lecrapaud 0.11.2__tar.gz → 0.11.3__tar.gz

{lecrapaud-0.11.2 → lecrapaud-0.11.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: lecrapaud
-Version: 0.11.2
+Version: 0.11.3
 Summary: Framework for machine and deep learning, with regression, classification and time series analysis
 License: Apache License
 Author: Pierre H. Gallet

{lecrapaud-0.11.2 → lecrapaud-0.11.3}/lecrapaud/api.py

@@ -27,11 +27,11 @@ Basic Usage:
 
 import joblib
 import pandas as pd
+import ast
 import logging
 import seaborn as sns
 import numpy as np
 import matplotlib.pyplot as plt
-from lecrapaud.utils import logger
 from lecrapaud.db.session import init_db
 from lecrapaud.feature_selection import FeatureSelectionEngine, PreprocessModel
 from lecrapaud.model_selection import (
@@ -46,6 +46,7 @@ from lecrapaud.feature_engineering import FeatureEngineeringEngine, PreprocessFe
 from lecrapaud.experiment import create_experiment
 from lecrapaud.db import Experiment
 from lecrapaud.search_space import normalize_models_idx
+from lecrapaud.utils import logger
 
 
 class LeCrapaud:
@@ -344,9 +345,13 @@ class ExperimentEngine:
         return pd.read_csv(f"{self.experiment.path}/feature_summary.csv")
 
     def get_threshold(self, target_number: int):
-        return joblib.load(
+        thresholds = joblib.load(
             f"{self.experiment.path}/TARGET_{target_number}/thresholds.pkl"
         )
+        if isinstance(thresholds, str):
+            thresholds = ast.literal_eval(thresholds)
+
+        return thresholds
 
     def load_model(self, target_number: int, model_name: str = None):
 

{lecrapaud-0.11.2 → lecrapaud-0.11.3}/lecrapaud/model_selection.py

@@ -1328,46 +1328,131 @@ def load_model(target_dir: str):
 # plots
 def plot_evaluation_for_classification(prediction: dict):
     """
-    Args
-        prediction (pd.DataFrame): Should be a df with TARGET, PRED, 0, 1 columns for y_true value (TARGET), y_pred (PRED), and probabilities (for classification only : 0 and 1)
+    Plot evaluation metrics for classification tasks (both binary and multiclass).
+
+    Args:
+        prediction (pd.DataFrame): Should be a df with:
+            - TARGET: true labels
+            - PRED: predicted labels
+            - For binary: column '1' or 1 for positive class probabilities
+            - For multiclass: columns 2 onwards for class probabilities
     """
     y_true = prediction["TARGET"]
     y_pred = prediction["PRED"]
-    y_pred_proba = prediction[1] if 1 in prediction.columns else prediction["1"]
 
     # Plot confusion matrix
     plot_confusion_matrix(y_true, y_pred)
 
-    # Compute ROC curve and ROC area
-    fpr, tpr, thresholds = roc_curve(y_true, y_pred_proba)
-    roc_auc = auc(fpr, tpr)
+    # Determine if binary or multiclass
+    unique_labels = np.unique(y_true)
+    unique_labels = np.sort(unique_labels)
+    n_classes = len(unique_labels)
+
+    if n_classes <= 2:
+        # Binary classification
+        y_pred_proba = prediction[1] if 1 in prediction.columns else prediction["1"]
+
+        # Compute and plot ROC curve
+        fpr, tpr, _ = roc_curve(y_true, y_pred_proba)
+        roc_auc = auc(fpr, tpr)
+
+        plt.figure(figsize=(8, 8))
+        plt.plot(
+            fpr,
+            tpr,
+            color="darkorange",
+            lw=2,
+            label=f"ROC curve (area = {roc_auc:0.2f})",
+        )
+        plt.plot([0, 1], [0, 1], color="navy", lw=2, linestyle="--")
+        plt.xlim([0.0, 1.0])
+        plt.ylim([0.0, 1.05])
+        plt.xlabel("False Positive Rate")
+        plt.ylabel("True Positive Rate")
+        plt.title("ROC Curve")
+        plt.legend(loc="lower right")
+        plt.show()
+
+        # Compute and plot precision-recall curve
+        precision, recall, _ = precision_recall_curve(y_true, y_pred_proba)
+        average_precision = average_precision_score(y_true, y_pred_proba)
+
+        plt.figure(figsize=(8, 8))
+        plt.step(recall, precision, color="b", alpha=0.2, where="post")
+        plt.fill_between(recall, precision, step="post", alpha=0.2, color="b")
+        plt.xlabel("Recall")
+        plt.ylabel("Precision")
+        plt.ylim([0.0, 1.05])
+        plt.xlim([0.0, 1.0])
+        plt.title(f"Precision-Recall Curve: AP={average_precision:0.2f}")
+        plt.show()
 
-    plt.figure(figsize=(8, 8))
-    plt.plot(
-        fpr, tpr, color="darkorange", lw=2, label="ROC curve (area = %0.2f)" % roc_auc
-    )
-    plt.plot([0, 1], [0, 1], color="navy", lw=2, linestyle="--")
-    plt.xlim([0.0, 1.0])
-    plt.ylim([0.0, 1.05])
-    plt.xlabel("False Positive Rate")
-    plt.ylabel("True Positive Rate")
-    plt.title("ROC Curve")
-    plt.legend(loc="lower right")
-    plt.show()
+    else:
+        # Multiclass classification
+        # Get class probabilities
+        pred_cols = [
+            col for col in prediction.columns if col not in ["ID", "TARGET", "PRED"]
+        ]
+        y_pred_proba = prediction[pred_cols].values
 
-    # Compute precision-recall curve
-    precision, recall, _ = precision_recall_curve(y_true, y_pred_proba)
-    average_precision = average_precision_score(y_true, y_pred_proba)
-
-    plt.figure(figsize=(8, 8))
-    plt.step(recall, precision, color="b", alpha=0.2, where="post")
-    plt.fill_between(recall, precision, step="post", alpha=0.2, color="b")
-    plt.xlabel("Recall")
-    plt.ylabel("Precision")
-    plt.ylim([0.0, 1.05])
-    plt.xlim([0.0, 1.0])
-    plt.title("Precision-Recall Curve: AP={0:0.2f}".format(average_precision))
-    plt.show()
+        # Compute ROC curve and ROC area for each class
+        fpr = dict()
+        tpr = dict()
+        roc_auc = dict()
+
+        plt.figure(figsize=(10, 8))
+        colors = plt.cm.get_cmap("tab10")(np.linspace(0, 1, n_classes))
+
+        for i, (label, color) in enumerate(zip(unique_labels, colors)):
+            y_true_binary = (y_true == label).astype(int)
+            y_score = y_pred_proba[:, i]
+
+            fpr[i], tpr[i], _ = roc_curve(y_true_binary, y_score)
+            roc_auc[i] = auc(fpr[i], tpr[i])
+
+            plt.plot(
+                fpr[i],
+                tpr[i],
+                color=color,
+                lw=2,
+                label=f"Class {label} (area = {roc_auc[i]:0.2f})",
+            )
+
+        plt.plot([0, 1], [0, 1], "k--", lw=2)
+        plt.xlim([0.0, 1.0])
+        plt.ylim([0.0, 1.05])
+        plt.xlabel("False Positive Rate")
+        plt.ylabel("True Positive Rate")
+        plt.title("Multiclass ROC Curves (One-vs-Rest)")
+        plt.legend(loc="lower right")
+        plt.show()
+
+        # Compute PR curve for each class
+        plt.figure(figsize=(10, 8))
+
+        for i, (label, color) in enumerate(zip(unique_labels, colors)):
+            y_true_binary = (y_true == label).astype(int)
+            y_score = y_pred_proba[:, i]
+
+            precision, recall, _ = precision_recall_curve(y_true_binary, y_score)
+            average_precision = average_precision_score(y_true_binary, y_score)
+
+            plt.step(
+                recall,
+                precision,
+                color=color,
+                alpha=0.8,
+                where="post",
+                label=f"Class {label} (AP = {average_precision:0.2f})",
+            )
+
+        plt.xlabel("Recall")
+        plt.ylabel("Precision")
+        plt.ylim([0.0, 1.05])
+        plt.xlim([0.0, 1.0])
+        plt.title("Multiclass Precision-Recall Curves")
+        plt.legend(loc="lower left")
+        plt.show()
 
 
 def plot_confusion_matrix(y_true, y_pred):

{lecrapaud-0.11.2 → lecrapaud-0.11.3}/lecrapaud/utils.py

@@ -9,7 +9,7 @@ from ftfy import fix_text
 import unicodedata
 import re
 import string
-
+import sys
 from lecrapaud.directories import logger_dir
 from lecrapaud.config import LOGGING_LEVEL, PYTHON_ENV, LECRAPAUD_LOCAL
 
@@ -57,6 +57,11 @@ def setup_logger():
         file_handler.setLevel(log_level)
         logger.addHandler(file_handler)
 
+    stream_handler = logging.StreamHandler(sys.stdout)
+    stream_handler.setFormatter(formatter)
+    stream_handler.setLevel(log_level)
+    logger.addHandler(stream_handler)
+
     _LECRAPAUD_LOGGER_ALREADY_CONFIGURED = True
     return logger
 

{lecrapaud-0.11.2 → lecrapaud-0.11.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "lecrapaud"
-version = "0.11.2"
+version = "0.11.3"
 description = "Framework for machine and deep learning, with regression, classification and time series analysis"
 authors = [
     {name = "Pierre H. Gallet"}