Fixes sklearn pickling issues

nextmv-gurobipy/pyproject.toml

@@ -18,7 +18,7 @@ classifiers = [
 ]
 dependencies = [
     "gurobipy>=12.0.1",
-    "nextmv>=0.25.0"
+    "nextmv>=0.28.1"
 ]
 description = "An SDK for integrating Gurobi with the Nextmv platform"
 dynamic = [

nextmv-scikit-learn/nextmv_sklearn/__about__.py

@@ -1 +1 @@
-__version__ = "v0.3.0"
+__version__ = "v0.3.1.dev1"

nextmv-scikit-learn/nextmv_sklearn/linear_model/options.py

@@ -19,21 +19,25 @@
     nextmv.Option(
         name="fit_intercept",
         option_type=bool,
+        default=True,
         description="Whether to calculate the intercept for this model.",
     ),
     nextmv.Option(
         name="copy_X",
         option_type=bool,
+        default=True,
         description="If True, X will be copied; else, it may be overwritten.",
     ),
     nextmv.Option(
         name="n_jobs",
         option_type=int,
+        default=1,
         description="The number of jobs to use for the computation.",
     ),
     nextmv.Option(
         name="positive",
         option_type=bool,
+        default=False,
         description="When set to True, forces the coefficients to be positive.",
     ),
 ]

nextmv-scikit-learn/pyproject.toml

@@ -18,7 +18,7 @@ classifiers = [
 ]
 dependencies = [
     "scikit-learn>=1.6.1",
-    "nextmv>=0.25.0"
+    "nextmv>=0.28.1"
 ]
 description = "An SDK for integrating scikit-learn with the Nextmv platform"
 dynamic = [

nextmv-scikit-learn/tests/test_pickle.py

@@ -0,0 +1,60 @@
+import os
+import unittest
+
+from nextmv_sklearn.linear_model import LinearRegression, LinearRegressionOptions
+
+import nextmv
+
+
+class MLRegressorModel(nextmv.Model):
+    def solve(self, input: nextmv.Input) -> nextmv.Output:
+        if input.options.mode == "linear":
+            model = LinearRegression(input.options)
+            _ = model
+            return nextmv.Output(solution={}, options=input.options)
+        else:
+            raise ValueError(f"Unsupported mode: {input.options.mode}")
+
+
+class TestPickle(unittest.TestCase):
+    def tearDown(self):
+        """Removes the mlflow elements created during the test."""
+        model_configuration = nextmv.ModelConfiguration(
+            name="reg",
+        )
+        nextmv.model._cleanup_python_model(model_dir="export", model_configuration=model_configuration)
+
+    def test_options(self):
+        model = MLRegressorModel()
+        # Define options (custom and sklearn).
+        sklearn_opts = LinearRegressionOptions().to_nextmv()
+        custom_options = nextmv.Options(
+            nextmv.Option(
+                name="mode",
+                option_type=str,
+                default="linear",
+                description="ML mode (linear or xgboost).",
+                required=False,
+            )
+        )
+        options = custom_options.merge(sklearn_opts)
+        # Create a model configuration so that we can pickle the model.
+        model_configuration = nextmv.ModelConfiguration(
+            name="reg",
+            requirements=[
+                "nextmv",
+                "nextmv-scikit-learn",
+            ],
+            options=options,
+        )
+
+        # Run the model with some input data.
+        input = nextmv.Input(data={}, options=options)
+        output = model.solve(input)
+        self.assertIsInstance(output, nextmv.Output)
+
+        # Save (pickle) the model to a directory.
+        os.makedirs("export", exist_ok=True)
+        model.save("export", model_configuration)
+        # Assert that the "export" directory is not empty
+        self.assertTrue(len(os.listdir("export")) > 0)