returns
         -------
-            tuple: (current_state, action_pred, reward, next_action, done)
+            `tuple` : (current_state, action_pred, reward, next_action, done)
         """
         if self.done:
             return None, None, 0, None, True
@@ -120,9 +120,9 @@ class AutoQL:
 
         Parameters
         ----------
-        env : Any
+        env : `Any`
             The environment to interact with
-        model : tf.keras.Model
+        model : `tf.keras.Model`
             The Q-network model
         """
 
@@ -137,16 +137,16 @@ class AutoQL:
 
         Parameters
         ----------
-            state : `np.ndarray`
-                Current state.
-            action : int
-                Expected action to process.
-            epsilon : float
-                Exploration probability. By default it is set to `0`
+        state : `np.ndarray`
+            Current state.
+        action : `int`
+            Expected action to process.
+        epsilon : `float`
+            Exploration probability. By default it is set to `0`
 
         Returns
         -------
-            tuple: (state, action, reward, next_action, done)
+            `tuple` : (state, action, reward, next_action, done)
         """
         current_state, value, reward, next_action, done = self.env.step(state, action)
 
@@ -164,17 +164,17 @@ class AutoQL:
 
         Parameters
         ----------
-            state : `np.ndarray`
-                Current state
-            action : int
-                Expected action to process.
+        state : `np.ndarray`
+            Current state
+        action : `int`
+            Expected action to process.
 
-            epsilon : float
-                Exploration probability.
+        epsilon : `float`
+            Exploration probability.
 
         Returns
         -------
-            tuple: (state, action, reward, next_action, done)
+            `tuple` : (state, action, reward, next_action, done)
         """
         current_state, greedy_action, reward, next_action, done = self.epsilon_greedy_policy(
             state, action, epsilon
@@ -202,7 +202,7 @@ class AutoQL:
 
         Returns
         -------
-            float: Training loss
+            `float` : Training loss
         """
 
         batch_ = random.sample(self.replay_buffer, self.batch_size)
@@ -250,21 +250,21 @@ class AutoQL:
 
         Parameters
         ----------
-        optimizer : str
+        optimizer : `str`
             The optimizer for training (e.g., `sgd`). By default it is set to `adam`.
-        loss_fn : str
+        loss_fn : `str`
             The loss function. By default it is set to `mse`.
-        num_episodes : int
+        num_episodes : `int`
             Total number of episodes to train. By default it is set to `50`.
-        num_steps : int
+        num_steps : `int`
             Steps per episode. By default it is set to `100`. If `num_steps` is less than `self.env.maxlen`, then the second will be chosen.
-        gamma : float
+        gamma : `float`
             Discount factor. By default it is set to `0.7`.
-        batch_size : int
+        batch_size : `int`
             Size of training batches. By default it is set to `32`.
-        patience : int
+        patience : `int`
             How many episodes to wait for improvement.
-        alpha : float
+        alpha : `float`
             Trade-off factor between loss and reward.
         """
         rewards = []

likelihood/tools/models_tools.py

@@ -11,7 +11,7 @@ logging.getLogger("tensorflow").setLevel(logging.ERROR)
 import sys
 import warnings
 from functools import wraps
-from typing import Dict
+from typing import Dict, List, Optional, Tuple, Union
 
 import numpy as np
 import tensorflow as tf
@@ -40,6 +40,189 @@ def suppress_warnings(func):
     return wrapper
 
 
+class TransformRange:
+    """
+    Generates a new DataFrame with ranges represented as strings.
+
+    Transforms numerical columns into categorical range bins with descriptive labels.
+    """
+
+    def __init__(self, df: pd.DataFrame) -> None:
+        """Initializes the class with the original DataFrame.
+
+        Parameters
+        ----------
+        df : `pd.DataFrame`
+            The original DataFrame to transform.
+
+        Raises
+        ------
+        TypeError
+            If df is not a pandas DataFrame.
+        """
+        if not isinstance(df, pd.DataFrame):
+            raise TypeError("df must be a pandas DataFrame")
+        self.df = df.copy()  # Create a copy to avoid modifying the original
+
+    def _create_bins_and_labels(
+        self, min_val: Union[int, float], max_val: Union[int, float], bin_size: int
+    ) -> Tuple[np.ndarray, List[str]]:
+        """
+        Creates the bin edges and their labels.
+
+        Parameters
+        ----------
+        min_val : `int` or `float`
+            The minimum value for the range.
+        max_val : `int` or `float`
+            The maximum value for the range.
+        bin_size : `int`
+            The size of each bin.
+
+        Returns
+        -------
+        bins : `np.ndarray`
+            The bin edges.
+        labels : `list`
+            The labels for the bins.
+
+        Raises
+        ------
+        ValueError
+            If bin_size is not positive or if min_val >= max_val.
+        """
+        if bin_size <= 0:
+            raise ValueError("bin_size must be positive")
+        if min_val >= max_val:
+            raise ValueError("min_val must be less than max_val")
+
+        start = int(min_val)
+        end = int(max_val) + bin_size
+
+        bins = np.arange(start, end + 1, bin_size)
+
+        if bins[-1] <= max_val:
+            bins = np.append(bins, max_val + 1)
+
+        labels = [f"{int(bins[i])}-{int(bins[i+1] - 1)}" for i in range(len(bins) - 1)]
+        return bins, labels
+
+    def _transform_column_to_ranges(self, column: str, bin_size: int) -> pd.Series:
+        """
+        Transforms a column in the DataFrame into range bins.
+
+        Parameters
+        ----------
+        column : `str`
+            The name of the column to transform.
+        bin_size : `int`
+            The size of each bin.
+
+        Returns
+        -------
+        `pd.Series`
+            A Series with the range labels.
+
+        Raises
+        ------
+        KeyError
+            If column is not found in the DataFrame.
+        ValueError
+            If bin_size is not positive or if column contains non-numeric data.
+        """
+        if column not in self.df.columns:
+            raise KeyError(f"Column '{column}' not found in DataFrame")
+
+        if bin_size <= 0:
+            raise ValueError("bin_size must be positive")
+
+        numeric_series = pd.to_numeric(self.df[column], errors="coerce")
+        if numeric_series.isna().all():
+            raise ValueError(f"Column '{column}' contains no valid numeric data")
+
+        min_val = numeric_series.min()
+        max_val = numeric_series.max()
+
+        if min_val == max_val:
+            return pd.Series(
+                [f"{int(min_val)}-{int(max_val)}"] * len(self.df), name=f"{column}_range"
+            )
+
+        bins, labels = self._create_bins_and_labels(min_val, max_val, bin_size)
+
+        return pd.cut(numeric_series, bins=bins, labels=labels, right=False, include_lowest=True)
+
+    def transform_dataframe(
+        self, columns_bin_sizes: Dict[str, int], drop_original: bool = False
+    ) -> pd.DataFrame:
+        """
+        Creates a new DataFrame with range columns.
+
+        Parameters
+        ----------
+        columns_bin_sizes : `dict`
+            A dictionary where the keys are column names and the values are the bin sizes.
+        drop_original : `bool`, optional
+            If True, drops original columns from the result, by default False
+
+        Returns
+        -------
+        `pd.DataFrame`
+            A DataFrame with the transformed range columns.
+
+        Raises
+        ------
+        TypeError
+            If columns_bin_sizes is not a dictionary.
+        """
+        if not isinstance(columns_bin_sizes, dict):
+            raise TypeError("columns_bin_sizes must be a dictionary")
+
+        if not columns_bin_sizes:
+            return pd.DataFrame()
+
+        range_columns = {}
+        for column, bin_size in columns_bin_sizes.items():
+            range_columns[f"{column}_range"] = self._transform_column_to_ranges(column, bin_size)
+
+        result_df = pd.DataFrame(range_columns)
+
+        if not drop_original:
+            original_cols = [col for col in self.df.columns if col not in columns_bin_sizes]
+            if original_cols:
+                result_df = pd.concat([self.df[original_cols], result_df], axis=1)
+
+        return result_df
+
+    def get_range_info(self, column: str) -> Dict[str, Union[int, float, List[str]]]:
+        """
+        Get information about the range transformation for a specific column.
+
+        Parameters
+        ----------
+        column : `str`
+            The name of the column to analyze.
+
+        Returns
+        -------
+        `dict`
+            Dictionary containing min_val, max_val, bin_size, and labels.
+        """
+        if column not in self.df.columns:
+            raise KeyError(f"Column '{column}' not found in DataFrame")
+
+        numeric_series = pd.to_numeric(self.df[column], errors="coerce")
+        min_val = numeric_series.min()
+        max_val = numeric_series.max()
+
+        return {
+            "min_value": min_val,
+            "max_value": max_val,
+            "range": max_val - min_val,
+            "column": column,
+        }
+
+
 def remove_collinearity(df: DataFrame, threshold: float = 0.9):
     """
     Removes highly collinear features from the DataFrame based on a correlation threshold.
@@ -56,8 +239,8 @@ def remove_collinearity(df: DataFrame, threshold: float = 0.9):
         The correlation threshold above which features will be removed. Default is `0.9`.
 
     Returns
-    ----------
-        DataFrame: A DataFrame with highly collinear features removed.
+    -------
+        DataFrame : A DataFrame with highly collinear features removed.
     """
     corr_matrix = df.corr().abs()
     upper_triangle = corr_matrix.where(np.triu(np.ones(corr_matrix.shape), k=1).astype(bool))
@@ -97,11 +280,11 @@ def train_and_insights(
         Fraction of data to use (default is 1.0).
 
     Keyword Arguments:
-    ----------
+    ------------------
     Additional keyword arguments passed to the `model.fit` function, such as validation split and callbacks.
 
     Returns
-    ----------
+    -------
     `tf.keras.Model`
         The trained model after fitting.
     """
@@ -207,7 +390,7 @@ def graph_metrics(adj_matrix: np.ndarray, eigenvector_threshold: float = 1e-6) -
         A threshold for the eigenvector centrality calculation, used to determine the cutoff for small eigenvalues. Default is `1e-6`.
 
     Returns
-    ----------
+    -------
     DataFrame : A DataFrame containing the following graph metrics as columns.
         - `Degree Centrality`: Degree centrality values for each node, indicating the number of direct connections each node has.
         - `Clustering Coefficient`: Clustering coefficient values for each node, representing the degree to which nodes cluster together.
@@ -218,7 +401,7 @@ def graph_metrics(adj_matrix: np.ndarray, eigenvector_threshold: float = 1e-6) -
         - `Assortativity`: The assortativity coefficient of the graph, measuring the tendency of nodes to connect to similar nodes.
 
     Notes
-    ----------
+    -----
     The returned DataFrame will have one row for each node and one column for each of the computed metrics.
     """
     adj_matrix = adj_matrix.astype(int)
@@ -251,3 +434,7 @@ def graph_metrics(adj_matrix: np.ndarray, eigenvector_threshold: float = 1e-6) -
     metrics_df["Assortativity"] = assortativity
 
     return metrics_df
+
+
+if __name__ == "__main__":
+    pass

likelihood/tools/numeric_tools.py

@@ -154,7 +154,7 @@ def xicor(X: np.ndarray, Y: np.ndarray, ties: bool = True, random_seed: int = No
         The first variable to be correlated. Must have at least one dimension.
     Y : `np.ndarray`
         The second variable to be correlated. Must have at least one dimension.
-    ties : bool
+    ties : `bool`
         Whether to handle ties using randomization.
     random_seed : int, optional
         Seed for the random number generator for reproducibility.
@@ -356,9 +356,9 @@ def find_multiples(target: int) -> tuple[int, int] | None:
     Returns
     -------
     tuple[int, int] | None
-        If i and i+1 both divide target, returns (i, i+1).
-        Otherwise, returns (i, target // i).
-        Returns None if no factors are found.
+        If `i` and `i+1` both divide target, returns (i, i+1).
+        Otherwise, returns `(i, target // i)`.
+        Returns `None` if no factors are found.
     """
     for i in range(2, target + 1):
         if target % i == 0:

likelihood/tools/tools.py

@@ -861,7 +861,7 @@ class DataFrameEncoder:
         """Encodes the `object` type columns of the dataframe
 
         Keyword Arguments:
-        ----------
+        ------------------
         - save_mode (`bool`): An optional integer parameter. By default it is set to `True`
         - dictionary_name (`str`): An optional string parameter. By default it is set to `labelencoder_dictionary`
         - norm_method (`str`): An optional string parameter to perform normalization. By default it is set to `None`

{likelihood-2.0.0.dist-info → likelihood-2.0.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: likelihood
-Version: 2.0.0
+Version: 2.0.1
 Summary: A package that performs the maximum likelihood algorithm.
 Home-page: https://github.com/jzsmoreno/likelihood/
 Author: J. A. Moreno-Guerra

{likelihood-2.0.0.dist-info → likelihood-2.0.1.dist-info}/RECORD

@@ -15,16 +15,16 @@ likelihood/models/deep/_predictor.py,sha256=XI4QfVM7PS_60zYtmi-V8UzNDrASFiDMVPmV
 likelihood/models/deep/autoencoders.py,sha256=muUBH9BclOK8ViI7PijyMOBBLVox6uwuIabyJvpU5qw,30729
 likelihood/models/deep/gan.py,sha256=rTnaLmIPjsKg6_0B8JZOVwPxdx59rHmqvzDitdJMCQ4,10924
 likelihood/models/deep/predictor.py,sha256=q5tPaAbF7s5XIcxVr6fyHTQdZa9tlixO9vb9a9Cw0wM,27831
-likelihood/models/deep/rl.py,sha256=9dhhnVTIETi9zvVeyOXYo1hl-LQJezmv0rgsUq11Qwc,11611
+likelihood/models/deep/rl.py,sha256=VVuwHwK24d2fe3uNHliE1QJsKGZAPhx_pdgj3jqN5rQ,11565
 likelihood/tools/__init__.py,sha256=N1IhMDzacsGQT2MIYBMBC0zTxes78vC_0gGrwkuPgmg,78
 likelihood/tools/cat_embed.py,sha256=SJ7o1vbrNYp21fLLcjRnWpUDcz1nVSe8TmMvsLIz5CI,7346
 likelihood/tools/figures.py,sha256=waF0NHIMrctCmaLhcuz5DMcXyRKynmn6aG0XITYCTLc,10940
 likelihood/tools/impute.py,sha256=n87Tv-xLUAdPl7BQLFcLWSsXBZbXksahyCayJWMydXc,9485
-likelihood/tools/models_tools.py,sha256=c3-vac-1MYSarYDtfR6XfVC7X_WY9auS7y2_3Z973IQ,8875
-likelihood/tools/numeric_tools.py,sha256=Hwf-lbqROqPPZ9N7eVzKIDyZxFGQdP53isWxPqpG0eo,12254
-likelihood/tools/tools.py,sha256=GKZsqjyO5tGXWGSfn3jlQBTjRlmBv2byfvpu-QclUx0,42188
-likelihood-2.0.0.dist-info/licenses/LICENSE,sha256=XWHWt9egYEUHGPTnlcZfJKLPmysacOwdiLj_-J7Z9ew,1066
-likelihood-2.0.0.dist-info/METADATA,sha256=Ziysy1MQuW77OHHd1UzMtlfeUT9wsdgCl6rxW3uLBEE,2917
-likelihood-2.0.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-likelihood-2.0.0.dist-info/top_level.txt,sha256=KDiBLr870YTxqLFqObTOSrTK10uw8dFsITSNLlte3PA,11
-likelihood-2.0.0.dist-info/RECORD,,
+likelihood/tools/models_tools.py,sha256=-QAfvCy9mw-ZyeJHzJJ7O6eDfUXghtA7KfFtTc-Tp0A,14607
+likelihood/tools/numeric_tools.py,sha256=JeLECoVS3ayFH53kUYkAMs0fzALZV1M22-tBLM-Q34g,12264
+likelihood/tools/tools.py,sha256=5vPUHrm8D4ODsg-MP4uZ3NgXV9fNbs0Olx7RWtUdVDU,42196
+likelihood-2.0.1.dist-info/licenses/LICENSE,sha256=XWHWt9egYEUHGPTnlcZfJKLPmysacOwdiLj_-J7Z9ew,1066
+likelihood-2.0.1.dist-info/METADATA,sha256=3mLJAcVO4jzu4IoCVVaSBPMxBWV-xnHs_f_DvvN9G0c,2917
+likelihood-2.0.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+likelihood-2.0.1.dist-info/top_level.txt,sha256=KDiBLr870YTxqLFqObTOSrTK10uw8dFsITSNLlte3PA,11
+likelihood-2.0.1.dist-info/RECORD,,