adding test, throwing on unsupported

Signed-off-by: Nicholas Parente <parentenickj@gmail.com>

dowhy/causal_identifier/auto_identifier.py

@@ -909,7 +909,7 @@ def identify_complete_adjustment_set(
             return []
         return [AdjustmentSet(AdjustmentSet.GENERAL, adjustment_set)]
 
-    return [AdjustmentSet(AdjustmentSet.GENERAL, [])]
+    raise ValueError("Exhaustive identification of general adjustment sets is not yet supported.")
 
 
 def identify_mediation(graph: nx.DiGraph, action_nodes: List[str], outcome_nodes: List[str]):

dowhy/graph.py

@@ -204,7 +204,9 @@ def get_descendants(graph: nx.DiGraph, nodes):
 
 def get_proper_causal_path_nodes(graph: nx.DiGraph, action_nodes, outcome_nodes):
     # Process is described in Van Der Zander et al. "Constructing Separators and
-    # Adjustment Sets in Ancestral Graphs", Section 4.1
+    # Adjustment Sets in Ancestral Graphs", Section 4.1.
+
+    # We cannot user do_surgery() since we require deep copies of the given graph.
 
     # 1) Create modified graphs removing inbound and outbound arrows from the action nodes, respectively.
     graph_post_interv = copy.deepcopy(graph)  # remove incoming arrows to our action nodes
@@ -216,17 +218,17 @@ def get_proper_causal_path_nodes(graph: nx.DiGraph, action_nodes, outcome_nodes)
 
     # 2) Use the modified graphs to identify the nodes which lie on proper causal paths from the
     # action nodes to the outcome nodes.
-    de_x = get_descendants(graph_post_interv, action_nodes)
-    an_y = get_ancestors(graph_with_action_nodes_as_sinks, outcome_nodes)
+    de_x = get_descendants(graph_post_interv, action_nodes).union(action_nodes)
+    an_y = get_ancestors(graph_with_action_nodes_as_sinks, outcome_nodes).union(outcome_nodes)
     return (set(de_x) - set(action_nodes)) & an_y
 
 
 def get_proper_backdoor_graph(graph: nx.DiGraph, action_nodes, outcome_nodes):
     # Process is described in Van Der Zander et al. "Constructing Separators and
-    # Adjustment Sets in Ancestral Graphs", Section 4.1
+    # Adjustment Sets in Ancestral Graphs", Section 4.1.
 
     # First we can just call get_proper_causal_path_nodes, then
-    # we remove edges from the action_nodes to the proper causal path nodes
+    # we remove edges from the action_nodes to the proper causal path nodes.
     graph_pbd = copy.deepcopy(graph)
     graph_pbd.remove_edges_from(
         [(u, v) for u in action_nodes for v in get_proper_causal_path_nodes(graph, action_nodes, outcome_nodes)]

tests/causal_identifiers/base.py

@@ -52,12 +52,14 @@ def __init__(
         self,
         graph_str,
         observed_variables,
+        action_nodes,
+        outcome_nodes,
         minimal_adjustment_sets,
         exhaustive_adjustment_sets,
     ):
         self.graph = build_graph_from_str(graph_str)
-        self.action_nodes = ["X"]
-        self.outcome_nodes = ["Y"]
+        self.action_nodes = action_nodes
+        self.outcome_nodes = outcome_nodes
         self.observed_nodes = observed_variables
         self.minimal_adjustment_sets = minimal_adjustment_sets
         self.exhaustive_adjustment_sets = exhaustive_adjustment_sets

tests/causal_identifiers/example_graphs.py

@@ -394,9 +394,21 @@
     "shpitser_simple_non_backdoor_adjustment_set": dict(
         graph_str="digraph{Z;X;Y; X->Z;X->Y}",
         observed_variables=["Z", "X", "Y"],
+        action_nodes=["X"],
+        outcome_nodes=["Y"],
         minimal_adjustment_sets=[{}],
         exhaustive_adjustment_sets=[{"Z"}, {}],
-    )
+    ),
+    # Example is selected from van der Zander et al. "Constructing Separators and Adjustment Sets in Ancestral
+    # Graphs", figure 2.
+    "van_der_zander_minimal_non_backdoor_adjustment_set": dict(
+        graph_str="digraph{Z1;Z2;X1;X2;Y1;Y2; X1->Y1;X1->Z1;Z1->Z2;Z2->X2;Y2->Z2}",
+        observed_variables=["Z1", "Z2", "X1", "X2", "Y1", "Y2"],
+        action_nodes=["X1", "X2"],
+        outcome_nodes=["Y1", "Y2"],
+        minimal_adjustment_sets=[{"Z1", "Z2"}],
+        exhaustive_adjustment_sets=["Z1", "Z2"],
+    ),
 }
 
 

tests/causal_identifiers/test_complete_adjustment_identifier.py

@@ -8,35 +8,15 @@
 
 
 class TestGeneralAdjustmentIdentification(object):
-    def test_identify_exhaustive_adjustment(
-        self, example_complete_adjustment_graph_solution: IdentificationTestGeneralCovariateAdjustmentGraphSolution
-    ):
-        graph = example_complete_adjustment_graph_solution.graph
-        expected_sets = example_complete_adjustment_graph_solution.exhaustive_adjustment_sets
-        adjustment_set_results = identify_complete_adjustment_set(
-            graph,
-            action_nodes=["X"],
-            outcome_nodes=["Y"],
-            observed_nodes=example_complete_adjustment_graph_solution.observed_nodes,
-            covariate_adjustment=CovariateAdjustment.COVARIATE_ADJUSTMENT_EXHAUSTIVE,
-        )
-        adjustment_sets = [
-            set(adjustment_set.get_variables())
-            for adjustment_set in adjustment_set_results
-            if len(adjustment_set.get_variables()) > 0
-        ]
-
-        assert all((len(s) == 0 and len(adjustment_sets) == 0) or set(s) in adjustment_sets for s in expected_sets)
-
     def test_identify_minimal_adjustment(
         self, example_complete_adjustment_graph_solution: IdentificationTestGeneralCovariateAdjustmentGraphSolution
     ):
         graph = example_complete_adjustment_graph_solution.graph
         expected_set = example_complete_adjustment_graph_solution.minimal_adjustment_sets[0]
         adjustment_set_results = identify_complete_adjustment_set(
             graph,
-            action_nodes=["X"],
-            outcome_nodes=["Y"],
+            action_nodes=example_complete_adjustment_graph_solution.action_nodes,
+            outcome_nodes=example_complete_adjustment_graph_solution.outcome_nodes,
             observed_nodes=example_complete_adjustment_graph_solution.observed_nodes,
             covariate_adjustment=CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT,
         )