fix provided arguments check, special case searcher-pipeline_space

neps/utils/run_args_from_yaml.py

@@ -2,11 +2,55 @@
 import logging
 import sys
 from collections.abc import Callable
+from optimizers import BaseOptimizer
 
 import yaml
 
 logger = logging.getLogger("neps")
 
+# Define the allowed parameters based on the arguments of neps.run(), that have a
+# simple value like a string or int type
+# [searcher_kwargs, run_pipeline, preload_hooks] require special handling due to their
+# values necessitating distinct treatment, that's why they are not listed
+EXPECTED_PARAMETERS = [
+    "pipeline_space",
+    "root_directory",
+    "max_evaluations_total",
+    "max_cost_total",
+    "overwrite_working_directory",
+    "post_run_summary",
+    "development_stage_id",
+    "task_id",
+    "max_evaluations_per_run",
+    "continue_until_max_evaluation_completed",
+    "loss_value_on_error",
+    "cost_value_on_error",
+    "ignore_errors",
+    "searcher",
+    "searcher_path",
+]
+
+# Mapping parameter names to their allowed types
+# [task_id, development_stage_id, pre_load_hooks] require special handling of type,
+# that's why they are not listed
+ALLOWED_TYPES = {
+        "run_pipeline": Callable,
+        "root_directory": str,
+        "pipeline_space": str,
+        "max_evaluations_total": int,
+        "max_cost_total": (int, float),
+        "overwrite_working_directory": bool,
+        "post_run_summary": bool,
+        "max_evaluations_per_run": int,
+        "continue_until_max_evaluation_completed": bool,
+        "loss_value_on_error": float,
+        "cost_value_on_error": float,
+        "ignore_errors": bool,
+        "searcher": str,
+        "searcher_path": str,
+        "searcher_kwargs": dict,
+    }
+
 
 def get_run_args_from_yaml(path):
     """
@@ -30,35 +74,16 @@ def get_run_args_from_yaml(path):
     # Load the YAML configuration file
     config = config_loader(path)
 
-    # Define the allowed parameters based on the arguments of neps.run()
-    allowed_parameters = [
-        "pipeline_space",
-        "root_directory",
-        "max_evaluations_total",
-        "max_cost_total",
-        "overwrite_working_directory",
-        "post_run_summary",
-        "development_stage_id",
-        "task_id",
-        "max_evaluations_per_run",
-        "continue_until_max_evaluation_completed",
-        "loss_value_on_error",
-        "cost_value_on_error",
-        "ignore_errors",
-        "searcher",
-        "searcher_path",
-    ]
-
     # Initialize an empty dictionary to hold the extracted settings
     settings = {}
 
     # Flatten yaml file and ignore hierarchical structure, only consider parameters(keys)
-    # with a value
+    # with an explicit value
     flat_config, special_configs = extract_leaf_keys(config)
 
-    # Check if just neps arguments are provided
+    # Check if flatten dict just contains neps arguments
     for parameter, value in flat_config.items():
-        if parameter in allowed_parameters:
+        if parameter in EXPECTED_PARAMETERS:
             settings[parameter] = value
         else:
             raise KeyError(f"Parameter '{parameter}' is not an argument of neps.run().")
@@ -243,24 +268,7 @@ def check_run_args(settings):
     Raises:
         TypeError: For mismatched setting value types.
     """
-    # Mapping parameter names to their allowed types
-    allowed_types = {
-        "run_pipeline": Callable,
-        "root_directory": str,
-        "pipeline_space": str,
-        "max_evaluations_total": int,
-        "max_cost_total": (int, float),
-        "overwrite_working_directory": bool,
-        "post_run_summary": bool,
-        "max_evaluations_per_run": int,
-        "continue_until_max_evaluation_completed": bool,
-        "loss_value_on_error": float,
-        "cost_value_on_error": float,
-        "ignore_errors": bool,
-        "searcher": str,
-        "searcher_path": str,
-        "searcher_kwargs": dict,
-    }
+
     for param, value in settings.items():
         if param == "development_stage_id" or param == "task_id":
             # this argument can be Any
@@ -270,7 +278,7 @@ def check_run_args(settings):
             if not all(callable(item) for item in value):
                 raise TypeError("All items in 'pre_load_hooks' must be callable.")
         else:
-            expected_type = allowed_types[param]
+            expected_type = ALLOWED_TYPES[param]
             if not isinstance(value, expected_type):
                 raise TypeError(
                     f"Parameter '{param}' expects a value of type {expected_type}, got "
@@ -305,7 +313,7 @@ def check_essential_arguments(
     if not root_directory:
         raise ValueError("'root_directory' is required but was not provided.")
     if not pipeline_space:
-        if searcher is not Callable:
+        if not isinstance(searcher, BaseOptimizer):
             raise ValueError("'pipeline_space' is required but was not provided.")
     if not max_evaluation_total and not max_cost_total:
         raise ValueError(