def test_custom_stat_report_for_customised_index(self):
mock_df = pd.DataFrame(data=np.random.rand(10, 2),
columns=["attr1", "attr2"])
mock_object = data_exploration.Exploration(data=mock_df)
expected = "given_name attr1"
outcome = mock_object.custom_stat_report(data=mock_df,
name="given_name")
self.assertEqual(outcome.index[0], expected)
def test_descriptive_stats_report_for_output_data_type(self):
mock_df = pd.DataFrame(data=np.random.rand(10, 2),
columns=["attr1", "attr2"])
mock_object = data_exploration.Exploration(data=mock_df)
expected = pd.DataFrame
self.assertEqual(
type(mock_object.descriptive_stats_report(given_data=mock_df)),
expected)
def test_distribution_comparison_for_3_pandas_Series(self):
mock_df = pd.DataFrame(data=np.random.rand(10, 4),
columns=["attr1", "attr2", "attr3", "attr4"])
mock_object = data_exploration.Exploration(data=mock_df)
self.assertIsNone(
mock_object.distribution_comparison(series1=mock_df["attr1"],
series2=mock_df["attr2"],
series3=mock_df["attr3"]))
def test_split_train_test_for_full_and_partial_mode(self):
mock_df = pd.DataFrame(data=np.random.rand(100, 2),
columns=["attr1", "attr2"])
mock_object_full = data_exploration.Exploration(data=mock_df,
running_mode="full")
mock_object_partial = data_exploration.Exploration(
data=mock_df, running_mode="partial")
expected_full = [70, 30]
expected_partial = [5, 2]
self.assertEqual(len(mock_object_full.split_train_test()[0]),
expected_full[0])
self.assertEqual(len(mock_object_full.split_train_test()[1]),
expected_full[1])
self.assertEqual(len(mock_object_partial.split_train_test()[0]),
expected_partial[0])
self.assertEqual(len(mock_object_partial.split_train_test()[1]),
expected_partial[1])
def test_descriptive_stats_report_for_statistics_results(self):
mock_df = pd.DataFrame(data=np.random.rand(10, 2),
columns=["attr1", "attr2"])
mock_object = data_exploration.Exploration(data=mock_df)
expected = mock_df["attr1"].mean()
self.assertEqual(
mock_object.descriptive_stats_report(given_data=mock_df).iloc[0,
0],
expected)
def test_scaler_min_max(self):
mock_df = pd.DataFrame(data=[10, 20, 30], columns=["attr1"])
mock_object = data_exploration.Exploration(data=mock_df)
expected = [1, 0.5, 0]
self.assertAlmostEqual(
mock_object.scaler_min_max(data=mock_df)[:, 0].max(), expected[0])
self.assertTrue(expected[1] in mock_object.scaler_min_max(
data=mock_df))
self.assertAlmostEqual(
mock_object.scaler_min_max(data=mock_df)[:, 0].min(), expected[2])
def test_plot_series_but_ignore_date_for_2_datasets(self):
date_times = pd.date_range(start="04-Jan-2021",
end="15-Jan-2021",
freq="H")
mock_df = pd.DataFrame(data=np.array([
np.random.uniform(1, 3, len(date_times)),
np.random.uniform(3, 5, len(date_times))
]).T,
columns=["attr1", "attr2"],
index=date_times)
mock_object = data_exploration.Exploration(data=mock_df)
self.assertIsNone(
mock_object.plot_series_but_ignore_date(series1=mock_df["attr1"],
series2=mock_df["attr2"]))
def test_apply_sliding_window_for_type_outcome_length_and_links_data_target(
self):
data_length = 200
window_length = parameters.General_Params().slid_win_length
mock_df = pd.DataFrame(data=np.random.rand(data_length, 3),
columns=["attr1", "Close", "attr2"])
mock_object = data_exploration.Exploration(data=mock_df)
df_out_data, df_out_targets = mock_object.apply_sliding_window(
scaled_data=np.array(mock_df))
array_out_data, array_out_targets = mock_object.apply_sliding_window(
scaled_data=np.array(mock_df))
self.assertEqual(len(df_out_data[0]), window_length)
self.assertEqual(len(df_out_data[-1]), window_length)
self.assertEqual(df_out_data[1]["Close"].iloc[-1], df_out_targets[0])
self.assertEqual(df_out_data[5]["Close"].iloc[-1], df_out_targets[4])
self.assertTrue(all(df_out_data[0] == array_out_data[0]))
from tensorflow.keras.losses import MeanAbsoluteError, MeanSquaredError
from core import data_preparation, data_exploration, learning_rate_strategy, model_developement
df = pd.read_csv(os.path.join(os.getcwd(), "data", "GSK per min.csv"))
""" Data Preparation """
prepare = data_preparation.Preparation(data=df)
prepare.set_timestamps_as_index()
prepare.sort_by_timestamp()
prepare.drop_unused()
prepare.transform_to_float()
prepare.time_series_fillna()
prepare.plot_daily_prices(show=True)
prepare.plot_prices_and_projection(show=True)
""" Data Exploration """
explore = data_exploration.Exploration(data=prepare.data)
train_set, test_set = explore.split_train_test()
explore.distribution_comparison(train_set=train_set["Close"],
test_set=test_set["Close"],
show=True)
explore.plot_series_but_ignore_date(train_set=train_set["Close"],
test_set=test_set["Close"],
show=True)
train_report = explore.custom_stat_report(data=train_set, name="Train")
test_report = explore.custom_stat_report(data=train_set, name="Test")
explore.box_plots(data=train_set, show=True)
scaled_train_data = explore.scaler_min_max(data=train_set)
scaled_test_data = explore.scaler_min_max(data=test_set)
sliding_window_train_data, sliding_window_train_target = explore.turn_dfs_into_arrays(
given_data=scaled_train_data)
sliding_window_test_data, sliding_window_test_target = explore.turn_dfs_into_arrays(
def test_box_plots_for_dataframe_with_4_attributes(self):
mock_df = pd.DataFrame(data=np.random.rand(10, 4),
columns=["attr1", "attr2", "attr3", "attr4"])
mock_object = data_exploration.Exploration(data=mock_df)
self.assertIsNone(mock_object.box_plots(data=mock_df))
def test_color_list_for_output_colors(self):
mock_list = data_exploration.Exploration(data=pd.DataFrame(
data=np.random.rand(10), columns=["attr1"])).color_list()
expected = ["red", "aqua"]
self.assertEqual(mock_list[2], expected[0])
self.assertEqual(mock_list[5], expected[1])