diff --git a/miplearn/problems/tsp.py b/miplearn/problems/tsp.py
index 209d24c..eb494e1 100644
--- a/miplearn/problems/tsp.py
+++ b/miplearn/problems/tsp.py
@@ -149,21 +149,8 @@ def build_tsp_model_gurobipy(
     )
 
     def lazy_separate(model: GurobiModel) -> List[Any]:
-        violations = []
-        x = model.inner.cbGetSolution(model.inner._x)
-        selected_edges = [e for e in model.inner._edges if x[e] > 0.5]
-        graph = nx.Graph()
-        graph.add_edges_from(selected_edges)
-        for component in list(nx.connected_components(graph)):
-            if len(component) < model.inner._n_cities:
-                cut_edges = tuple(
-                    (e[0], e[1])
-                    for e in model.inner._edges
-                    if (e[0] in component and e[1] not in component)
-                    or (e[0] not in component and e[1] in component)
-                )
-                violations.append(cut_edges)
-        return violations
+        x_val = model.inner.cbGetSolution(model.inner._x)
+        return _tsp_separate(x_val, edges, data.n_cities)
 
     def lazy_enforce(model: GurobiModel, violations: List[Any]) -> None:
         for violation in violations:
@@ -212,22 +199,9 @@ def build_tsp_model_pyomo(
     model.subtour_eqs = pe.ConstraintList()
 
     def lazy_separate(m: PyomoModel) -> List[Any]:
-        violations = []
         m.solver.cbGetSolution([model.x[e] for e in edges])
         x_val = {e: model.x[e].value for e in edges}
-        selected_edges = [e for e in edges if x_val[e] > 0.5]
-        graph = nx.Graph()
-        graph.add_edges_from(selected_edges)
-        for component in list(nx.connected_components(graph)):
-            if len(component) < data.n_cities:
-                cut_edges = tuple(
-                    (e[0], e[1])
-                    for e in edges
-                    if (e[0] in component and e[1] not in component)
-                    or (e[0] not in component and e[1] in component)
-                )
-                violations.append(cut_edges)
-        return violations
+        return _tsp_separate(x_val, edges, data.n_cities)
 
     def lazy_enforce(m: PyomoModel, violations: List[Any]) -> None:
         logger.warning(f"Adding {len(violations)} subtour elimination constraints...")
@@ -251,3 +225,24 @@ def _tsp_read(data: Union[str, TravelingSalesmanData]) -> TravelingSalesmanData:
         data = read_pkl_gz(data)
     assert isinstance(data, TravelingSalesmanData)
     return data
+
+
+def _tsp_separate(
+    x_val: dict[Tuple[int, int], float],
+    edges: List[Tuple[int, int]],
+    n_cities: int,
+) -> List[Tuple[Tuple[int, int], ...]]:
+    violations = []
+    selected_edges = [e for e in edges if x_val[e] > 0.5]
+    graph = nx.Graph()
+    graph.add_edges_from(selected_edges)
+    for component in list(nx.connected_components(graph)):
+        if len(component) < n_cities:
+            cut_edges = tuple(
+                (e[0], e[1])
+                for e in edges
+                if (e[0] in component and e[1] not in component)
+                or (e[0] not in component and e[1] in component)
+            )
+            violations.append(cut_edges)
+    return violations