Add x0 as alias for params.

src/optimagic/exceptions.py

@@ -27,6 +27,10 @@ class MissingInputError(OptimagicError):
     """Exception for missing user provided input."""
 
 
+class AliasError(OptimagicError):
+    """Exception for aliasing errors."""
+
+
 class InvalidKwargsError(OptimagicError):
     """Exception for invalid user provided keyword arguments."""
 

src/optimagic/optimization/optimize.py

@@ -7,6 +7,7 @@
     InvalidFunctionError,
     InvalidKwargsError,
     MissingInputError,
+    AliasError,
 )
 from optimagic.logging.create_tables import (
     make_optimization_iteration_table,
@@ -63,6 +64,8 @@ def maximize(
     multistart_options=None,
     collect_history=True,
     skip_checks=False,
+    # scipy aliases
+    x0=None,
     # deprecated arguments
     criterion=None,
     criterion_kwargs=None,
@@ -99,6 +102,8 @@ def maximize(
         multistart_options=multistart_options,
         collect_history=collect_history,
         skip_checks=skip_checks,
+        # scipy aliases
+        x0=x0,
         # deprecated arguments
         criterion=criterion,
         criterion_kwargs=criterion_kwargs,
@@ -136,6 +141,8 @@ def minimize(
     multistart_options=None,
     collect_history=True,
     skip_checks=False,
+    # scipy aliases
+    x0=None,
     # deprecated arguments
     criterion=None,
     criterion_kwargs=None,
@@ -173,6 +180,8 @@ def minimize(
         multistart_options=multistart_options,
         collect_history=collect_history,
         skip_checks=skip_checks,
+        # scipy aliases
+        x0=x0,
         # deprecated arguments
         criterion=criterion,
         criterion_kwargs=criterion_kwargs,
@@ -211,6 +220,8 @@ def _optimize(
     multistart_options,
     collect_history,
     skip_checks,
+    # scipy aliases
+    x0,
     # deprecated arguments
     criterion,
     criterion_kwargs,
@@ -239,7 +250,7 @@ def _optimize(
         )
         raise MissingInputError(msg)
 
-    if params is None:
+    if params is None and x0 is None:
         msg = (
             "Missing start parameters. Please provide start parameters as the second "
             "positional argument or as the keyword argument `params`."
@@ -328,6 +339,20 @@ def _optimize(
         if fun_and_jac_kwargs is None:
             fun_and_jac_kwargs = criterion_and_derivative_kwargs
 
+    # ==================================================================================
+    # handle scipy aliases
+    # ==================================================================================
+
+    if x0 is not None:
+        if params is not None:
+            msg = (
+                "x0 is an alias for params (for better compatibility with scipy). "
+                "Do not use both x0 and params."
+            )
+            raise AliasError(msg)
+        else:
+            params = x0
+
     # ==================================================================================
     # Set default values and check options
     # ==================================================================================

tests/optimagic/optimization/test_scipy_aliases.py

@@ -0,0 +1,43 @@
+import optimagic as om
+import numpy as np
+from numpy.testing import assert_array_almost_equal as aaae
+from optimagic.exceptions import AliasError
+import pytest
+
+
+def test_x0_works_in_minimize():
+    res = om.minimize(
+        fun=lambda x: x @ x,
+        x0=np.arange(3),
+        algorithm="scipy_lbfgsb",
+    )
+    aaae(res.params, np.zeros(3))
+
+
+def test_x0_works_in_maximize():
+    res = om.maximize(
+        fun=lambda x: -x @ x,
+        x0=np.arange(3),
+        algorithm="scipy_lbfgsb",
+    )
+    aaae(res.params, np.zeros(3))
+
+
+def test_x0_and_params_do_not_work_together_in_minimize():
+    with pytest.raises(AliasError, match="x0 is an alias"):
+        om.minimize(
+            fun=lambda x: x @ x,
+            x0=np.arange(3),
+            params=np.arange(3),
+            algorithm="scipy_lbfgsb",
+        )
+
+
+def test_x0_and_params_do_not_work_together_in_maximize():
+    with pytest.raises(AliasError, match="x0 is an alias"):
+        om.maximize(
+            fun=lambda x: -x @ x,
+            x0=np.arange(3),
+            params=np.arange(3),
+            algorithm="scipy_lbfgsb",
+        )