Raise when trying to sample a Multinomial variable

pymc/sampling/mcmc.py

@@ -51,6 +51,7 @@
 from pymc.backends.base import IBaseTrace, MultiTrace, _choose_chains
 from pymc.backends.zarr import ZarrChain, ZarrTrace
 from pymc.blocking import DictToArrayBijection
+from pymc.distributions.multivariate import Multinomial
 from pymc.exceptions import SamplingError
 from pymc.initial_point import PointType, StartDict, make_initial_point_fns_per_chain
 from pymc.model import Model, modelcontext
@@ -63,6 +64,7 @@
 )
 from pymc.step_methods import NUTS, CompoundStep
 from pymc.step_methods.arraystep import BlockedStep, PopulationArrayStepShared
+from pymc.step_methods.cannot_sample import CannotSampleRV
 from pymc.step_methods.hmc import quadpotential
 from pymc.util import (
     ProgressBarManager,
@@ -144,6 +146,13 @@
         if initial_point is None:
             initial_point = model.initial_point()
 
+        for rv in model.free_RVs:
+            if isinstance(rv.owner.op, Multinomial) and getattr(rv.tag, "observed", None) is None:
+                for step_class in list(selected_steps.keys()):
+                    if rv in selected_steps[step_class]:
+                        selected_steps[step_class].remove(rv)
+                selected_steps.setdefault(CannotSampleRV, []).append(rv)
+
         for step_class, vars in selected_steps.items():
             if vars:
                 name = getattr(step_class, "name")

pymc/step_methods/cannot_sample.py

@@ -0,0 +1,32 @@
+#   Copyright 2025 - present The PyMC Developers
+#
+#   Licensed under the Apache License, Version 2.0 (the "License");
+#   you may not use this file except in compliance with the License.
+#   You may obtain a copy of the License at
+#
+#       http://www.apache.org/licenses/LICENSE-2.0
+#
+#   Unless required by applicable law or agreed to in writing, software
+#   distributed under the License is distributed on an "AS IS" BASIS,
+#   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#   See the License for the specific language governing permissions and
+#   limitations under the License.
+from pymc.step_methods.arraystep import ArrayStep
+
+
+class CannotSampleRV(ArrayStep):
+    """A step method that raises an error when sampling a latent Multinomial variable."""
+
+    name = "cannot_sample_rv"
+
+    def __init__(self, vars, **kwargs):
+        # Remove keys that ArrayStep.__init__ does not accept.
+        kwargs.pop("model", None)
+        kwargs.pop("initial_point", None)
+        kwargs.pop("compile_kwargs", None)
+        self.vars = vars
+        super().__init__(vars=vars, fs=[], **kwargs)
+
+    def astep(self, q0):
+        # This method is required by the abstract base class.
+        raise ValueError("Latent Multinomial variables are not supported")

tests/distributions/test_distribution.py

@@ -83,6 +83,15 @@ def test_issue_4499(self):
             x = pm.DiracDelta("x", 1, size=10)
         npt.assert_almost_equal(m.compile_logp()({"x": np.ones(10)}), 0 * 10)
 
+    def test_issue_7548(self):
+        # Test for bug in Multinomial, it should raise when trying to sample a Multinomial variable
+        with pm.Model() as model:
+            p = [0.3, 0.4, 0.3]
+            n = 10
+            x = pm.Multinomial("x", n=n, p=p)
+            with pytest.raises(ValueError, match="Latent Multinomial variables are not supported"):
+                pm.sample(draws=100, chains=1)
+
 
 def test_all_distributions_have_support_points():
     import pymc.distributions as dist_module