Merge pull request #140 from DataResponsibly/development

Development

Author: denysgerasymuk799

.github/workflows/build-virny.yml

@@ -48,7 +48,7 @@ jobs:
       - name: Build Virny
         run: |
           source ~/.venv/bin/activate
-          pip install -e ".[dev,docs]"
+          pip install -e ".[test,docs]"
           pip install requests-toolbelt==1.0.0
 
       # We should delete the git project from the build cache to avoid conflicts

.gitignore

@@ -9,6 +9,9 @@ notebooks
 .ipynb_checkpoints
 docs/examples/test.py
 tests/results
+.pt_tmp
+lightning_logs
+saved_models
 
 # Remove big files from GitHub repo
 virny/datasets/2018

README.md

@@ -92,6 +92,7 @@ In contrast to existing fairness software libraries and model card generating fr
 * Metric static and interactive visualizations
 * Data loaders with subsampling for popular fair-ML benchmark datasets
 * User-friendly parameters input via config yaml files 
+* Integration with PyTorch Tabular
 
 Check out [our documentation](https://dataresponsibly.github.io/Virny/) for a comprehensive overview.
 
@@ -118,4 +119,4 @@ If Virny has been useful to you, and you would like to cite it in a scientific p
 
 ## 📝 License
 
-**Virny** is free and open-source software licensed under the [3-clause BSD license](https://github.com/DataResponsibly/Virny/blob/main/LICENSE).
+**Virny** is free and open-source software licensed under the [3-clause BSD license](https://github.com/DataResponsibly/Virny/blob/main/LICENSE).

docs/examples/.pages

@@ -6,3 +6,4 @@ nav:
     - Multiple_Models_Interface_With_Multiple_Test_Sets.md
     - Multiple_Models_Interface_With_Inprocessor.md
     - Multiple_Models_Interface_With_Postprocessor.md
+    - Multiple_Models_Interface_With_PyTorch_Tabular.md

docs/examples/Multiple_Models_Interface_Use_Case.ipynb

@@ -409,7 +409,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_Use_Case.md

@@ -267,7 +267,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```

docs/examples/Multiple_Models_Interface_With_DB_Writer.ipynb

@@ -288,7 +288,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_With_DB_Writer.md

@@ -183,7 +183,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```

docs/examples/Multiple_Models_Interface_With_Error_Analysis.ipynb

@@ -318,7 +318,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_With_Error_Analysis.md

@@ -189,7 +189,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```

docs/examples/Multiple_Models_Interface_With_Inprocessor.ipynb

@@ -291,7 +291,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_With_Inprocessor.md

@@ -176,7 +176,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```

docs/examples/Multiple_Models_Interface_With_Multiple_Test_Sets.ipynb

@@ -266,7 +266,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_With_Multiple_Test_Sets.md

@@ -175,7 +175,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```

docs/examples/Multiple_Models_Interface_With_Postprocessor.ipynb

@@ -334,7 +334,7 @@
    "outputs": [],
    "source": [
     "column_transformer = ColumnTransformer(transformers=[\n",
-    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),\n",
+    "    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),\n",
     "    ('numerical_features', StandardScaler(), data_loader.numerical_columns),\n",
     "])"
    ],

docs/examples/Multiple_Models_Interface_With_Postprocessor.md

@@ -208,7 +208,7 @@ data_loader.X_data[data_loader.X_data.columns[:5]].head()
 
 ```python
 column_transformer = ColumnTransformer(transformers=[
-    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse=False), data_loader.categorical_columns),
+    ('categorical_features', OneHotEncoder(handle_unknown='ignore', sparse_output=False), data_loader.categorical_columns),
     ('numerical_features', StandardScaler(), data_loader.numerical_columns),
 ])
 ```