LEADS-2: Fix Path Object Serialization in Amended YAML Files

src/lightspeed_evaluation/core/output/data_persistence.py

@@ -36,7 +36,7 @@ def save_evaluation_data(
         # Save amended data to output directory
         with open(amended_data_path, "w", encoding="utf-8") as f:
             yaml.dump(
-                [conv_data.model_dump() for conv_data in evaluation_data],
+                [conv_data.model_dump(mode="json") for conv_data in evaluation_data],
                 f,
                 default_flow_style=False,
                 sort_keys=False,

tests/unit/core/output/test_data_persistence.py

@@ -0,0 +1,182 @@
+"""Tests for data persistence utilities."""
+
+from pathlib import Path
+
+import yaml
+
+from lightspeed_evaluation.core.models import EvaluationData, TurnData
+from lightspeed_evaluation.core.output.data_persistence import save_evaluation_data
+
+
+def test_save_evaluation_data_with_path_objects(tmp_path: Path) -> None:
+    """Test that Path objects are properly serialized as strings in YAML."""
+    # Create test data with Path objects
+    setup_script_path = Path("config/sample_scripts/setup.sh")
+    cleanup_script_path = Path("config/sample_scripts/cleanup.sh")
+    verify_script_path = Path("config/sample_scripts/verify.sh")
+
+    evaluation_data = [
+        EvaluationData(
+            conversation_group_id="test_conv_1",
+            description="Test conversation with script paths",
+            setup_script=setup_script_path,
+            cleanup_script=cleanup_script_path,
+            turns=[
+                TurnData(
+                    turn_id="turn_1",
+                    query="Test query",
+                    response="Test response",
+                    verify_script=verify_script_path,
+                )
+            ],
+        )
+    ]
+
+    # Create temporary evaluation data file
+    temp_eval_file = tmp_path / "test_evaluation_data.yaml"
+    temp_eval_file.write_text("# Test evaluation data\n")
+
+    # Save evaluation data
+    output_file = save_evaluation_data(
+        evaluation_data=evaluation_data,
+        original_data_path=str(temp_eval_file),
+        output_dir=str(tmp_path),
+    )
+
+    # Verify output file was created
+    assert output_file is not None
+    output_path = Path(output_file)
+    assert output_path.exists()
+
+    # Load the saved YAML and verify Path objects are strings
+    with open(output_path, "r", encoding="utf-8") as f:
+        saved_data = yaml.safe_load(f)
+
+    # Verify the structure
+    assert len(saved_data) == 1
+    conversation = saved_data[0]
+
+    # Verify script paths are strings, not PosixPath objects
+    assert isinstance(conversation["setup_script"], str)
+    assert isinstance(conversation["cleanup_script"], str)
+    assert conversation["setup_script"] == "config/sample_scripts/setup.sh"
+    assert conversation["cleanup_script"] == "config/sample_scripts/cleanup.sh"
+
+    # Verify turn script path is also a string
+    assert len(conversation["turns"]) == 1
+    turn = conversation["turns"][0]
+    assert isinstance(turn["verify_script"], str)
+    assert turn["verify_script"] == "config/sample_scripts/verify.sh"
+
+    # Verify no PosixPath serialization artifacts in the file
+    with open(output_path, "r", encoding="utf-8") as f:
+        file_content = f.read()
+    assert "!!python/object/apply:pathlib.PosixPath" not in file_content
+    assert "PosixPath" not in file_content
+
+
+def test_save_evaluation_data_with_string_paths(tmp_path: Path) -> None:
+    """Test that string paths remain as strings in YAML."""
+    # Create test data with string paths
+    evaluation_data = [
+        EvaluationData(
+            conversation_group_id="test_conv_2",
+            description="Test conversation with string paths",
+            setup_script="sample_scripts/setup.sh",
+            cleanup_script="sample_scripts/cleanup.sh",
+            turns=[
+                TurnData(
+                    turn_id="turn_1",
+                    query="Test query",
+                    response="Test response",
+                    verify_script="sample_scripts/verify.sh",
+                )
+            ],
+        )
+    ]
+
+    # Create temporary evaluation data file
+    temp_eval_file = tmp_path / "test_evaluation_data.yaml"
+    temp_eval_file.write_text("# Test evaluation data\n")
+
+    # Save evaluation data
+    output_file = save_evaluation_data(
+        evaluation_data=evaluation_data,
+        original_data_path=str(temp_eval_file),
+        output_dir=str(tmp_path),
+    )
+
+    # Verify output file was created
+    assert output_file is not None
+    output_path = Path(output_file)
+    assert output_path.exists()
+
+    # Load the saved YAML
+    with open(output_path, "r", encoding="utf-8") as f:
+        saved_data = yaml.safe_load(f)
+
+    # Verify paths are still strings
+    conversation = saved_data[0]
+    assert conversation["setup_script"] == "sample_scripts/setup.sh"
+    assert conversation["cleanup_script"] == "sample_scripts/cleanup.sh"
+    assert conversation["turns"][0]["verify_script"] == "sample_scripts/verify.sh"
+
+
+def test_save_evaluation_data_without_scripts(tmp_path: Path) -> None:
+    """Test that evaluation data without scripts saves correctly."""
+    evaluation_data = [
+        EvaluationData(
+            conversation_group_id="test_conv_3",
+            description="Test conversation without scripts",
+            turns=[
+                TurnData(
+                    turn_id="turn_1",
+                    query="Test query",
+                    response="Test response",
+                )
+            ],
+        )
+    ]
+
+    # Create temporary evaluation data file
+    temp_eval_file = tmp_path / "test_evaluation_data.yaml"
+    temp_eval_file.write_text("# Test evaluation data\n")
+
+    # Save evaluation data
+    output_file = save_evaluation_data(
+        evaluation_data=evaluation_data,
+        original_data_path=str(temp_eval_file),
+        output_dir=str(tmp_path),
+    )
+
+    # Verify output file was created
+    assert output_file is not None
+    assert Path(output_file).exists()
+
+
+def test_save_evaluation_data_error_handling(tmp_path: Path) -> None:
+    """Test error handling when saving evaluation data fails."""
+    evaluation_data = [
+        EvaluationData(
+            conversation_group_id="test_conv_4",
+            turns=[
+                TurnData(
+                    turn_id="turn_1",
+                    query="Test query",
+                )
+            ],
+        )
+    ]
+
+    # Try to save to an invalid path (file instead of directory)
+    invalid_file = tmp_path / "invalid.txt"
+    invalid_file.write_text("Invalid")
+
+    result = save_evaluation_data(
+        evaluation_data=evaluation_data,
+        original_data_path="test.yaml",
+        output_dir=str(invalid_file),  # This is a file, not a directory
+    )
+
+    # Should return None on error
+    assert result is None