Merge branch 'main' into implement_grid_search

.github/workflows/ci.yml

@@ -47,7 +47,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        os: [ubuntu-latest, macos-latest]
+        os: [ubuntu-latest, macos-latest, windows-latest]
         env: ["py310", "py311", "py312"]
     steps:
       - name: Checkout branch

docs/development.rst

@@ -0,0 +1,38 @@
+Development
+===========
+
+The ``metalearners`` repository can be cloned as follows
+
+.. code-block:: console
+
+  git clone https://github.com/Quantco/metalearners.git
+
+The dependencies are managed via
+`pixi <https://pixi.sh/latest/>`_. Please make sure to install ``pixi`` on
+your system. Once pixi is installed, you can install and run the
+pre-commit hooks as follows:
+
+
+.. code-block:: console
+
+  pixi run pre-commit-install
+  pixi run pre-commit-run
+
+
+You can run the tests as follows:
+
+.. code-block:: console
+
+  pixi run postinstall
+  pixi run pytest tests
+
+You can build the documentation locally by running
+
+.. code-block:: console
+
+  pixi run -e docs postinstall
+  pixi run -e docs docs
+
+You can then inspect the locally built docs by opening ``docs/_build/index.html``.
+
+You can find all pixi tasks in ``pixi.toml``.

docs/examples/example_reuse.ipynb

@@ -253,6 +253,8 @@
     "drlearner = DRLearner(\n",
     "    nuisance_model_factory=LGBMRegressor,\n",
     "    treatment_model_factory=LGBMRegressor,\n",
+    "    nuisance_model_params={\"verbose\": -1},\n",
+    "    treatment_model_params={\"verbose\": -1},\n",
     "    fitted_propensity_model=trained_propensity_model,\n",
     "    is_classification=False,\n",
     "    n_variants=2,\n",

docs/index.rst

@@ -10,6 +10,7 @@ Welcome to metalearners's documentation!
    Glossary <glossary.rst>
    What about parallelism? <parallelism.rst>
    Examples <examples/index.rst>
+   Development <development.rst>
    FAQ <faq.rst>
    API Reference <api/modules>
    Change Log <changelog.rst>

metalearners/_utils.py

@@ -243,9 +243,10 @@ def convert_treatment(treatment: Vector) -> np.ndarray:
         new_treatment = treatment.to_numpy()
     if new_treatment.dtype == bool:
         return new_treatment.astype(int)
-    elif new_treatment.dtype == float and all(x.is_integer() for x in new_treatment):
+    if new_treatment.dtype == float and all(x.is_integer() for x in new_treatment):
         return new_treatment.astype(int)
-    elif new_treatment.dtype != int:
+
+    if not pd.api.types.is_integer_dtype(new_treatment):
         raise TypeError(
             "Treatment must be boolean, integer or float with integer values."
         )