Merge branch 'main' into marcelo/ctgan

Author: lotif

.github/workflows/code_checks.yml

@@ -32,7 +32,7 @@ jobs:
       - uses: actions/[email protected]
 
       - name: Install uv
-        uses: astral-sh/setup-uv@ed21f2f24f8dd64503750218de024bcf64c7250a
+        uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
           version: "0.5.21"

.github/workflows/docs.yml

@@ -45,7 +45,7 @@ jobs:
         uses: actions/[email protected]
 
       - name: Install uv
-        uses: astral-sh/[email protected].5
+        uses: astral-sh/[email protected].6
         with:
           version: "0.5.21"
           enable-cache: true
@@ -65,7 +65,7 @@ jobs:
         run: touch site/.nojekyll
 
       - name: Upload artifact
-        uses: actions/upload-artifact@v5
+        uses: actions/upload-artifact@v6
         with:
           name: docs-site
           path: site/
@@ -85,7 +85,7 @@ jobs:
           git config user.email 41898282+github-actions[bot]@users.noreply.github.com
 
       - name: Download artifact
-        uses: actions/download-artifact@v6
+        uses: actions/download-artifact@v7
         with:
           name: docs-site
           path: site
@@ -94,7 +94,7 @@ jobs:
         run: touch site/.nojekyll
 
       - name: Deploy to Github pages
-        uses: JamesIves/[email protected].4
+        uses: JamesIves/[email protected].6
         with:
           branch: gh-pages
           folder: site

.github/workflows/integration_tests.yml

@@ -44,7 +44,7 @@ jobs:
       - uses: actions/[email protected]
 
       - name: Install uv
-        uses: astral-sh/setup-uv@ed21f2f24f8dd64503750218de024bcf64c7250a
+        uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
           version: "0.5.21"

.github/workflows/publish.yml

@@ -19,7 +19,7 @@ jobs:
       - uses: actions/[email protected]
 
       - name: Install uv
-        uses: astral-sh/setup-uv@ed21f2f24f8dd64503750218de024bcf64c7250a
+        uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
           version: "0.5.21"

.github/workflows/unit_tests.yml

@@ -44,7 +44,7 @@ jobs:
       - uses: actions/[email protected]
 
       - name: Install uv
-        uses: astral-sh/setup-uv@ed21f2f24f8dd64503750218de024bcf64c7250a
+        uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
           version: "0.5.21"

.gitignore

@@ -44,6 +44,20 @@ outputs/
 # mkdocs site
 site/
 
+# Training examples
+examples/training/single_table/data/**
+examples/training/single_table/results/**
+examples/training/multi_table/data/**
+examples/training/multi_table/results/**
+examples/synthesizing/single_table/data/**
+examples/synthesizing/single_table/results/**
+examples/synthesizing/multi_table/data/**
+examples/synthesizing/multi_table/results/**
+
+# Test artifacts
+tests/integration/attacks/tartan_federer/assets/tabddpm_models/**/challenge_label_predictions.csv
+tests/integration/attacks/tartan_federer/assets/tartan_federer_attack_results
+
 # Training Logs
 *.err
 *.out

examples/ensemble_attack/compute_attack_success.py

@@ -10,8 +10,8 @@
 import pandas as pd
 from omegaconf import DictConfig
 
-from midst_toolkit.attacks.ensemble.metric_utils import get_tpr_at_fpr
 from midst_toolkit.common.logger import log
+from midst_toolkit.evaluation.privacy.mia_scoring import TprAtFpr
 
 
 def load_target_challenge_labels_and_probabilities(
@@ -91,7 +91,7 @@ def compute_attack_success_for_given_targets(
     assert len(predictions) == len(targets), "Number of predictions must match number of targets."
 
     # Compute TPR@FPR for all the target models
-    tpr_at_fpr = get_tpr_at_fpr(targets, predictions, max_fpr=0.1)
+    tpr_at_fpr = TprAtFpr.get_tpr_at_fpr(targets, predictions, fpr_threshold=0.1)
 
     # Save the final attack success rate into a text file.
     metric_save_path = experiment_directory / f"attack_success_for_{metaclassifier_model_name}.txt"

src/midst_toolkit/attacks/ensemble/blending.py

@@ -12,10 +12,10 @@
 from sklearn.linear_model import LogisticRegression
 
 from midst_toolkit.attacks.ensemble.distance_features import calculate_domias_score, calculate_gower_features
-from midst_toolkit.attacks.ensemble.metric_utils import get_tpr_at_fpr
 from midst_toolkit.attacks.ensemble.rmia.rmia_calculation import calculate_rmia_signals
 from midst_toolkit.attacks.ensemble.xgboost_tuner import XgBoostHyperparameterTuner
 from midst_toolkit.common.logger import log
+from midst_toolkit.evaluation.privacy.mia_scoring import TprAtFpr
 
 
 class MetaClassifierType(Enum):
@@ -251,6 +251,6 @@ def predict(
         score = None
 
         if y_test is not None:
-            score = get_tpr_at_fpr(true_membership=y_test, predictions=probabilities, max_fpr=0.1)
+            score = TprAtFpr.get_tpr_at_fpr(true_membership=y_test, predictions=probabilities, max_fpr=0.1)
 
         return probabilities, score

src/midst_toolkit/attacks/ensemble/metric_utils.py

@@ -9,8 +9,8 @@
 from sklearn.pipeline import Pipeline
 from sklearn.preprocessing import StandardScaler
 
-from midst_toolkit.attacks.ensemble.metric_utils import get_tpr_at_fpr
 from midst_toolkit.common.variables import DEVICE
+from midst_toolkit.evaluation.privacy.mia_scoring import TprAtFpr
 
 
 # Setting the logging level to WARNING, suppressing INFO and DEBUG messages from Optuna
@@ -103,7 +103,7 @@ def _evaluate_pipeline_cv(self, trial: Trial, num_kfolds: int) -> float:
             Mean TPR at FPR=0.1 across all folds.
         """
         pipeline = self._create_xgb_pipeline(trial)
-        tpr_scorer = make_scorer(get_tpr_at_fpr)
+        tpr_scorer = make_scorer(TprAtFpr.get_tpr_at_fpr)
 
         cv_scores = cross_val_score(
             pipeline,