Merge pull request #250 from asamal4/temp-fix-multi-provider-script

fix: quick fix to multi-run script as per latest method

Author: asamal4

script/run_multi_provider_eval.py

@@ -28,7 +28,7 @@
 import numpy as np
 import yaml
 
-from lightspeed_evaluation.runner.evaluation import run_evaluation
+from lightspeed_evaluation.runner.evaluation import create_eval_parser, run_evaluation
 
 # Configure logging
 logging.basicConfig(
@@ -128,11 +128,17 @@ def _run_evaluation_worker(
             temp_file.close()
 
         # Run evaluation
-        summary = run_evaluation(
-            system_config_path=str(temp_config_path),
-            evaluation_data_path=eval_data_path,
-            output_dir=str(output_dir),
+        eval_args = create_eval_parser().parse_args(
+            [
+                "--system-config",
+                str(temp_config_path),
+                "--eval-data",
+                eval_data_path,
+                "--output-dir",
+                str(output_dir),
+            ]
         )
+        summary = run_evaluation(eval_args)
 
         # Check result
         if summary is not None:
@@ -473,11 +479,17 @@ def _run_single_evaluation(
             logger.debug(f"Output directory: {output_dir}")
 
             # Run evaluation by calling the function directly
-            summary = run_evaluation(
-                system_config_path=str(temp_config_path),
-                evaluation_data_path=str(self.eval_data_path),
-                output_dir=str(output_dir),
+            eval_args = create_eval_parser().parse_args(
+                [
+                    "--system-config",
+                    str(temp_config_path),
+                    "--eval-data",
+                    str(self.eval_data_path),
+                    "--output-dir",
+                    str(output_dir),
+                ]
             )
+            summary = run_evaluation(eval_args)
 
             # Check result
             if summary is not None:

src/lightspeed_evaluation/runner/evaluation.py

@@ -235,8 +235,8 @@ def run_evaluation(  # pylint: disable=too-many-locals
         return None
 
 
-def main() -> int:
-    """Command line interface."""
+def create_eval_parser() -> argparse.ArgumentParser:
+    """Create the argument parser for the evaluation runner."""
     parser = argparse.ArgumentParser(
         description="LightSpeed Evaluation Framework / Tool",
     )
@@ -274,8 +274,12 @@ def main() -> int:
         action="store_true",
         help="Enable cache warmup mode - rebuild caches without reading existing entries",
     )
+    return parser
+
 
-    eval_args = parser.parse_args()
+def main() -> int:
+    """Command line interface."""
+    eval_args = create_eval_parser().parse_args()
 
     summary = run_evaluation(eval_args)
     return 0 if summary is not None else 1