[GSK-2625] Fix deprecations in tests fixtures from sklearn 1.4.0 (#1743)

* GSK-2625

* GSK-2625: fix OneHotEncoder

* Regenerating pdm.lock

* Revert " GSK-2625: fix OneHotEncoder"

This reverts commit 0bfa24a.

* Fix deprecated `sparse` argument for sklearn OneHotEncoder

* Regenerating pdm.lock

* Little adjustment of scores due to updated solver in sklearn

* Temporarily disable drug classification fixture tests

Revert this once imbalanced-learn is updated

* Fix test to be selective to specific warning

* Revert "Temporarily disable drug classification fixture tests"

This reverts commit a33a3a8.

* Regenerating pdm.lock

* Regenerating pdm.lock

* Remove sklearn higher bounds

* Regenerating pdm.lock

* Fix mlflow behaviour for windows

* Fix mlflow behaviour for windows

* Fast failing

---------

Co-authored-by: BotLocker <bot.locker@users.noreply.github.com>
Co-authored-by: Matteo Dora <matteo@giskard.ai>
Co-authored-by: Hartorn <bazire@giskard.ai>

Author: 4 people

.github/workflows/build-python.yml

@@ -10,29 +10,29 @@ on:
   workflow_dispatch:
     inputs:
       run-integration-tests:
-        description: 'If integration test should be run'
+        description: "If integration test should be run"
         required: true
         type: boolean
         default: false
       use-cache:
-        description: 'If cache should be used'
+        description: "If cache should be used"
         required: true
         type: boolean
         default: true
       is-dispatch:
-        description: 'Just to identify manual dispatch'
+        description: "Just to identify manual dispatch"
         required: true
         type: boolean
         default: true
   workflow_call:
     inputs:
       run-integration-tests:
-        description: 'If integration test should be run'
+        description: "If integration test should be run"
         required: true
         type: boolean
         default: false
       use-cache:
-        description: 'If cache should be used'
+        description: "If cache should be used"
         required: true
         type: boolean
         default: false
@@ -53,7 +53,7 @@ jobs:
     strategy:
       fail-fast: false # Do not stop when any job fails
       matrix:
-        python-version: [ "3.9", "3.10", "3.11" ]
+        python-version: ["3.9", "3.10", "3.11"]
         os: [ubuntu-latest]
         pydantic_v1: [false]
         pandas_v1: [false]
@@ -116,8 +116,6 @@ jobs:
           pdm install -G :all
           pdm list
 
-      - name: Lint code
-        run: pdm run lint
       - name: Re-install lightgbm from sources for MacOS
         if: ${{ matrix.os == 'macos-latest' }}
         run: |
@@ -157,6 +155,9 @@ jobs:
           pdm run pip freeze | grep '^pandas'
           pdm run pip freeze | grep -q '^pandas==${{ matrix.pandas_v1 && '1' || '2' }}\.'
 
+      - name: Lint code
+        run: pdm run lint
+
       - name: Test code (concurrency)
         run: pdm test-worker
 
@@ -205,7 +206,7 @@ jobs:
       - name: Setup PDM
         uses: pdm-project/setup-pdm@v3
         with:
-          python-version: '3.10'
+          python-version: "3.10"
           cache: false
       - name: Build wheel
         run: pdm build
@@ -233,7 +234,7 @@ jobs:
       - name: Setup PDM
         uses: pdm-project/setup-pdm@v3
         with:
-          python-version: '3.10'
+          python-version: "3.10"
           cache: false
       - name: Build wheel
         run: pdm build
@@ -252,8 +253,8 @@ jobs:
       - name: Setup PDM
         uses: pdm-project/setup-pdm@v3
         with:
-          python-version: '3.10'
-          version: '2.10.4' # Fix to repair the CI, use latest version when fixed on pdm
+          python-version: "3.10"
+          version: "2.10.4" # Fix to repair the CI, use latest version when fixed on pdm
           cache: false
       - name: Build wheel
         run: pdm build
@@ -291,7 +292,7 @@ jobs:
       - name: Set up Pandoc (needed for doc)
         uses: r-lib/actions/setup-pandoc@v2
         with:
-          pandoc-version: '3.1.7' # https://github.com/jgm/pandoc/releases
+          pandoc-version: "3.1.7" # https://github.com/jgm/pandoc/releases
 
       - name: Install dependencies
         run: pdm install -G :all

docs/reference/notebooks/credit_scoring.ipynb

@@ -268,7 +268,7 @@
     "\n",
     "categorical_transformer = Pipeline([\n",
     "    (\"imputer\", SimpleImputer(strategy=\"constant\", fill_value=\"missing\")),\n",
-    "    (\"onehot\", OneHotEncoder(handle_unknown=\"ignore\", sparse=False)),\n",
+    "    (\"onehot\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False)),\n",
     "])\n",
     "\n",
     "preprocessor = ColumnTransformer(transformers=[\n",

docs/reference/notebooks/insurance_prediction_lgbm.ipynb

@@ -346,7 +346,7 @@
    "source": [
     "preprocessor = ColumnTransformer(transformers=[\n",
     "    (\"scaler\", StandardScaler(), NUMERICAL_COLS),\n",
-    "    (\"one_hot_encoder\", OneHotEncoder(handle_unknown=\"ignore\", sparse=False), CATEGORICAL_COLS),\n",
+    "    (\"one_hot_encoder\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False), CATEGORICAL_COLS),\n",
     "])"
    ]
   },

docs/reference/notebooks/newspaper_classification_pytorch.ipynb

@@ -285,7 +285,7 @@
     "class TextClassificationModel(nn.Module):\n",
     "    def __init__(self, vocab_size, embed_dim, num_class):\n",
     "        super(TextClassificationModel, self).__init__()\n",
-    "        self.embedding = nn.EmbeddingBag(vocab_size, embed_dim, sparse=False)\n",
+    "        self.embedding = nn.EmbeddingBag(vocab_size, embed_dim, sparse_output=False)\n",
     "        self.fc = nn.Linear(embed_dim, num_class)\n",
     "        self.init_weights()\n",
     "\n",

docs/reference/notebooks/wage_classification.ipynb

@@ -296,7 +296,7 @@
    "source": [
     "preprocessor = ColumnTransformer(transformers=[\n",
     "    (\"num\", StandardScaler(), NUMERICAL_FEATURES),\n",
-    "    (\"cat\", OneHotEncoder(handle_unknown=\"ignore\", sparse=False), CATEGORICAL_FEATURES),\n",
+    "    (\"cat\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False), CATEGORICAL_FEATURES),\n",
     "])"
    ]
   },

giskard/demo/titanic_classification.py

@@ -60,7 +60,7 @@ def get_model_and_df(model: str = ModelTypes.LOGISTIC_REGRESSION, max_iter: int
     cat_transormer = Pipeline(
         [
             ("imputer", SimpleImputer(strategy="constant", fill_value="missing")),
-            ("onehot", OneHotEncoder(handle_unknown="ignore", sparse=False)),
+            ("onehot", OneHotEncoder(handle_unknown="ignore", sparse_output=False)),
         ]
     )