feat(tests): add tests for models and options and api

tests/eval/test_models.py

@@ -0,0 +1,342 @@
+# SPDX-FileCopyrightText: Copyright (c) 2023 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+# http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+
+import pytest
+from pydantic import ValidationError
+
+from nemoguardrails.eval.models import (
+    ComplianceCheckLog,
+    EvalConfig,
+    EvalOutput,
+    ExpectedOutput,
+    InteractionSet,
+    Policy,
+)
+
+ROOT = os.path.dirname(__file__)
+
+
+def test_interaction_set_expected_output_instantiation():
+    """Test that ExpectedOutput is correctly instantiated based on type."""
+
+    # string type
+    data = {"type": "string", "policy": "test_policy"}
+    interaction_set = InteractionSet.model_validate(
+        {"id": "test_id", "inputs": ["test input"], "expected_output": [data]}
+    )
+    assert len(interaction_set.expected_output) == 1
+    assert interaction_set.expected_output[0].type == "string"
+    assert interaction_set.expected_output[0].policy == "test_policy"
+
+    # dict type
+    data = {"type": "dict", "policy": "test_policy"}
+    interaction_set = InteractionSet.model_validate(
+        {"id": "test_id", "inputs": ["test input"], "expected_output": [data]}
+    )
+    assert len(interaction_set.expected_output) == 1
+    assert interaction_set.expected_output[0].type == "dict"
+    assert interaction_set.expected_output[0].policy == "test_policy"
+
+
+def test_eval_config_from_path():
+    """Test loading config from path."""
+
+    config = EvalConfig.from_path(os.path.join(ROOT, "config_yml"))
+    assert config is not None
+    assert len(config.policies) > 0
+
+
+def test_compliance_check_log():
+    """Test ComplianceCheckLog model."""
+    log = ComplianceCheckLog.model_validate({"id": "test_id", "llm_calls": []})
+    assert log.id == "test_id"
+    assert log.llm_calls == []
+
+
+def test_eval_output():
+    """Test EvalOutput model."""
+    output = EvalOutput.model_validate(
+        {
+            "results": [
+                {
+                    "id": "test_id",
+                    "input": "test input",
+                    "output": "test_output",
+                    "compliance_checks": [
+                        {
+                            "id": "check_id",
+                            "created_at": "2024-01-01T00:00:00",
+                            "method": "test_method",
+                            "compliance": {"policy1": True},
+                            "details": "test details",
+                            "interaction_id": "test_id",
+                        }
+                    ],
+                }
+            ],
+            "logs": [],
+        }
+    )
+    assert len(output.results) == 1
+    assert output.results[0].id == "test_id"
+    assert output.results[0].input == "test input"
+    assert output.results[0].output == "test_output"
+    assert len(output.results[0].compliance_checks) == 1
+    assert output.results[0].compliance_checks[0].id == "check_id"
+    assert output.results[0].compliance_checks[0].interaction_id == "test_id"
+
+
+def test_eval_config_policy_validation_empty_lists():
+    """Test that empty policies and interactions lists are handled correctly."""
+    config = EvalConfig.model_validate(
+        {
+            "policies": [],
+            "interactions": [],
+        }
+    )
+    assert len(config.policies) == 0
+    assert len(config.interactions) == 0
+
+
+def test_eval_config_policy_validation_invalid_policy_format_missing_description():
+    """Test that invalid policy formats are rejected."""
+    with pytest.raises(ValueError):
+        EvalConfig.model_validate(
+            {
+                "policies": [{"id": "policy1"}],
+                "interactions": [],
+            }
+        )
+
+
+def test_eval_config_policy_validation_invalid_interaction_format_missing_inputs():
+    """Test that invalid interaction formats are rejected."""
+    with pytest.raises(ValueError):
+        EvalConfig.model_validate(
+            {
+                "policies": [{"id": "policy1", "description": "Test policy"}],
+                "interactions": [
+                    {
+                        "id": "test_id",
+                        "expected_output": [{"type": "string", "policy": "policy1"}],
+                    }
+                ],
+            }
+        )
+
+
+def test_interaction_set_empty_expected_output():
+    """Test that empty expected_output list is handled correctly."""
+    interaction_set = InteractionSet.model_validate(
+        {"id": "test_id", "inputs": ["test input"], "expected_output": []}
+    )
+    assert len(interaction_set.expected_output) == 0
+
+
+def test_interaction_set_invalid_format():
+    """Test that invalid expected_output format is rejected."""
+    with pytest.raises(ValueError):
+        InteractionSet.model_validate(
+            {
+                "id": "test_id",
+                "inputs": ["test input"],
+                "expected_output": [{"type": "string"}],
+            }
+        )
+
+    # TODO: The model currently doesn't validate the type field values.
+    # This test should pass once type validation is implemented.
+    # with pytest.raises(ValueError):
+    #     InteractionSet.model_validate(
+    #         {
+    #             "id": "test_id",
+    #             "inputs": ["test input"],
+    #             "expected_output": [{"type": "invalid_type", "policy": "test_policy"}],
+    #         }
+    #     )
+
+
+def test_compliance_check_log_invalid_format():
+    """Test that invalid ComplianceCheckLog format is rejected."""
+    with pytest.raises(ValueError):
+        ComplianceCheckLog.model_validate({})
+
+    # invalid llm_calls format
+    with pytest.raises(ValueError):
+        ComplianceCheckLog.model_validate({"id": "test_id", "llm_calls": "invalid"})
+
+
+def test_policy_creation():
+    policy = Policy(
+        id="policy_1",
+        description="Test policy description",
+        weight=50,
+        apply_to_all=False,
+    )
+    assert policy.id == "policy_1"
+    assert policy.description == "Test policy description"
+    assert policy.weight == 50
+    assert not policy.apply_to_all
+
+
+def test_policy_default_values():
+    policy = Policy(
+        id="policy_2",
+        description="Another test policy",
+    )
+    assert policy.weight == 100
+    assert policy.apply_to_all
+
+
+def test_policy_invalid_weight():
+    with pytest.raises(ValidationError):
+        Policy(
+            id="policy_3",
+            description="Invalid weight test",
+            weight="invalid_weight",
+        )
+
+
+def test_expected_output_creation():
+    output = ExpectedOutput(
+        type="refusal",
+        policy="policy_1",
+    )
+    assert output.type == "refusal"
+    assert output.policy == "policy_1"
+
+
+def test_expected_output_missing_field():
+    with pytest.raises(ValidationError):
+        ExpectedOutput(
+            type="refusal",
+        )
+
+
+def test_eval_config_policy_validation_valid():
+    """Test that policy validation works correctly."""
+
+    config = EvalConfig.model_validate(
+        {
+            "policies": [{"id": "policy1", "description": "Test policy"}],
+            "interactions": [
+                {
+                    "id": "test_id",
+                    "inputs": ["test input"],
+                    "expected_output": [{"type": "string", "policy": "policy1"}],
+                }
+            ],
+        }
+    )
+    assert len(config.policies) == 1
+    assert len(config.interactions) == 1
+
+
+def test_eval_config_policy_validation_invalid_policy_not_found():
+    # invalid case, policy not found
+    with pytest.raises(
+        ValueError, match="Invalid policy id policy2 used in interaction set"
+    ):
+        EvalConfig.model_validate(
+            {
+                "policies": [{"id": "policy1", "description": "Test policy"}],
+                "interactions": [
+                    {
+                        "id": "test_id",
+                        "inputs": ["test input"],
+                        "expected_output": [
+                            {
+                                "type": "string",
+                                "policy": "policy2",
+                            }
+                        ],
+                    }
+                ],
+            }
+        )
+
+
+def test_eval_config_policy_validation_multiple_interactions():
+    """Test that policy validation works with multiple interactions."""
+    config = EvalConfig.model_validate(
+        {
+            "policies": [{"id": "policy1", "description": "Test policy"}],
+            "interactions": [
+                {
+                    "id": "test_id1",
+                    "inputs": ["test input 1"],
+                    "expected_output": [{"type": "string", "policy": "policy1"}],
+                },
+                {
+                    "id": "test_id2",
+                    "inputs": ["test input 2"],
+                    "expected_output": [{"type": "string", "policy": "policy1"}],
+                },
+            ],
+        }
+    )
+    assert len(config.interactions) == 2
+
+
+def test_eval_config_policy_validation_multiple_policies():
+    """Test that policy validation works with multiple policies."""
+    config = EvalConfig.model_validate(
+        {
+            "policies": [
+                {"id": "policy1", "description": "Test policy 1"},
+                {"id": "policy2", "description": "Test policy 2"},
+            ],
+            "interactions": [
+                {
+                    "id": "test_id",
+                    "inputs": ["test input"],
+                    "expected_output": [
+                        {"type": "string", "policy": "policy1"},
+                        {"type": "string", "policy": "policy2"},
+                    ],
+                }
+            ],
+        }
+    )
+    assert len(config.policies) == 2
+    assert len(config.interactions[0].expected_output) == 2
+
+
+def test_eval_config_policy_validation_duplicate_policy_ids():
+    """Test that duplicate policy IDs are handled.
+
+    Note: The model currently doesn't validate for duplicate policy IDs.
+    This test should be updated if duplicate policy ID validation is added.
+    """
+    config = EvalConfig.model_validate(
+        {
+            "policies": [
+                {"id": "policy1", "description": "Test policy 1"},
+                {"id": "policy1", "description": "Test policy 2"},
+            ],
+            "interactions": [
+                {
+                    "id": "test_id",
+                    "inputs": ["test input"],
+                    "expected_output": [{"type": "string", "policy": "policy1"}],
+                }
+            ],
+        }
+    )
+    assert len(config.policies) == 2
+    assert config.policies[0].id == "policy1"
+    assert config.policies[1].id == "policy1"