poe format

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

dowhy/causal_identifier/auto_identifier.py

@@ -1,8 +1,8 @@
+import copy
 import itertools
 import logging
 from enum import Enum
 from typing import Dict, List, Optional, Union
-import copy
 
 import networkx as nx
 import sympy as sp
@@ -21,9 +21,9 @@
     get_backdoor_paths,
     get_descendants,
     get_instruments,
-    has_directed_path,
+    get_proper_backdoor_graph,
     get_proper_causal_path_nodes,
-    get_proper_backdoor_graph
+    has_directed_path,
 )
 from dowhy.utils.api import parse_state
 
@@ -51,11 +51,13 @@ class BackdoorAdjustment(Enum):
     BACKDOOR_MIN_EFFICIENT = "efficient-minimal-adjustment"
     BACKDOOR_MINCOST_EFFICIENT = "efficient-mincost-adjustment"
 
+
 class CovariateAdjustment(Enum):
     # Covariate adjustment method names
     COVARIATE_ADJUSTMENT_DEFAULT = "default"
     COVARIATE_ADJUSTMENT_EXHAUSTIVE = "exhaustive-search"
 
+
 MAX_BACKDOOR_ITERATIONS = 100000
 
 METHOD_NAMES = {
@@ -119,7 +121,7 @@ def identify_effect(
             self.backdoor_adjustment,
             self.optimize_backdoor,
             self.costs,
-            self.covariate_adjustment
+            self.covariate_adjustment,
         )
 
         estimand.identifier = self
@@ -158,7 +160,7 @@ def identify_effect_auto(
     backdoor_adjustment: BackdoorAdjustment = BackdoorAdjustment.BACKDOOR_DEFAULT,
     optimize_backdoor: bool = False,
     costs: Optional[List] = None,
-    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT
+    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT,
 ) -> IdentifiedEstimand:
     """Main method that returns an identified estimand (if one exists).
 
@@ -198,7 +200,7 @@ def identify_effect_auto(
             estimand_type,
             costs,
             conditional_node_names,
-            covariate_adjustment
+            covariate_adjustment,
         )
     elif estimand_type == EstimandType.NONPARAMETRIC_NDE:
         return identify_nde_effect(
@@ -233,7 +235,7 @@ def identify_ate_effect(
     estimand_type: EstimandType,
     costs: List,
     conditional_node_names: List[str] = None,
-    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT
+    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT,
 ):
     estimands_dict = {}
     mediation_first_stage_confounders = None
@@ -308,7 +310,9 @@ def identify_ate_effect(
     ### 4. GENERAL ADJUSTMENT IDENTIFICATION
     # This generalizes the backdoor criterion, identifying other valid covariate adjustment sets that might not
     # satisfy the backdoor criterion.
-    adjustment_sets = identify_complete_adjustment_set(graph, action_nodes, outcome_nodes, observed_nodes, covariate_adjustment)
+    adjustment_sets = identify_complete_adjustment_set(
+        graph, action_nodes, outcome_nodes, observed_nodes, covariate_adjustment
+    )
     logger.info("Number of general adjustment sets found: " + str(len(adjustment_sets)))
     estimands_dict, adjusment_variables_dict = build_adjustment_set_estimands_dict(
         action_nodes, outcome_nodes, observed_nodes, adjustment_sets, estimands_dict
@@ -762,7 +766,8 @@ def get_default_adjustment_set_id(
     # Default set contains minimum possible number of instrumental variables, to prevent lowering variance in the treatment variable.
     instrument_names = set(get_instruments(graph, action_nodes, outcome_nodes))
     iv_count_dict = {
-        key: len(set(adjustment_set).intersection(instrument_names)) for key, adjustment_set in adjustment_sets_dict.items()
+        key: len(set(adjustment_set).intersection(instrument_names))
+        for key, adjustment_set in adjustment_sets_dict.items()
     }
     min_iv_count = min(iv_count_dict.values())
     min_iv_keys = {key for key, iv_count in iv_count_dict.items() if iv_count == min_iv_count}
@@ -783,7 +788,7 @@ def build_adjustment_set_estimands_dict(
     outcome_names: List[str],
     observed_nodes: List[str],
     adjustment_sets: List[AdjustmentSet],
-    estimands_dict: Dict
+    estimands_dict: Dict,
 ):
     """Build the final dict for adjustment sets by filtering unobserved variables if needed."""
     adjustment_variables_dict = {}
@@ -884,7 +889,7 @@ def identify_complete_adjustment_set(
     action_nodes: List[str],
     outcome_nodes: List[str],
     observed_nodes: List[str],
-    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT
+    covariate_adjustment: CovariateAdjustment = CovariateAdjustment.COVARIATE_ADJUSTMENT_DEFAULT,
 ) -> List[AdjustmentSet]:
 
     graph_pbd = get_proper_backdoor_graph(graph, action_nodes, outcome_nodes)
@@ -897,7 +902,7 @@ def identify_complete_adjustment_set(
             set(action_nodes),
             set(outcome_nodes),
             # Require the adjustment set to consist only of observed nodes
-            restricted=((set(graph_pbd.nodes) - set(pcp_nodes)) & set(observed_nodes))
+            restricted=((set(graph_pbd.nodes) - set(pcp_nodes)) & set(observed_nodes)),
         )
         if adjustment_set is None:
             logger.info("No adjustment sets found.")

dowhy/causal_identifier/backdoor.py

@@ -118,7 +118,7 @@ def get_backdoor_vars(self):
                         AdjustmentSet(
                             _type=AdjustmentSet.BACKDOOR,
                             variables=tuple(obj.find_set()),
-                            num_paths_blocked_by_observed_nodes=obj.num_sets()
+                            num_paths_blocked_by_observed_nodes=obj.num_sets(),
                         )
                     )
 

dowhy/causal_identifier/identified_estimand.py

@@ -26,7 +26,7 @@ def __init__(
         default_backdoor_id=None,
         identifier_method=None,
         no_directed_path=False,
-        default_adjustment_set_id=None
+        default_adjustment_set_id=None,
     ):
         self.identifier = identifier
         self.treatment_variable = parse_state(treatment_variable)

dowhy/graph.py

@@ -1,11 +1,11 @@
 """This module defines the fundamental interfaces and functions related to causal graphs."""
 
+import copy
 import itertools
 import logging
 import re
 from abc import abstractmethod
 from typing import Any, List, Protocol
-import copy
 
 import networkx as nx
 from networkx.algorithms.dag import has_cycle
@@ -234,7 +234,6 @@ def get_proper_backdoor_graph(graph: nx.DiGraph, action_nodes, outcome_nodes):
     return graph_pbd
 
 
-
 def check_dseparation(graph: nx.DiGraph, nodes1, nodes2, nodes3, new_graph=None, dseparation_algo="default"):
     if dseparation_algo == "default":
         if new_graph is None:

tests/causal_identifiers/base.py

@@ -2,7 +2,11 @@
 
 from dowhy.graph import build_graph_from_str
 
-from .example_graphs import TEST_FRONTDOOR_GRAPH_SOLUTIONS, TEST_GRAPH_SOLUTIONS, TEST_GRAPH_SOLUTIONS_COMPLETE_ADJUSTMENT
+from .example_graphs import (
+    TEST_FRONTDOOR_GRAPH_SOLUTIONS,
+    TEST_GRAPH_SOLUTIONS,
+    TEST_GRAPH_SOLUTIONS_COMPLETE_ADJUSTMENT,
+)
 
 
 class IdentificationTestGraphSolution(object):
@@ -68,7 +72,9 @@ def example_graph_solution(request):
 def example_frontdoor_graph_solution(request):
     return IdentificationTestFrontdoorGraphSolution(**TEST_FRONTDOOR_GRAPH_SOLUTIONS[request.param])
 
+
 @pytest.fixture(params=TEST_GRAPH_SOLUTIONS_COMPLETE_ADJUSTMENT.keys())
 def example_complete_adjustment_graph_solution(request):
-    return IdentificationTestGeneralCovariateAdjustmentGraphSolution(**TEST_GRAPH_SOLUTIONS_COMPLETE_ADJUSTMENT[request.param])
-
+    return IdentificationTestGeneralCovariateAdjustmentGraphSolution(
+        **TEST_GRAPH_SOLUTIONS_COMPLETE_ADJUSTMENT[request.param]
+    )

tests/causal_identifiers/test_complete_adjustment_identifier.py

@@ -1,15 +1,16 @@
 import pytest
 
 from dowhy.causal_identifier import AutoIdentifier, CovariateAdjustment
-from dowhy.causal_identifier.identify_effect import EstimandType
-
 from dowhy.causal_identifier.auto_identifier import identify_complete_adjustment_set
+from dowhy.causal_identifier.identify_effect import EstimandType
 
 from .base import IdentificationTestGeneralCovariateAdjustmentGraphSolution, example_complete_adjustment_graph_solution
 
 
 class TestGeneralAdjustmentIdentification(object):
-    def test_identify_exhaustive_adjustment(self, example_complete_adjustment_graph_solution: IdentificationTestGeneralCovariateAdjustmentGraphSolution):
+    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(
@@ -25,12 +26,11 @@ def test_identify_exhaustive_adjustment(self, example_complete_adjustment_graph_
             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
-        )
+        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):
+    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(
@@ -47,6 +47,3 @@ def test_identify_minimal_adjustment(self, example_complete_adjustment_graph_sol
         ]
 
         assert (len(expected_set) == 0 and len(adjustment_sets) == 0) or set(expected_set) in adjustment_sets
-
-
-