PyTest added

0_meal_identification/meal_identification/meal_identification/tests/test_modeling2.py

@@ -0,0 +1,142 @@
+#import pytest
+from unittest.mock import MagicMock
+from pathlib import Path
+import pandas as pd
+from pathlib import Path
+import sys
+import os
+
+
+# Mocking external dependencies
+sys.modules['dataset_operations'] = MagicMock()
+sys.modules['dataset_cleaner'] = MagicMock()
+
+# Importing from the project
+from meal_identification.modeling.train import (
+    ScaledLogitTransformer, GMMHMM, train_model_instance, 
+    load_data, xy_split, process_labels, load_model, save_model
+)
+from meal_identification.config import (
+    MODELS_DIR,
+    PROCESSED_DATA_DIR,
+    INTERIM_DATA_DIR
+)
+
+# Define a fixture for sample data
+import pytest
+import pandas as pd
+
+@pytest.fixture
+def sample_data():
+    """
+    Fixture to provide a sample dataframe loaded from a CSV file.
+    """
+    file_path = "0_meal_identification/meal_identification/data/interim/2024-11-15_500030__i5mins_d4hrs_c5g_l3hrs_n4.csv"
+    return pd.read_csv(file_path)
+
+
+@pytest.fixture
+def mock_model_paths(mocker):
+    """
+    Mock the paths for the model and data directories to use actual paths from the project structure.
+    """
+    mocker.patch("meal_identification.config.MODELS_DIR", Path("models/GaussianHMM_model"))
+    mocker.patch("meal_identification.config.INTERIM_DATA_DIR", Path("data/interim"))
+    mocker.patch("meal_identification.config.PROCESSED_DATA_DIR", Path("data/processed"))
+
+
+def test_load_data(mocker, mock_model_paths, sample_data):
+    """
+    Test the load_data function for loading data from a CSV file.
+    """
+    # Mock the load_data function to return sample_data
+    mocker.patch("meal_identification.modeling.train.load_data", return_value=sample_data)
+
+    # Test function call
+    data_path = INTERIM_DATA_DIR / "test_data.csv"
+    result = load_data(data_path)
+
+    # Assertions
+    assert result.shape[1] == 10, "The dataframe does not have the expected number of columns."
+    assert "bgl" in result.columns, "'bgl' column is missing in the dataframe."
+
+
+def test_xy_split(sample_data):
+    """
+    Test splitting data into features (X) and target (Y).
+    """
+    X, Y = xy_split(sample_data)
+    assert X.shape == (3, 1), "Features (X) do not have the correct shape."
+    assert Y.shape == (3, 1), "Target (Y) does not have the correct shape."
+
+
+def test_process_labels(sample_data):
+    """
+    Test the label processing for binary classification.
+    """
+    _, Y = xy_split(sample_data)
+    processed_Y = process_labels(Y)
+    assert set(processed_Y["msg_type"]) == {0, 1}, "Labels are not correctly processed."
+
+
+def test_save_model(mocker):
+    """
+    Test saving the model to the specified path.
+    """
+    # Mocking the model and save function
+    mock_model = MagicMock()
+    mocker.patch("meal_identification.modeling.train.save_model")
+
+    # Call save_model
+    model_path = MODELS_DIR / "test_model.pkl"
+    save_model(mock_model, model_path)
+
+    # Verify the function was called
+    save_model.assert_called_once_with(mock_model, model_path)
+
+
+def test_train_model_instance(mocker, mock_model_paths):
+    """
+    Test training the model using train_model_instance.
+    """
+    # Mock the train_model_instance function
+    mocker.patch("meal_identification.modeling.train.train_model_instance", return_value=MagicMock())
+
+    # Call train_model_instance
+    data_path = INTERIM_DATA_DIR / "test_data.csv"
+    model_path = MODELS_DIR / "test_model.pkl"
+    model = train_model_instance(
+        model="GMMHMM",
+        data_path=data_path,
+        model_path=model_path,
+        transformer=ScaledLogitTransformer()
+    )
+
+    # Verify the return type and functionality
+    assert model is not None, "Model training returned None."
+    assert isinstance(model, MagicMock), "Model is not of the expected type."
+
+
+def test_train_full_model_pipeline(mocker, mock_model_paths):
+    """
+    Test the full model training process, including saving the model.
+    """
+    # Mock the train_model_instance and save_model functions
+    mock_train_model_instance = mocker.patch("meal_identification.modeling.train.train_model_instance", return_value=MagicMock())
+    mock_save_model = mocker.patch("meal_identification.modeling.train.save_model")
+
+    # Perform full model training
+    data_path = INTERIM_DATA_DIR / "test_data.csv"
+    model_path = MODELS_DIR / "test_model.pkl"
+    model = train_model_instance(
+        model="GMMHMM",
+        data_path=data_path,
+        model_path=model_path,
+        transformer=ScaledLogitTransformer()
+    )
+    save_model(model, model_path)
+
+    # Assertions
+    mock_train_model_instance.assert_called_once()
+    mock_save_model.assert_called_once_with(model, model_path)
+    assert model is not None, "Full training pipeline failed to return a trained model."