SIENTIAPDE-1252: Implement artifact generation and MLflow logging for model training results
This commit introduces artifact generation and MLflow logging capabilities to the model training process. It includes the following changes: - Added methods to generate reports, save data files, and log model parameters, metrics, models, and artifacts to MLflow. - Implemented error handling for various scenarios, such as missing files, invalid data, and MLflow connection errors. - Created a new 'header.html' file for report styling and navigation. - Modified the 'model_repository.py' file to include the new artifact generation and MLflow logging methods. - Added comprehensive unit tests to ensure the functionality and robustness of the new features.
This commit is contained in:
Bruno Domingues
@@ -497,3 +497,680 @@ def test_update_production_model(mlflow_repository):
|
||||
'mlflow_run_id': '0',
|
||||
'mlflow_experiment_id': '0',
|
||||
}
|
||||
|
||||
|
||||
# ========== Tests for Model Artifact Generation Methods ==========
|
||||
|
||||
|
||||
def test_get_next_run_name_new(mlflow_repository):
|
||||
"""Test get_next_run_name generates correct run name based on existing runs."""
|
||||
mlflow_repository.model_serving.search_runs_by_name.return_value = [
|
||||
MagicMock(),
|
||||
MagicMock(),
|
||||
MagicMock(),
|
||||
]
|
||||
|
||||
result = mlflow_repository.get_next_run_name_new('test_experiment')
|
||||
|
||||
mlflow_repository.model_serving.search_runs_by_name.assert_called_once_with(
|
||||
experiment_names=['test_experiment'], order_by=['start_time desc']
|
||||
)
|
||||
assert result == 'test_experiment-4'
|
||||
|
||||
|
||||
def test_get_next_run_name_new_first_run(mlflow_repository):
|
||||
"""Test get_next_run_name for first run (no existing runs)."""
|
||||
mlflow_repository.model_serving.search_runs_by_name.return_value = []
|
||||
|
||||
result = mlflow_repository.get_next_run_name_new('new_experiment')
|
||||
|
||||
assert result == 'new_experiment-1'
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_artifacts_success(mock_path, mlflow_repository):
|
||||
"""Test generate_artifacts successfully creates all artifacts."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
# Mock data
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.target_variable = 'target'
|
||||
params.variable_columns = ['feat1', 'feat2']
|
||||
params.experiment_name = 'test_exp'
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = 'test_run-1'
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1, 2], 'feat2': [3, 4]})
|
||||
data.y_train = DataFrame({'target': [5, 6]})
|
||||
data.x_test = DataFrame({'feat1': [7, 8], 'feat2': [9, 10]})
|
||||
data.y_test = DataFrame({'target': [11, 12]})
|
||||
data.regr = MagicMock()
|
||||
data.regr.predict = MagicMock(return_value=np.array([5.1, 6.1]))
|
||||
data.y_pred = np.array([11.1, 12.1])
|
||||
|
||||
# Mock path operations
|
||||
mock_path.exists.return_value = True
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
|
||||
# Mock private methods
|
||||
mlflow_repository._get_reports_directory = MagicMock(return_value='/reports')
|
||||
mlflow_repository._create_run_directory = MagicMock(return_value='/reports/test_run-1_20231010')
|
||||
mlflow_repository._setup_run_directory = MagicMock()
|
||||
mlflow_repository._generate_report = MagicMock(return_value=data)
|
||||
|
||||
result = mlflow_repository.generate_artifacts(data)
|
||||
|
||||
# Assertions
|
||||
mlflow_repository._get_reports_directory.assert_called_once()
|
||||
mlflow_repository._create_run_directory.assert_called_once_with('/reports', 'test_run-1')
|
||||
mlflow_repository._setup_run_directory.assert_called_once()
|
||||
mlflow_repository._generate_report.assert_called_once()
|
||||
assert result == data
|
||||
|
||||
|
||||
def test_generate_artifacts_missing_run_name(mlflow_repository):
|
||||
"""Test generate_artifacts raises ValueError when run_name is not set."""
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = None
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository.generate_artifacts(data)
|
||||
|
||||
assert 'run_name must be set before generating artifacts' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_artifacts_reports_directory_not_exists(mock_path, mlflow_repository):
|
||||
"""Test generate_artifacts raises FileNotFoundError when reports directory doesn't exist."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.target_variable = 'target'
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = 'test_run-1'
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame({'target': [2]})
|
||||
data.x_test = DataFrame({'feat1': [3]})
|
||||
data.y_test = DataFrame({'target': [4]})
|
||||
data.regr = MagicMock()
|
||||
data.y_pred = np.array([4.1])
|
||||
|
||||
mlflow_repository._get_reports_directory = MagicMock(return_value='/reports')
|
||||
mock_path.exists.return_value = False
|
||||
|
||||
with pytest.raises(FileNotFoundError) as exc_info:
|
||||
mlflow_repository.generate_artifacts(data)
|
||||
|
||||
assert 'Reports directory does not exist' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_artifacts_header_file_not_exists(mock_path, mlflow_repository):
|
||||
"""Test generate_artifacts raises FileNotFoundError when header.html doesn't exist."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.target_variable = 'target'
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = 'test_run-1'
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame({'target': [2]})
|
||||
data.x_test = DataFrame({'feat1': [3]})
|
||||
data.y_test = DataFrame({'target': [4]})
|
||||
data.regr = MagicMock()
|
||||
data.regr.predict = MagicMock(return_value=np.array([2.1]))
|
||||
data.y_pred = np.array([4.1])
|
||||
|
||||
mlflow_repository._get_reports_directory = MagicMock(return_value='/reports')
|
||||
mlflow_repository._create_run_directory = MagicMock(return_value='/reports/test_run-1_20231010')
|
||||
|
||||
# First call returns True (reports dir exists), second returns False (header.html doesn't exist)
|
||||
mock_path.exists.side_effect = [True, False]
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
|
||||
with pytest.raises(FileNotFoundError) as exc_info:
|
||||
mlflow_repository.generate_artifacts(data)
|
||||
|
||||
assert 'Header file does not exist' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_success(mock_path, mlflow_repository):
|
||||
"""Test save_run successfully logs all parameters, metrics, models, and artifacts."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.train_size = 80
|
||||
params.removed_intervals = [(1, 10), (20, 30)]
|
||||
params.experiment_name = 'test_exp'
|
||||
params.target_variable = 'target'
|
||||
params.variable_columns = ['feat1', 'feat2']
|
||||
params.lag_train = 5
|
||||
params.lag_val = 3
|
||||
params.window = 10
|
||||
params.low_lim = 0.0
|
||||
params.upp_lim = 1.0
|
||||
params.include_ar = True
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = 'test_run-1'
|
||||
data.params = params
|
||||
data.report_path = '/reports/report.html'
|
||||
data.train_data_path = '/reports/train.csv'
|
||||
data.test_data_path = '/reports/test.csv'
|
||||
data.mse_val = 0.123
|
||||
data.r2_val = 0.987
|
||||
data.mae_val = 0.456
|
||||
data.scaler_dict = {'scaler': 'minmax'}
|
||||
data.process_data = MagicMock()
|
||||
data.regr = MagicMock()
|
||||
|
||||
mock_path.exists.return_value = True
|
||||
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
# Verify experiment was set
|
||||
mlflow_repository.model_serving.set_experiment.assert_called_once_with('test_exp')
|
||||
|
||||
# Verify parameters were logged
|
||||
assert mlflow_repository.model_serving.log_param.call_count == 13
|
||||
|
||||
# Verify metrics were logged
|
||||
mlflow_repository.model_serving.log_metric.assert_any_call('MSE', 0.123)
|
||||
mlflow_repository.model_serving.log_metric.assert_any_call('R2', 0.987)
|
||||
mlflow_repository.model_serving.log_metric.assert_any_call('MAE', 0.456)
|
||||
|
||||
# Verify models were logged
|
||||
mlflow_repository.model_serving.log_model.assert_any_call(data.process_data, 'data_model')
|
||||
mlflow_repository.model_serving.log_model.assert_any_call(data.regr, 'prediction_model')
|
||||
|
||||
# Verify artifacts were logged
|
||||
mlflow_repository.model_serving.log_artifact.assert_any_call('/reports/report.html')
|
||||
mlflow_repository.model_serving.log_artifact.assert_any_call('/reports/train.csv')
|
||||
mlflow_repository.model_serving.log_artifact.assert_any_call('/reports/test.csv')
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_missing_report_path(mock_path, mlflow_repository):
|
||||
"""Test save_run raises ValueError when report_path is missing."""
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.report_path = None
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
assert 'Report file does not exist' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_missing_metrics(mock_path, mlflow_repository):
|
||||
"""Test save_run raises ValueError when metrics are None."""
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.report_path = '/reports/report.html'
|
||||
data.train_data_path = '/reports/train.csv'
|
||||
data.test_data_path = '/reports/test.csv'
|
||||
data.mse_val = None
|
||||
data.r2_val = 0.987
|
||||
data.mae_val = 0.456
|
||||
|
||||
mock_path.exists.return_value = True
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
assert 'One or more metrics (MSE, R2, MAE) are None' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_mlflow_error(mock_path, mlflow_repository):
|
||||
"""Test save_run handles MLflow errors gracefully."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.train_size = 80
|
||||
params.removed_intervals = []
|
||||
params.experiment_name = 'test_exp'
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.run_name = 'test_run-1'
|
||||
data.params = params
|
||||
data.report_path = '/reports/report.html'
|
||||
data.train_data_path = '/reports/train.csv'
|
||||
data.test_data_path = '/reports/test.csv'
|
||||
data.mse_val = 0.123
|
||||
data.r2_val = 0.987
|
||||
data.mae_val = 0.456
|
||||
|
||||
mock_path.exists.return_value = True
|
||||
mlflow_repository.model_serving.set_experiment.side_effect = Exception(
|
||||
'MLflow connection error'
|
||||
)
|
||||
|
||||
with pytest.raises(Exception) as exc_info:
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
assert 'Failed to save run' in str(exc_info.value)
|
||||
assert 'MLflow connection error' in str(exc_info.value)
|
||||
|
||||
|
||||
# ========== Additional Tests for 100% Coverage ==========
|
||||
|
||||
|
||||
def test_init_artifacts_data_empty_x_train(mlflow_repository):
|
||||
"""Test _init_artifacts_data raises ValueError when x_train is empty."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame() # Empty DataFrame
|
||||
data.y_train = DataFrame({'target': [1]})
|
||||
data.x_test = DataFrame({'feat1': [1]})
|
||||
data.y_test = DataFrame({'target': [1]})
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'Training features (x_train) are empty' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_init_artifacts_data_empty_y_train(mlflow_repository):
|
||||
"""Test _init_artifacts_data raises ValueError when y_train is empty."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame() # Empty DataFrame
|
||||
data.x_test = DataFrame({'feat1': [1]})
|
||||
data.y_test = DataFrame({'target': [1]})
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'Training target (y_train) is empty' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_init_artifacts_data_empty_x_test(mlflow_repository):
|
||||
"""Test _init_artifacts_data raises ValueError when x_test is empty."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame({'target': [1]})
|
||||
data.x_test = DataFrame() # Empty DataFrame
|
||||
data.y_test = DataFrame({'target': [1]})
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'Test features (x_test) are empty' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_init_artifacts_data_empty_y_test(mlflow_repository):
|
||||
"""Test _init_artifacts_data raises ValueError when y_test is empty."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame({'target': [1]})
|
||||
data.x_test = DataFrame({'feat1': [1]})
|
||||
data.y_test = DataFrame() # Empty DataFrame
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'Test target (y_test) is empty' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_init_artifacts_data_none_y_pred(mlflow_repository):
|
||||
"""Test _init_artifacts_data raises ValueError when y_pred is None."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1]})
|
||||
data.y_train = DataFrame({'target': [1]})
|
||||
data.x_test = DataFrame({'feat1': [1]})
|
||||
data.y_test = DataFrame({'target': [1]})
|
||||
data.y_pred = None
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'Test predictions (y_pred) are None' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_init_artifacts_data_success(mlflow_repository):
|
||||
"""Test _init_artifacts_data successfully prepares data."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.target_variable = 'target'
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.x_train = DataFrame({'feat1': [1, 2]})
|
||||
data.y_train = DataFrame({'target': [3, 4]})
|
||||
data.x_test = DataFrame({'feat1': [5, 6]})
|
||||
data.y_test = DataFrame({'target': [7, 8]})
|
||||
data.regr = MagicMock()
|
||||
data.regr.predict = MagicMock(return_value=np.array([3.1, 4.1]))
|
||||
data.y_pred = np.array([7.1, 8.1])
|
||||
|
||||
reference_data, current_data = mlflow_repository._init_artifacts_data(data)
|
||||
|
||||
assert 'target' in reference_data.columns
|
||||
assert 'prediction' in reference_data.columns
|
||||
assert 'target' in current_data.columns
|
||||
assert 'prediction' in current_data.columns
|
||||
assert len(reference_data) == 2
|
||||
assert len(current_data) == 2
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.makedirs')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_create_run_directory_success(mock_path, mock_makedirs, mlflow_repository):
|
||||
"""Test _create_run_directory successfully creates directory."""
|
||||
mock_path.join.return_value = '/reports/test_run_20231010_123456_123456'
|
||||
|
||||
result = mlflow_repository._create_run_directory('/reports', 'test_run')
|
||||
|
||||
mock_makedirs.assert_called_once_with('/reports/test_run_20231010_123456_123456', exist_ok=True)
|
||||
assert result == '/reports/test_run_20231010_123456_123456'
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.makedirs')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_create_run_directory_permission_error(mock_path, mock_makedirs, mlflow_repository):
|
||||
"""Test _create_run_directory handles PermissionError."""
|
||||
mock_path.join.return_value = '/reports/test_run_20231010'
|
||||
mock_makedirs.side_effect = PermissionError('Permission denied')
|
||||
|
||||
with pytest.raises(PermissionError) as exc_info:
|
||||
mlflow_repository._create_run_directory('/reports', 'test_run')
|
||||
|
||||
assert 'Permission denied when creating directory' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.makedirs')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_create_run_directory_os_error(mock_path, mock_makedirs, mlflow_repository):
|
||||
"""Test _create_run_directory handles OSError."""
|
||||
mock_path.join.return_value = '/reports/test_run_20231010'
|
||||
mock_makedirs.side_effect = OSError('Disk full')
|
||||
|
||||
with pytest.raises(OSError) as exc_info:
|
||||
mlflow_repository._create_run_directory('/reports', 'test_run')
|
||||
|
||||
assert 'Failed to create directory' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.shutil')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_setup_run_directory_success(mock_path, mock_shutil, mlflow_repository):
|
||||
"""Test _setup_run_directory successfully sets up directory."""
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
mock_open = MagicMock()
|
||||
|
||||
with patch('builtins.open', mock_open):
|
||||
mlflow_repository._setup_run_directory('/run_dir', '/reports/header.html')
|
||||
|
||||
assert mock_open.call_count == 3 # 3 empty files
|
||||
mock_shutil.copy.assert_called_once_with('/reports/header.html', '/run_dir/header.html')
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.shutil')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_setup_run_directory_file_not_found(mock_path, mock_shutil, mlflow_repository):
|
||||
"""Test _setup_run_directory handles FileNotFoundError."""
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
mock_shutil.copy.side_effect = FileNotFoundError('Header not found')
|
||||
|
||||
mock_open = MagicMock()
|
||||
with patch('builtins.open', mock_open):
|
||||
with pytest.raises(FileNotFoundError) as exc_info:
|
||||
mlflow_repository._setup_run_directory('/run_dir', '/reports/header.html')
|
||||
|
||||
assert 'Header file not found' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.shutil')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_setup_run_directory_permission_error(mock_path, mock_shutil, mlflow_repository):
|
||||
"""Test _setup_run_directory handles PermissionError."""
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
|
||||
mock_open = MagicMock()
|
||||
mock_open.side_effect = PermissionError('Permission denied')
|
||||
|
||||
with patch('builtins.open', mock_open):
|
||||
with pytest.raises(PermissionError) as exc_info:
|
||||
mlflow_repository._setup_run_directory('/run_dir', '/reports/header.html')
|
||||
|
||||
assert 'Permission denied when setting up directory' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.shutil')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_setup_run_directory_os_error(mock_path, mock_shutil, mlflow_repository):
|
||||
"""Test _setup_run_directory handles OSError."""
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
|
||||
mock_open = MagicMock()
|
||||
mock_open.side_effect = OSError('Disk error')
|
||||
|
||||
with patch('builtins.open', mock_open):
|
||||
with pytest.raises(OSError) as exc_info:
|
||||
mlflow_repository._setup_run_directory('/run_dir', '/reports/header.html')
|
||||
|
||||
assert 'Failed to setup run directory' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.Reports')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_report_success(mock_path, mock_reports, mlflow_repository):
|
||||
"""Test _generate_report successfully generates all reports."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.variable_columns = ['feat1', 'feat2']
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.run_dir = '/run_dir'
|
||||
|
||||
reference_data = DataFrame(
|
||||
{'feat1': [1.0], 'feat2': [2.0], 'target': [3.0], 'prediction': [3.1]}
|
||||
)
|
||||
current_data = DataFrame({'feat1': [4.0], 'feat2': [5.0], 'target': [6.0], 'prediction': [6.1]})
|
||||
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
mock_report_instance = MagicMock()
|
||||
mock_reports.return_value = mock_report_instance
|
||||
|
||||
# Mock DataFrame.to_csv to avoid actual file writing
|
||||
with patch.object(DataFrame, 'to_csv'):
|
||||
result = mlflow_repository._generate_report(reference_data, current_data, data)
|
||||
|
||||
mock_reports.assert_called_once()
|
||||
mock_report_instance.add_data_quality_section.assert_called_once()
|
||||
mock_report_instance.add_data_drift_section.assert_called_once()
|
||||
mock_report_instance.add_regression_section.assert_called_once()
|
||||
mock_report_instance.save_all_sections_html.assert_called_once()
|
||||
assert result == data
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_report_value_error(mock_path, mlflow_repository):
|
||||
"""Test _generate_report handles ValueError from data conversion."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.run_dir = '/run_dir'
|
||||
|
||||
# DataFrame with non-numeric data
|
||||
reference_data = DataFrame({'feat1': ['a', 'b']})
|
||||
current_data = DataFrame({'feat1': ['c', 'd']})
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._generate_report(reference_data, current_data, data)
|
||||
|
||||
assert 'Failed to convert data to float64' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.Reports')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_report_permission_error(mock_path, mock_reports, mlflow_repository):
|
||||
"""Test _generate_report handles PermissionError."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.variable_columns = ['feat1']
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.run_dir = '/run_dir'
|
||||
|
||||
reference_data = DataFrame({'feat1': [1.0]})
|
||||
current_data = DataFrame({'feat1': [2.0]})
|
||||
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
mock_report_instance = MagicMock()
|
||||
mock_reports.return_value = mock_report_instance
|
||||
mock_report_instance.save_all_sections_html.side_effect = PermissionError('Permission denied')
|
||||
|
||||
with pytest.raises(PermissionError) as exc_info:
|
||||
mlflow_repository._generate_report(reference_data, current_data, data)
|
||||
|
||||
assert 'Permission denied when writing report files' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.Reports')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_report_os_error(mock_path, mock_reports, mlflow_repository):
|
||||
"""Test _generate_report handles OSError."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.variable_columns = ['feat1']
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.run_dir = '/run_dir'
|
||||
|
||||
reference_data = DataFrame({'feat1': [1.0]})
|
||||
current_data = DataFrame({'feat1': [2.0]})
|
||||
|
||||
mock_path.join.side_effect = lambda *args: '/'.join(args)
|
||||
mock_report_instance = MagicMock()
|
||||
mock_reports.return_value = mock_report_instance
|
||||
mock_report_instance.save_all_sections_html.side_effect = OSError('Disk error')
|
||||
|
||||
with pytest.raises(OSError) as exc_info:
|
||||
mlflow_repository._generate_report(reference_data, current_data, data)
|
||||
|
||||
assert 'Failed to generate report' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.Reports')
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_generate_report_run_dir_none(mock_path, mock_reports, mlflow_repository):
|
||||
"""Test _generate_report raises ValueError when run_dir is None."""
|
||||
from model_manager.utils.models.train_model_params import TrainModelParams
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
params = MagicMock(spec=TrainModelParams)
|
||||
params.variable_columns = ['feat1']
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.params = params
|
||||
data.run_dir = None # Not set
|
||||
|
||||
reference_data = DataFrame({'feat1': [1.0]})
|
||||
current_data = DataFrame({'feat1': [2.0]})
|
||||
|
||||
mock_report_instance = MagicMock()
|
||||
mock_reports.return_value = mock_report_instance
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository._generate_report(reference_data, current_data, data)
|
||||
|
||||
assert 'run_dir is not set after directory creation' in str(exc_info.value)
|
||||
|
||||
|
||||
def test_get_reports_directory(mlflow_repository):
|
||||
"""Test _get_reports_directory returns correct path."""
|
||||
result = mlflow_repository._get_reports_directory()
|
||||
|
||||
assert result.endswith('model_manager/reports')
|
||||
assert 'model_manager' in result
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_missing_train_data_path(mock_path, mlflow_repository):
|
||||
"""Test save_run raises ValueError when train_data_path is missing."""
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.report_path = '/reports/report.html'
|
||||
data.train_data_path = None
|
||||
|
||||
mock_path.exists.return_value = True
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
assert 'Training data file does not exist' in str(exc_info.value)
|
||||
|
||||
|
||||
@patch('model_manager.utils.repository.model_repository.path')
|
||||
def test_save_run_missing_test_data_path(mock_path, mlflow_repository):
|
||||
"""Test save_run raises ValueError when test_data_path is missing."""
|
||||
from model_manager.utils.models.train_model_result import TrainModelResult
|
||||
|
||||
data = MagicMock(spec=TrainModelResult)
|
||||
data.report_path = '/reports/report.html'
|
||||
data.train_data_path = '/reports/train.csv'
|
||||
data.test_data_path = None
|
||||
|
||||
mock_path.exists.return_value = True
|
||||
|
||||
with pytest.raises(ValueError) as exc_info:
|
||||
mlflow_repository.save_run(data)
|
||||
|
||||
assert 'Test data file does not exist' in str(exc_info.value)
|
||||
|
||||
Reference in New Issue
Block a user