[Repo Assist] improve: replace generic Exception with specific exception types in remaining modules

dowhy/api/causal_data_frame.py

@@ -118,7 +118,7 @@ def do(
                     variable_types[key] = self.convert_to_custom_type(all_variables[key].name)
 
         elif len(self._obj.columns) < len(variable_types):
-            raise Exception("Number of variables in the DataFrame is lesser than the variable_types dict")
+            raise ValueError("Number of variables in the DataFrame is lesser than the variable_types dict")
 
         if not self._sampler:
             self._method = method
@@ -160,7 +160,7 @@ def convert_to_custom_type(self, input_type):
         elif "category" in input_type:
             return "d"
         else:
-            raise Exception("{} format is not supported".format(input_type))
+            raise TypeError("{} format is not supported".format(input_type))
 
     def parse_x(self, x):
         if type(x) == str:
@@ -169,4 +169,4 @@ def parse_x(self, x):
             return {xi: None for xi in x}, True
         if type(x) == dict:
             return x, False
-        raise Exception("x format not recognized: {}".format(type(x)))
+        raise TypeError("x format not recognized: {}".format(type(x)))

dowhy/causal_identifier/id_identifier.py

@@ -47,7 +47,7 @@ def get_val(self, return_type: str):
         elif return_type == "sum":
             return self._sum
         else:
-            raise Exception("Provide correct return type.")
+            raise ValueError("Provide correct return type. Must be 'prod' or 'sum'.")
 
     def _print_estimator(self, prefix, estimator: Union[Dict, "IDExpression"] = None, start: bool = False):
         """
@@ -126,7 +126,7 @@ def identify_effect_id(
     try:
         tsort_node_names = OrderedSet(list(nx.topological_sort(graph)))  # topological sorting of graph nodes
     except nx.NetworkXUnfeasible:
-        raise Exception("The graph must be a directed acyclic graph (DAG).")
+        raise ValueError("The graph must be a directed acyclic graph (DAG).")
 
     return __adjacency_matrix_identify_effect(
         adjacency_matrix,

dowhy/causal_prediction/algorithms/base_algorithm.py

@@ -28,7 +28,7 @@ def __init__(self, model, optimizer, lr, weight_decay, betas, momentum):
         # Check if the optimizer is currently supported
         if self.optimizer not in ["Adam", "SGD"]:
             error_msg = self.optimizer + " is not implemented currently. Try Adam or SGD."
-            raise Exception(error_msg)
+            raise ValueError(error_msg)
 
     def training_step(self, train_batch, batch_idx):
         """

dowhy/do_samplers/kernel_density_sampler.py

@@ -40,7 +40,7 @@ def _fit_conditional(self):
         )
 
     def _infer_variable_types(self):
-        raise Exception(
+        raise NotImplementedError(
             "Variable type inference not implemented. Specify variable_types={var_name: var_type}, "
             "where var_type is 'o', 'c', or 'd' for ordered, continuous, or discrete, respectively."
         )

dowhy/do_samplers/mcmc_sampler.py

@@ -87,7 +87,7 @@ def build_bayesian_network(self, g, df):
                 elif g.nodes()[node]["variable_type"] == "b":
                     g.nodes()[node]["variable"] = pm.Bernoulli("{}".format(node), logit_p=mu, observed=df[node])
                 else:
-                    raise Exception("Unrecognized variable type: {}".format(g.nodes()[node]["variable_type"]))
+                    raise ValueError("Unrecognized variable type: {}".format(g.nodes()[node]["variable_type"]))
         return g
 
     def fit_causal_model(self, g, df, data_types, initialization_trace=None):
@@ -99,7 +99,7 @@ def fit_causal_model(self, g, df, data_types, initialization_trace=None):
                 g = self.build_bayesian_network(g, df)
                 trace = pm.sample(1000, tune=1000)
         else:
-            raise Exception("Graph is not a DAG!")
+            raise ValueError("Graph is not a DAG!")
         return g, trace
 
     def sample_prior_causal_model(self, g, df, data_types, initialization_trace):
@@ -111,7 +111,7 @@ def sample_prior_causal_model(self, g, df, data_types, initialization_trace):
                 g = self.build_bayesian_network(g, df)
                 trace = pm.sample_prior_predictive(1)
         else:
-            raise Exception("Graph is not a DAG!")
+            raise ValueError("Graph is not a DAG!")
         return g, trace
 
     def do_x_surgery(self, g, x):

dowhy/gcm/equation_parser.py

@@ -151,7 +151,7 @@ def _extract_noise_model_components(noise_eq: str) -> Tuple[str, dict]:
         parsed_args = _parse_args(args)
         return noise_model_name, parsed_args
     else:
-        raise Exception("Unable to recognise the format or function specified")
+        raise ValueError("Unable to recognise the format or function specified")
 
 
 def _extract_equation_components(equation: str) -> Tuple[str, str]:
@@ -185,7 +185,7 @@ def _add_undefined_nodes_info(causal_nodes_info: dict, present_nodes: list) -> N
 
 def _check_node_redundancy(causal_nodes_info: dict, node_name: str) -> None:
     if node_name in causal_nodes_info:
-        raise Exception(f"The node {node_name} is specified twice which is not allowed.")
+        raise ValueError(f"The node {node_name} is specified twice which is not allowed.")
 
 
 def _sanitize_input_expression(expression: str, banned_characters: list) -> None:

dowhy/utils/api.py

@@ -7,4 +7,4 @@ def parse_state(state):
         return [xi for xi in state.keys()]
     if not state:
         return []
-    raise Exception("Input format for {} not recognized: {}".format(state, type(state)))
+    raise TypeError("Input format for {} not recognized: {}".format(state, type(state)))

dowhy/utils/propensity_score.py

@@ -20,7 +20,7 @@ def propensity_of_treatment_score(data, covariates, treatment, model="logistic",
 
 def state_propensity_score(data, covariates, treatments, variable_types=None):
     if len(set(covariates).intersection(treatments)) != 0:
-        raise Exception("Can't control for causal states. Remove treatment from covariates.")
+        raise ValueError("Can't control for causal states. Remove treatment from covariates.")
     log_propensities = {}
     for i, treatment in enumerate(treatments):
         if variable_types[treatment] in ["b"]:
@@ -36,7 +36,7 @@ def state_propensity_score(data, covariates, treatments, variable_types=None):
                 continuous_treatment_model(data.copy(), covariates + treatments[i + 1 :], treatment, variable_types)
             )
         else:
-            raise Exception(
+            raise ValueError(
                 "Variable type {} for variable {} is not a recognized format type.".format(
                     variable_types[treatment], treatment
                 )
@@ -104,7 +104,7 @@ def get_type_string(variables, variable_types):
         elif variable_types[variable] in ["c"]:
             var_types.append("c")
         else:
-            raise Exception(
+            raise ValueError(
                 "Variable type {} for variable {} not a recognized type.".format(variable_types[variable], variable)
             )
     return "".join(var_types)