CounterfactUS 0.1.0__tar.gz

counterfactus-0.1.0/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Antonio Navas Orozco
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

counterfactus-0.1.0/PKG-INFO

@@ -0,0 +1,142 @@
+Metadata-Version: 2.4
+Name: CounterfactUS
+Version: 0.1.0
+Summary: Optimal and robust counterfactual explanations
+Author-email: Antonio Navas Orozco <anavas1@us.es>
+License: MIT License
+        
+        Copyright (c) 2026 Antonio Navas Orozco
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: Homepage, https://github.com/antnavoro/CounterfactUS
+Project-URL: Repository, https://github.com/antnavoro/CounterfactUS
+Project-URL: Bug Tracker, https://github.com/antnavoro/CounterfactUS/issues
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Intended Audience :: Science/Research
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: matplotlib
+Requires-Dist: seaborn
+Requires-Dist: scipy
+Requires-Dist: rpy2
+Requires-Dist: gurobipy
+Dynamic: license-file
+
+# CounterfactUS: Optimal & Robust Counterfactual Explanations
+
+**CounterfactUS** is a Python library designed to generate optimal and robust counterfactual explanations for predictive models. It formulates the counterfactual generation as a Mixed-Integer Programming (MIP) problem, allowing for exact counterfactuals.
+
+## 🚀 Key Features
+
+*   **Model Support:** Designed for Generalized Partially Linear Single-Index Models (GPLSIM), including special subcases like Single-Index Models (SIM) and Generalized Linear Models (GLM). (Uses R's `mgcv` and `gplsim` internally).
+*   **Optimal Explanations:** Uses Gurobi as an optimization backend to find counterfactuals that minimize distance exactly, rather than relying on heuristics.
+*   **Statistical Robustness:** Provides ground-truth model guarantees by enforcing prediction thresholds within a user-defined *confidence level*.
+*   **Highly Customizable Constraints:**
+    *   **Distance Metrics:** Support for L1, L2, Gower, and combined L0+L1 distances.
+    *   **Sparsity (L0):** Limit the maximum number of features that can be modified.
+    *   **Immutability & Bounds:** Prevent specific features (e.g., *Age*, *Race*) from changing or restrict their allowable ranges.
+    *   **Directional Weights:** Assign different weights to increasing or decreasing specific features.
+    *   **Linear Constraints:** Define linear relationships (e.g., $X_1 + X_2 \le b$).
+*   **Automated Preprocessing:** Automatically infers feature types (binary, categorical, continuous), normalizes data, and handles one-hot encoding internally.
+
+## 📦 Installation
+
+This package requires Python 3.9+ and an active Gurobi license (or academic license) to solve the MIP problems. Additionally, it requires an R installation with the `mgcv` and `gplsim` packages available.
+
+You can install the package directly via pip:
+
+```bash
+pip install CounterfactUS
+```
+
+*Note: Ensure that your system's R environment is accessible to Python via `rpy2`.*
+
+## 💡 Quick Start
+
+Generating counterfactuals is as simple as calling the `compute_counterfactuals` function. 
+
+```python
+import pandas as pd
+from counterfactus import compute_counterfactuals
+
+# 1. Load your dataset
+X_train = pd.DataFrame(...) # Non-parametric features (e.g., continuous variables for GPLSIM)
+Z_train = pd.DataFrame(...) # Linear features (e.g., categorical variables or strict linear effects)
+Y_train = pd.Series(...)    # Target variable
+
+# 2. Define the query instances you want to explain
+x_queries = pd.DataFrame(...)
+z_queries = pd.DataFrame(...)
+
+# 3. Generate counterfactuals
+results_df = compute_counterfactuals(
+    X=X_train, 
+    Z=Z_train, 
+    Y=Y_train,
+    family_name="binomial", 
+    link_name="logit",
+    prediction_threshold=0.5,       # Target prediction threshold
+    x0s=x_queries,
+    z0s=z_queries,
+    distance="l1",                  # Distance metric to minimize
+    max_l0=2,                       # Allow a maximum of 2 feature changes
+    confidence_level=0.95,          # Ensure 95% statistical confidence in the recourse
+    inmutable_z=["Race", "Sex"],    # Prevent demographic features from changing
+    gurobi_verbose=False
+)
+
+# 4. View results
+print(results_df)
+```
+
+## ⚙️ Advanced Usage
+
+`compute_counterfactuals` handles complex parameter grids natively. You can pass lists to parameters like `prediction_threshold`, `max_l0`, or `confidence_level` to explore the Pareto frontier of counterfactuals. 
+
+By default, the library computes the Cartesian product of all list parameters. You can change this behavior to iterate through the lists concurrently by setting `use_zip_iteration=True`.
+
+```python
+# Explore trade-offs between sparsity and target threshold
+results = compute_counterfactuals(
+    # ... data ...
+    prediction_threshold=[0.5, 0.6, 0.7, 0.8],
+    max_l0=[1, 2, 3], 
+    confidence_level=[0.5, 0.90, 0.95],
+)
+# or
+# Automatically generate trade-off plots
+results = compute_counterfactuals(
+    # ... data ...
+    plot_figures=True
+)
+```
+
+## 📄 License
+
+MIT License

counterfactus-0.1.0/README.md

@@ -0,0 +1,92 @@
+# CounterfactUS: Optimal & Robust Counterfactual Explanations
+
+**CounterfactUS** is a Python library designed to generate optimal and robust counterfactual explanations for predictive models. It formulates the counterfactual generation as a Mixed-Integer Programming (MIP) problem, allowing for exact counterfactuals.
+
+## 🚀 Key Features
+
+*   **Model Support:** Designed for Generalized Partially Linear Single-Index Models (GPLSIM), including special subcases like Single-Index Models (SIM) and Generalized Linear Models (GLM). (Uses R's `mgcv` and `gplsim` internally).
+*   **Optimal Explanations:** Uses Gurobi as an optimization backend to find counterfactuals that minimize distance exactly, rather than relying on heuristics.
+*   **Statistical Robustness:** Provides ground-truth model guarantees by enforcing prediction thresholds within a user-defined *confidence level*.
+*   **Highly Customizable Constraints:**
+    *   **Distance Metrics:** Support for L1, L2, Gower, and combined L0+L1 distances.
+    *   **Sparsity (L0):** Limit the maximum number of features that can be modified.
+    *   **Immutability & Bounds:** Prevent specific features (e.g., *Age*, *Race*) from changing or restrict their allowable ranges.
+    *   **Directional Weights:** Assign different weights to increasing or decreasing specific features.
+    *   **Linear Constraints:** Define linear relationships (e.g., $X_1 + X_2 \le b$).
+*   **Automated Preprocessing:** Automatically infers feature types (binary, categorical, continuous), normalizes data, and handles one-hot encoding internally.
+
+## 📦 Installation
+
+This package requires Python 3.9+ and an active Gurobi license (or academic license) to solve the MIP problems. Additionally, it requires an R installation with the `mgcv` and `gplsim` packages available.
+
+You can install the package directly via pip:
+
+```bash
+pip install CounterfactUS
+```
+
+*Note: Ensure that your system's R environment is accessible to Python via `rpy2`.*
+
+## 💡 Quick Start
+
+Generating counterfactuals is as simple as calling the `compute_counterfactuals` function. 
+
+```python
+import pandas as pd
+from counterfactus import compute_counterfactuals
+
+# 1. Load your dataset
+X_train = pd.DataFrame(...) # Non-parametric features (e.g., continuous variables for GPLSIM)
+Z_train = pd.DataFrame(...) # Linear features (e.g., categorical variables or strict linear effects)
+Y_train = pd.Series(...)    # Target variable
+
+# 2. Define the query instances you want to explain
+x_queries = pd.DataFrame(...)
+z_queries = pd.DataFrame(...)
+
+# 3. Generate counterfactuals
+results_df = compute_counterfactuals(
+    X=X_train, 
+    Z=Z_train, 
+    Y=Y_train,
+    family_name="binomial", 
+    link_name="logit",
+    prediction_threshold=0.5,       # Target prediction threshold
+    x0s=x_queries,
+    z0s=z_queries,
+    distance="l1",                  # Distance metric to minimize
+    max_l0=2,                       # Allow a maximum of 2 feature changes
+    confidence_level=0.95,          # Ensure 95% statistical confidence in the recourse
+    inmutable_z=["Race", "Sex"],    # Prevent demographic features from changing
+    gurobi_verbose=False
+)
+
+# 4. View results
+print(results_df)
+```
+
+## ⚙️ Advanced Usage
+
+`compute_counterfactuals` handles complex parameter grids natively. You can pass lists to parameters like `prediction_threshold`, `max_l0`, or `confidence_level` to explore the Pareto frontier of counterfactuals. 
+
+By default, the library computes the Cartesian product of all list parameters. You can change this behavior to iterate through the lists concurrently by setting `use_zip_iteration=True`.
+
+```python
+# Explore trade-offs between sparsity and target threshold
+results = compute_counterfactuals(
+    # ... data ...
+    prediction_threshold=[0.5, 0.6, 0.7, 0.8],
+    max_l0=[1, 2, 3], 
+    confidence_level=[0.5, 0.90, 0.95],
+)
+# or
+# Automatically generate trade-off plots
+results = compute_counterfactuals(
+    # ... data ...
+    plot_figures=True
+)
+```
+
+## 📄 License
+
+MIT License

counterfactus-0.1.0/pyproject.toml

@@ -0,0 +1,42 @@
+[build-system]
+requires = ["setuptools>=68", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name            = "CounterfactUS"
+version         = "0.1.0"
+description     = "Optimal and robust counterfactual explanations"
+readme          = "README.md"
+requires-python = ">=3.9"
+authors = [
+  { name = "Antonio Navas Orozco", email = "anavas1@us.es" }
+]
+license = { file = "LICENSE" }
+classifiers = [
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "License :: OSI Approved :: MIT License",
+    "Operating System :: OS Independent",
+    "Intended Audience :: Science/Research",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+]
+
+dependencies = [
+    "numpy",
+    "pandas",
+    "matplotlib",
+    "seaborn",
+    "scipy",
+    "rpy2",
+    "gurobipy",
+]
+
+[project.urls]
+Homepage = "https://github.com/antnavoro/CounterfactUS"
+Repository = "https://github.com/antnavoro/CounterfactUS"
+"Bug Tracker" = "https://github.com/antnavoro/CounterfactUS/issues"
+
+[tool.setuptools.packages.find]
+where = ["src"]

counterfactus-0.1.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

counterfactus-0.1.0/src/CounterfactUS.egg-info/PKG-INFO

@@ -0,0 +1,142 @@
+Metadata-Version: 2.4
+Name: CounterfactUS
+Version: 0.1.0
+Summary: Optimal and robust counterfactual explanations
+Author-email: Antonio Navas Orozco <anavas1@us.es>
+License: MIT License
+        
+        Copyright (c) 2026 Antonio Navas Orozco
+        
+        Permission is hereby granted, free of charge, to any person obtaining a copy
+        of this software and associated documentation files (the "Software"), to deal
+        in the Software without restriction, including without limitation the rights
+        to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+        copies of the Software, and to permit persons to whom the Software is
+        furnished to do so, subject to the following conditions:
+        
+        The above copyright notice and this permission notice shall be included in all
+        copies or substantial portions of the Software.
+        
+        THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+        IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+        FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+        AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+        LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+        OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+        SOFTWARE.
+        
+Project-URL: Homepage, https://github.com/antnavoro/CounterfactUS
+Project-URL: Repository, https://github.com/antnavoro/CounterfactUS
+Project-URL: Bug Tracker, https://github.com/antnavoro/CounterfactUS/issues
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Intended Audience :: Science/Research
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: matplotlib
+Requires-Dist: seaborn
+Requires-Dist: scipy
+Requires-Dist: rpy2
+Requires-Dist: gurobipy
+Dynamic: license-file
+
+# CounterfactUS: Optimal & Robust Counterfactual Explanations
+
+**CounterfactUS** is a Python library designed to generate optimal and robust counterfactual explanations for predictive models. It formulates the counterfactual generation as a Mixed-Integer Programming (MIP) problem, allowing for exact counterfactuals.
+
+## 🚀 Key Features
+
+*   **Model Support:** Designed for Generalized Partially Linear Single-Index Models (GPLSIM), including special subcases like Single-Index Models (SIM) and Generalized Linear Models (GLM). (Uses R's `mgcv` and `gplsim` internally).
+*   **Optimal Explanations:** Uses Gurobi as an optimization backend to find counterfactuals that minimize distance exactly, rather than relying on heuristics.
+*   **Statistical Robustness:** Provides ground-truth model guarantees by enforcing prediction thresholds within a user-defined *confidence level*.
+*   **Highly Customizable Constraints:**
+    *   **Distance Metrics:** Support for L1, L2, Gower, and combined L0+L1 distances.
+    *   **Sparsity (L0):** Limit the maximum number of features that can be modified.
+    *   **Immutability & Bounds:** Prevent specific features (e.g., *Age*, *Race*) from changing or restrict their allowable ranges.
+    *   **Directional Weights:** Assign different weights to increasing or decreasing specific features.
+    *   **Linear Constraints:** Define linear relationships (e.g., $X_1 + X_2 \le b$).
+*   **Automated Preprocessing:** Automatically infers feature types (binary, categorical, continuous), normalizes data, and handles one-hot encoding internally.
+
+## 📦 Installation
+
+This package requires Python 3.9+ and an active Gurobi license (or academic license) to solve the MIP problems. Additionally, it requires an R installation with the `mgcv` and `gplsim` packages available.
+
+You can install the package directly via pip:
+
+```bash
+pip install CounterfactUS
+```
+
+*Note: Ensure that your system's R environment is accessible to Python via `rpy2`.*
+
+## 💡 Quick Start
+
+Generating counterfactuals is as simple as calling the `compute_counterfactuals` function. 
+
+```python
+import pandas as pd
+from counterfactus import compute_counterfactuals
+
+# 1. Load your dataset
+X_train = pd.DataFrame(...) # Non-parametric features (e.g., continuous variables for GPLSIM)
+Z_train = pd.DataFrame(...) # Linear features (e.g., categorical variables or strict linear effects)
+Y_train = pd.Series(...)    # Target variable
+
+# 2. Define the query instances you want to explain
+x_queries = pd.DataFrame(...)
+z_queries = pd.DataFrame(...)
+
+# 3. Generate counterfactuals
+results_df = compute_counterfactuals(
+    X=X_train, 
+    Z=Z_train, 
+    Y=Y_train,
+    family_name="binomial", 
+    link_name="logit",
+    prediction_threshold=0.5,       # Target prediction threshold
+    x0s=x_queries,
+    z0s=z_queries,
+    distance="l1",                  # Distance metric to minimize
+    max_l0=2,                       # Allow a maximum of 2 feature changes
+    confidence_level=0.95,          # Ensure 95% statistical confidence in the recourse
+    inmutable_z=["Race", "Sex"],    # Prevent demographic features from changing
+    gurobi_verbose=False
+)
+
+# 4. View results
+print(results_df)
+```
+
+## ⚙️ Advanced Usage
+
+`compute_counterfactuals` handles complex parameter grids natively. You can pass lists to parameters like `prediction_threshold`, `max_l0`, or `confidence_level` to explore the Pareto frontier of counterfactuals. 
+
+By default, the library computes the Cartesian product of all list parameters. You can change this behavior to iterate through the lists concurrently by setting `use_zip_iteration=True`.
+
+```python
+# Explore trade-offs between sparsity and target threshold
+results = compute_counterfactuals(
+    # ... data ...
+    prediction_threshold=[0.5, 0.6, 0.7, 0.8],
+    max_l0=[1, 2, 3], 
+    confidence_level=[0.5, 0.90, 0.95],
+)
+# or
+# Automatically generate trade-off plots
+results = compute_counterfactuals(
+    # ... data ...
+    plot_figures=True
+)
+```
+
+## 📄 License
+
+MIT License

counterfactus-0.1.0/src/CounterfactUS.egg-info/SOURCES.txt

@@ -0,0 +1,14 @@
+LICENSE
+README.md
+pyproject.toml
+src/CounterfactUS.egg-info/PKG-INFO
+src/CounterfactUS.egg-info/SOURCES.txt
+src/CounterfactUS.egg-info/dependency_links.txt
+src/CounterfactUS.egg-info/requires.txt
+src/CounterfactUS.egg-info/top_level.txt
+src/counterfactus/__init__.py
+src/counterfactus/covariance_GPLSIM.py
+src/counterfactus/fit_GPLSIM.py
+src/counterfactus/main.py
+src/counterfactus/solver.py
+tests/test_main.py

counterfactus-0.1.0/src/CounterfactUS.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

counterfactus-0.1.0/src/CounterfactUS.egg-info/requires.txt

@@ -0,0 +1,7 @@
+numpy
+pandas
+matplotlib
+seaborn
+scipy
+rpy2
+gurobipy

counterfactus-0.1.0/src/CounterfactUS.egg-info/top_level.txt

@@ -0,0 +1 @@
+counterfactus

counterfactus-0.1.0/src/counterfactus/__init__.py

@@ -0,0 +1,5 @@
+from .main import compute_counterfactuals
+
+__all__ = [
+    "compute_counterfactuals",
+]