autogaita 1.5.1__py3-none-any.whl → 1.5.2__py3-none-any.whl

tests/test_group_unit.py

@@ -0,0 +1,393 @@
+from autogaita.group.group_main import (
+    group,
+    import_data,
+    avg_and_std,
+    PCA_main,
+    create_stats_df,
+    cluster_extent_test,
+    anova_design_sanity_check,
+    tukeys_only_info_message,
+    ANOVA_main,
+    plot_results,
+    print_finish,
+)
+from autogaita.group.group_1_preparation import some_prep
+from autogaita.group.group_2_data_processing import (
+    load_previous_runs_dataframes,
+    check_PCA_and_stats_variables,
+)
+from autogaita.group.group_3_PCA import run_PCA, convert_PCA_bins_to_list
+from autogaita.group.group_4_stats import run_ANOVA, multcompare_SC_Percentages
+from autogaita.resources.utils import bin_num_to_percentages
+import os
+import math
+import pytest
+from sklearn import datasets
+import pandas as pd
+import pandas.testing as pdt
+import numpy as np
+
+
+# %%................................  fixtures  ........................................
+@pytest.fixture
+def extract_folderinfo(tmp_path):
+    return {
+        "group_names": ["group1", "group2"],
+        "group_dirs": ["/path/to/group1", "/path/to/group2"],
+        "results_dir": tmp_path,
+        "load_dir": "",
+    }
+
+
+@pytest.fixture
+def extract_cfg():
+    return {
+        "do_permtest": True,
+        "do_anova": True,
+        "permutation_number": 100,
+        "PCA_n_components": 3,
+        "PCA_custom_scatter_PCs": "",
+        "PCA_save_3D_video": False,  # True
+        "PCA_bins": "",
+        "stats_threshold": 0.05,
+        "plot_SE": False,
+        "color_palette": "viridis",
+        "dont_show_plots": True,
+        "legend_outside": True,
+        "which_leg": "left",
+        "anova_design": "Mixed ANOVA",
+        "PCA_variables": [],
+        "stats_variables": [],
+    }
+
+
+# %%..............................  1. preparation  ....................................
+def test_some_preps_cfg_files(extract_folderinfo, extract_cfg):
+    """Check for equivalence of the cfg file using load dir and the cfg dict generated by some_prep"""
+    # 1. Ensure that the cfg is equal to that used in group approval (with repo data)
+    extract_folderinfo["group_names"] = ["5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["group_dirs"] = [
+        "example data/5mm/Results/",
+        "example data/12mm/Results/",
+        "example data/25mm/Results/",
+    ]
+    extract_cfg["do_permtest"] = False
+    extract_cfg["anova_design"] = "RM ANOVA"
+    extract_cfg["PCA_variables"] = [
+        "Knee y",
+        "Ankle y",
+        "Hind paw tao y",
+        "Ankle Angle",
+        "Knee Angle",
+        "Hip Angle",
+    ]
+    extract_cfg["stats_variables"] = extract_cfg["PCA_variables"]
+    # 2. Run some_prep twice with and without load dir & test cfg match
+    extract_folderinfo, generated_cfg = some_prep(extract_folderinfo, extract_cfg)
+    extract_folderinfo["load_dir"] = "example data/group"
+    extract_folderinfo, loaded_cfg = some_prep(extract_folderinfo, extract_cfg)
+    # 3. Check that cfgs are result of different processes (being loaded or not) and
+    #    not exact copies (e.g. references to the same object) for some Python-reason
+    assert "loaded" in loaded_cfg.keys()
+    assert "loaded" not in generated_cfg.keys()
+    # 4. Check that the cfgs are equivalent
+    loaded_cfg.pop("loaded")  # except this key of course - get rid of it
+    generated_cfg == loaded_cfg
+    # 5. Finally change some cfg key and assert that it has not been overwritten by
+    # loading a previous run's cfg
+    # => might seem unnecessary but the way this was implemented initially did
+    #    overwrite cfgs due to how it was loading stuff
+    extract_cfg["do_permtest"] = True
+    extract_cfg["anova_design"] = "Mixed ANOVA"
+    extract_cfg["PCA_variables"] = ["Ankle Angle", "Knee Angle", "Hip Angle"]
+    extract_folderinfo, loaded_cfg_2 = some_prep(extract_folderinfo, extract_cfg)
+    differing_keys = ["do_permtest", "anova_design", "PCA_variables"]
+    for key in differing_keys:
+        assert loaded_cfg_2[key] != loaded_cfg[key]
+
+
+# %%............................  2. data processing  ..................................
+def test_check_PCA_and_stats_variables(extract_folderinfo, extract_cfg):
+    """Test this sanity check does not have the necessary columns"""
+    group_names = ["5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["group_names"] = group_names
+    extract_folderinfo["load_dir"] = "example data/group"
+    # changing PCA/stats vars like this already performs a first test because the load
+    # function only runs successfully, if the test_PCA_.. function does
+    # => i.e. the features are present in the all 3 kinds of dfs
+    extract_cfg["PCA_variables"] = ["Ankle Angle", "Knee Angle"]
+    extract_cfg["stats_variables"] = ["Ankle Angle", "Knee Angle"]
+    avg_dfs, _, _, extract_cfg = load_repos_group_data(extract_folderinfo, extract_cfg)
+    # test correct failure
+    # => code means: test that after removing the column our check correctly raises a
+    #    ValueError
+    avg_dfs[0].drop(columns=["Ankle Angle"], inplace=True)
+    with pytest.raises(ValueError):
+        check_PCA_and_stats_variables(
+            avg_dfs[0], "5 mm", "Average", extract_folderinfo, extract_cfg
+        )
+
+
+def test_smoke_load_dir(extract_folderinfo, extract_cfg):
+    """Test that load dir runs through without errors and produces Results"""
+    extract_folderinfo["group_names"] = ["5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["load_dir"] = "example data/group"
+    group(extract_folderinfo, extract_cfg)
+    # I checked and hardcoded the 39 here, shouldn't change.
+    # If it does I want to know and this should fail
+    assert len(os.listdir(extract_folderinfo["results_dir"])) == 39
+
+
+def test_load_previous_runs_dataframes(extract_folderinfo, extract_cfg):
+    """Testing if errors are raised correctly and if the loaded dfs are eqiuvalent to the ones import_data generates"""
+    # NOTE
+    # ----
+    # => last tests fail on git actions but not locally
+    # => quite sure it's because of how git actions' os lists the files when loading
+    # => we have a similar issue in test_group_approval see os.getenv("CI") there
+    # 1: fails as wanted if group name wrong (df not found in load dir)
+    extract_folderinfo["group_names"] = ["not 5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["load_dir"] = "example data/group"
+    with pytest.raises(FileNotFoundError):
+        avg_dfs, _, _, extract_cfg = load_repos_group_data(
+            extract_folderinfo, extract_cfg
+        )
+    # 2: avg_dfs equivalent to import_data's avg_dfs
+    extract_folderinfo["group_names"] = ["5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["group_dirs"] = [  # for "import_data" later
+        "example data/5mm/Results/",
+        "example data/12mm/Results/",
+        "example data/25mm/Results/",
+    ]
+    avg_dfs, _, _, extract_cfg = load_repos_group_data(extract_folderinfo, extract_cfg)
+    # some prep required for import data & avg_and_std
+    extract_cfg["sampling_rate"] = 100
+    extract_cfg["bin_num"] = 25
+    extract_cfg["save_to_xls"] = [True, True, True]
+    extract_cfg["tracking_software"] = "DLC"
+    extract_cfg["analyse_average_x"] = True
+    i_dfs, _, extract_cfg = import_data(extract_folderinfo, extract_cfg)
+    i_avg_dfs, _ = avg_and_std(i_dfs, extract_folderinfo, extract_cfg)
+    if not os.getenv("CI"):
+        for g in range(len(extract_folderinfo["group_names"])):
+            pdt.assert_frame_equal(
+                avg_dfs[g], i_avg_dfs[g], check_dtype=False, check_exact=False
+            )
+    # 3: check that it also works for Universal 3D
+    # => only do it locally, I don't want to upload the MoVi Data to git
+    if not os.getenv("CI"):
+        extract_folderinfo["group_names"] = ["crawling", "jogging"]
+        extract_folderinfo["group_dirs"] = [
+            "/Users/mahan/sciebo/Research/AutoGaitA/Showcase 3/MOVI/Final/crawling/Results/",
+            "/Users/mahan/sciebo/Research/AutoGaitA/Showcase 3/MOVI/Final/jogging/Results/",
+        ]
+        extract_folderinfo["load_dir"] = (
+            "/Users/mahan/sciebo/Research/AutoGaitA/Showcase 3/MOVI/Unit Test of load_prev_groupruns_dfs/"
+        )
+        avg_dfs, _, _, extract_cfg = load_repos_group_data(
+            extract_folderinfo, extract_cfg
+        )
+        extract_cfg["tracking_software"] = "Universal 3D"
+        extract_cfg["which_leg"] = "right"
+        extract_cfg["analyse_average_y"] = True
+        i_dfs, _, extract_cfg = import_data(extract_folderinfo, extract_cfg)
+        i_avg_dfs, _ = avg_and_std(i_dfs, extract_folderinfo, extract_cfg)
+        if not os.getenv("CI"):
+            for g in range(len(extract_folderinfo["group_names"])):
+                pdt.assert_frame_equal(
+                    avg_dfs[g], i_avg_dfs[g], check_dtype=False, check_exact=False
+                )
+
+
+# %%..............................  4. statistics  .....................................
+
+
+def test_load_dfs_stats_df_perm_test_smoke(extract_folderinfo, extract_cfg):
+    """Smoke test three functions:
+    1. load_previous_runs-dataframes
+    2. create_stats_df
+    3. cluster_extent_test
+    => Check if it all runs without errors on our repo's example data
+    """
+    # first, prepare approrpiately & overwrite the fixtures of this script as needed
+    # 1) for load_previous_runs_dataframes (called by load repos group data)
+    extract_folderinfo["group_names"] = ["5 mm", "12 mm", "25 mm"]
+    extract_folderinfo["load_dir"] = "example data/group"
+    avg_dfs, g_avg_dfs, g_std_dfs, extract_cfg = load_repos_group_data(
+        extract_folderinfo, extract_cfg
+    )
+    # 2) for stats_df & the permutation test functions
+    extract_cfg["sampling_rate"] = 100
+    extract_cfg["do_permtest"] = True
+    extract_cfg["dont_show_plots"] = True
+    extract_cfg["tracking_software"] = "DLC"
+    extract_cfg["group_color_dict"] = {"5 mm": "red", "12 mm": "blue", "25 mm": "green"}
+    extract_folderinfo["contrasts"] = ["5 mm & 12 mm", "5 mm & 25 mm", "12 mm & 25 mm"]
+    stats_df = create_stats_df(avg_dfs, extract_folderinfo, extract_cfg)
+    plot_panel_instance = None
+    cluster_extent_test(
+        stats_df,
+        g_avg_dfs,
+        g_std_dfs,
+        "Ankle Angle",
+        extract_folderinfo,
+        extract_cfg,
+        plot_panel_instance,
+    )
+
+
+def test_Mixed_ANOVA(extract_cfg):
+    # Adopted example data from https://real-statistics.com/
+    # See the Mixed ANOVA xlsx file for complete link
+    extract_cfg["do_anova"] = True
+    extract_cfg["anova_design"] = "Mixed ANOVA"
+    stats_df = pd.read_excel("tests/test_data/group_data/Mixed ANOVA Example Data.xlsx")
+    stats_var = "Value"
+    result = run_ANOVA(stats_df, stats_var, extract_cfg)
+    # no GG p vals here
+    assert math.isclose(result["p-unc"][0], stats_df["p(A)"][0], abs_tol=1e-05)
+    assert math.isclose(result["p-unc"][1], stats_df["p(B)"][0], abs_tol=1e-05)
+    assert math.isclose(result["p-unc"][2], stats_df["p(AxB)"][0], abs_tol=1e-05)
+
+
+def test_multcomp_df_with_scipy_example(extract_folderinfo, extract_cfg):
+    # Adopted example from https://docs.scipy.org/doc/scipy-1.15.0/reference/generated/
+    # scipy.stats.tukey_hsd.html
+    extract_folderinfo["group_names"] = ["group0", "group1", "group2"]
+    extract_folderinfo["contrasts"] = [
+        "group0 & group1",
+        "group0 & group2",
+        "group1 & group2",
+    ]
+    # fmt: off
+    scipy_example = [24.5, 23.5, 26.4, 27.1, 29.9, 28.4, 34.2, 29.5, 32.2, 30.1, 26.1, 28.3, 24.3, 26.2, 27.8]  # fmt: on
+    # prep data
+    data = []
+    bin_num = 25
+    for i in scipy_example:
+        data.append([i] * bin_num)
+    data = np.ravel(data)
+    # prep group_col
+    subject_num = 5
+    group_col = []
+    for group in extract_folderinfo["group_names"]:
+        for _ in range(subject_num):
+            group_col.append([group] * bin_num)
+    group_col = np.ravel(group_col)
+    # prep ID_col
+    total_subject_num = 15
+    IDs = np.arange(1, total_subject_num + 1)
+    ID_col = np.repeat(IDs, bin_num)
+    # prep SC percentage col
+    extract_cfg["bin_num"] = 25
+    SC_percentages = bin_num_to_percentages(extract_cfg["bin_num"])
+    SC_percentage_col = [SC_percentages] * total_subject_num
+    SC_percentage_col = np.ravel(SC_percentage_col)
+    # pytest.set_trace()
+    # manually create dummy stats df
+    stats_df = pd.DataFrame(data=data, columns=["Variable"])
+    stats_df["Group"] = group_col
+    stats_df["ID"] = ID_col
+    stats_df["SC Percentage"] = SC_percentage_col
+    # run the func and perform hard-coded check with vals of doc's example
+    # => multcomp_df's rows are all equal because our SC percentages were all equal here
+    # => correct results correspond to doc-example's rows 1, 4 and 2 first listing all
+    #    stats-values, then p then CI low & CI high since that corresponds to the order
+    #    of multcomp df's columns
+    # => i.e., col1= SC %, then 3 cols for each contrast's q, then p, then CI low, then
+    #    CI high in the order of groups0 & 1, 0 & 2, and 1 & 2
+    multcomp_df = multcompare_SC_Percentages(
+        stats_df, "Variable", extract_folderinfo, extract_cfg
+    )
+    # fmt: off
+    correct_results = [-4.6, -0.260, 4.34, 0.014, 0.980, 0.02, -8.249, -3.909, 0.691, -0.951, 3.389, 7.989]  
+    for r, result in enumerate(correct_results):
+        assert math.isclose(result, multcomp_df.iloc[0, r+1], abs_tol=0.001)
+
+
+# %%..................................  PCA  ...........................................
+def test_run_PCA(extract_folderinfo, extract_cfg):
+    # Replicate the example found in https://www.kdnuggets.com/2023/05/
+    # principal-component-analysis-pca-scikitlearn.html using our PCA df and PCA_info
+    # structure
+    # fmt: off
+    true_results_PCA_eigenvectors = [[0.1443294, -0.24518758, -0.00205106, -0.23932041,  0.14199204,  0.39466085, 0.4229343, -0.2985331, 0.31342949, -0.0886167,0.29671456,  0.37616741, 0.28675223],
+    [-0.48365155, -0.22493093, -0.31606881,  0.0105905,  -0.299634,   -0.06503951
+    , 0.00335981, -0.02877949, -0.03930172, -0.52999567,  0.27923515,  0.16449619,
+    -0.36490283],
+    [-0.20738262,  0.08901289,  0.6262239,   0.61208035,  0.13075693,  0.14617896,
+    0.1506819,   0.17036816,  0.14945431, -0.13730621,  0.08522192,  0.16600459,
+    -0.12674592]]
+    true_results_PCA_explained_var = [0.36198848, 0.1920749 , 0.11123631]
+    # fmt: on
+    wine_data = datasets.load_wine(as_frame=True)
+    wine_df = wine_data.data
+    features = wine_df.columns
+    PCA_df, PCA_info = run_PCA(wine_df, features, extract_folderinfo, extract_cfg)
+    for i in range(3):
+        # absolute values are compared because the signs can be different w. eigenvecs
+        assert np.allclose(
+            np.absolute(PCA_info["eigenvectors"][i]),
+            np.absolute(true_results_PCA_eigenvectors[i]),
+            atol=1e-05,
+        )
+    assert np.allclose(
+        PCA_info["explained_vars"], true_results_PCA_explained_var, atol=1e-05
+    )
+
+
+cases = (
+    (
+        "1-30,50,70-100",
+        100,
+        np.arange(1, 31).tolist() + [50] + np.arange(70, 101).tolist(),
+    ),
+    ("1-70", 25, [4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, 48, 52, 56, 60, 64, 68]),
+    ("2, 10, 22, 44, 50, 79", 50, [2, 10, 22, 44, 50]),
+    ("1-50,70-80", 10, [10, 20, 30, 40, 50, 70, 80]),
+)  # fmt: skip
+@pytest.mark.parametrize("PCA_bins, bin_num, expected_bins_list", cases)
+def test_convert_PCA_bins_to_list(
+    PCA_bins, bin_num, expected_bins_list, extract_folderinfo, extract_cfg
+):
+    extract_cfg["PCA_bins"] = PCA_bins
+    extract_cfg["bin_num"] = bin_num
+    updated_cfg = convert_PCA_bins_to_list(extract_folderinfo, extract_cfg)
+    assert np.array_equal(updated_cfg["PCA_bins"], np.array(expected_bins_list))
+
+
+def load_repos_group_data(extract_folderinfo, extract_cfg):
+    """Use load_previous_runs_dataframes to load example data from the repo"""
+    extract_cfg["sampling_rate"] = 100
+    extract_cfg["group_color_dict"] = {"5 mm": "red", "12 mm": "blue", "25 mm": "green"}
+    extract_folderinfo["contrasts"] = ["5 mm & 12 mm", "5 mm & 25 mm", "12 mm & 25 mm"]
+    avg_dfs, g_avg_dfs, g_std_dfs, extract_cfg = load_previous_runs_dataframes(
+        extract_folderinfo, extract_cfg
+    )
+    return avg_dfs, g_avg_dfs, g_std_dfs, extract_cfg
+
+
+# NOTE !!!
+# => This test is outdated and wrong because it is for two within subject factors (for
+#    example 3 separate behavioural tests and pre/post medication treatment or so -
+#    done by each subject)
+# => I will not delete this since I might want to re-use it if I should support such
+#    designs
+#
+# def test_twoway_RM_ANOVA(extract_cfg):
+#     # Adopted example data from https://real-statistics.com/
+#     # See the RM ANOVA xlsx file for complete link
+#     extract_cfg["do_anova"] = True
+#     extract_cfg["anova_design"] = "RM ANOVA"
+#     stats_df = pd.read_excel("tests/test_data/group_data/RM ANOVA Example Data.xlsx")
+#     stats_var = "Value"
+#     result = run_ANOVA(stats_df, stats_var, extract_cfg)
+#     # Note that the last 2 assert statements have a different tolerance because those
+#     # p-values differed a bit. Not sure why but it's a tiny amount and tolerable IMO
+#     pytest.set_trace()
+#     assert math.isclose(result["p-unc"][0], stats_df["p(A)"][0], abs_tol=1e-05)
+#     assert math.isclose(result["p-unc"][1], stats_df["p(B)"][0], abs_tol=1e-05)
+#     assert math.isclose(result["p-unc"][2], stats_df["p(AxB)"][0], abs_tol=1e-05)
+#     assert math.isclose(result["p-GG-corr"][0], stats_df["GG-p(A)"][0], abs_tol=1e-05)
+#     assert math.isclose(result["p-GG-corr"][1], stats_df["GG-p(B)"][0], abs_tol=1e-04)
+#     assert math.isclose(result["p-GG-corr"][2], stats_df["GG-p(AxB)"][0], abs_tol=1e-02)

tests/test_universal3D_approval.py

@@ -0,0 +1,89 @@
+from autogaita.resources.utils import try_to_run_gaita
+import pandas as pd
+import pandas.testing as pdt
+import os
+import shutil
+import pdb
+import pytest
+
+
+# ............................  SIMI APPROVAL TESTS STRUCTURE  .........................
+# 1. Run autogaita.simi for Subject (with the cfg used there)
+# 2. Load the "Average Stepcycles".xlsx file from the repo and compare for
+#    equivalence to average_data
+# 3. Do the same for "Standard Devs. Stepcycle.xlsx" and std_data
+# 4. Pass the test if the two df-pairs are equal
+
+
+# ...............................  PREPARE - FOUR FIXTURES   ...........................
+
+
+@pytest.fixture
+def extract_true_dir():
+    return "tests/test_data/universal3D_data/true_data/"
+
+
+@pytest.fixture
+def extract_info(tmp_path):
+    info = {}
+    info["name"] = "TestSubject"
+    info["results_dir"] = os.path.join(tmp_path, info["name"])
+    return info
+
+
+@pytest.fixture
+def extract_folderinfo():
+    folderinfo = {}
+    folderinfo["root_dir"] = "tests/test_data/universal3D_data/test_data/"
+    folderinfo["sctable_filename"] = "SC Latency Table"
+    folderinfo["postname_string"] = ""
+    return folderinfo
+
+
+@pytest.fixture
+def extract_cfg():
+    cfg = {}
+    cfg["sampling_rate"] = 100
+    cfg["dont_show_plots"] = True
+    cfg["y_acceleration"] = True
+    cfg["angular_acceleration"] = True
+    cfg["bin_num"] = 25
+    cfg["plot_SE"] = True
+    cfg["standardise_z_at_SC_level"] = True
+    cfg["standardise_z_to_a_joint"] = False
+    cfg["z_standardisation_joint"] = ["Midfoot, left"]
+    cfg["plot_joint_number"] = 7
+    cfg["legend_outside"] = True
+    cfg["flip_gait_direction"] = True
+    cfg["color_palette"] = "viridis"
+    cfg["analyse_average_y"] = False
+    cfg["standardise_y_coordinates"] = True
+    cfg["y_standardisation_joint"] = ["Midfoot, left"]
+    cfg["coordinate_standardisation_xls"] = ""
+    cfg["joints"] = ["Midfoot", "Ankle", "Knee", "Hip", "Pelvis "]
+    cfg["angles"] = {
+        "name": ["Ankle", "Knee", "Hip"],
+        "lower_joint": ["Midfoot", "Ankle", "Knee"],
+        "upper_joint": ["Knee", "Hip", "Pelvis "],
+    }
+    return cfg
+
+
+# ...............................  RUN - ONE APPROVAL TEST  ............................
+
+
+@pytest.mark.slow  # https://docs.pytest.org/en/7.1.x/example/markers.html
+def test_universal3D_approval(
+    extract_true_dir, extract_info, extract_folderinfo, extract_cfg
+):
+    # run
+    try_to_run_gaita(
+        "Universal 3D", extract_info, extract_folderinfo, extract_cfg, False
+    )
+    for true_df_file in os.listdir(extract_true_dir):
+        if true_df_file.endswith(".xlsx"):
+            true_df = pd.read_excel(os.path.join(extract_true_dir, true_df_file))
+            test_df = pd.read_excel(
+                os.path.join(extract_info["results_dir"], true_df_file)
+            )
+            pdt.assert_frame_equal(test_df, true_df)

tests/test_universal3D_unit_1_preparation.py

@@ -0,0 +1,69 @@
+from autogaita.universal3D.universal3D_datafile_preparation import (
+    clean_a_file,
+    rename_a_column,
+)
+from hypothesis import given, strategies as st
+import pandas as pd
+
+
+# %%..............................  preparation  .......................................
+
+
+# property test for removing a string from column names via clean a file
+column_names = st.lists(st.text(min_size=1), min_size=1, max_size=10, unique=True)
+strings_to_remove = st.text(min_size=1)
+
+
+@given(df_columns=column_names, string_to_remove=strings_to_remove)
+def test_clean_a_file(df_columns, string_to_remove):
+    df = pd.DataFrame(columns=df_columns)
+    cleaned_df = clean_a_file(df, string_to_remove)
+    for col in df_columns:
+        if string_to_remove in col:
+            assert col.replace(string_to_remove, "") in cleaned_df.columns
+        else:
+            assert col in cleaned_df.columns
+
+
+# Property test for renaming a given column according to 3 possible cases
+candidate_side_ids = st.sampled_from(
+    ["l", "L", "left", "LEFT", "Left", "r", "R", "right", "RIGHT", "Right"]
+)
+candidate_coord_ids = st.sampled_from(["x", "y", "z", "X", "Y", "Z"])
+candidate_joint_ids = st.text(min_size=2)
+separators = st.sampled_from(["_", "-", ":", "."])
+
+
+@given(
+    candidate_side_id=candidate_side_ids,
+    candidate_coord_id=candidate_coord_ids,
+    candidate_joint_id=candidate_joint_ids,
+    separator=separators,
+)
+def test_rename_a_column(
+    candidate_side_id, candidate_coord_id, candidate_joint_id, separator
+):
+    # only run main test below if separator is not the joint id anywhere
+    # => we require separators to be unique in column strings
+    if separator in candidate_joint_id:
+        return
+    # preparation - what output we are expecting
+    expected_side = "left" if candidate_side_id.lower() in ["l", "left"] else "right"
+    expected_coord = candidate_coord_id.capitalize()
+    # test 1 - column is side-specific
+    col_string = separator.join(
+        [candidate_side_id, candidate_coord_id, candidate_joint_id]
+    )
+    result = rename_a_column(col_string, separator)
+    assert isinstance(result, str)
+    assert (
+        result == f"{candidate_joint_id}, {expected_side} {expected_coord.capitalize()}"
+    )
+    # test 2 - column is central
+    col_string = separator.join([candidate_coord_id, candidate_joint_id])
+    result = rename_a_column(col_string, separator)
+    assert result == f"{candidate_joint_id} {expected_coord.capitalize()}"
+    # test 3 - column is not a coordinate column
+    col_string = separator.join([candidate_joint_id])
+    result = rename_a_column(col_string, separator)
+    assert result == col_string

tests/test_universal3D_unit_2_sc_extraction.py

@@ -0,0 +1,108 @@
+from autogaita.universal3D.universal3D_1_preparation import some_prep
+from autogaita.universal3D.universal3D_2_sc_extraction import (
+    check_different_angle_joint_coords,
+)
+import os
+import numpy as np
+import pytest
+
+
+# %%................................  fixtures  ........................................
+@pytest.fixture
+def extract_info(tmp_path):
+    info = {}
+    info["name"] = "TestSubject"
+    info["results_dir"] = os.path.join(tmp_path, info["name"])
+    return info
+
+
+@pytest.fixture
+def extract_folderinfo():
+    folderinfo = {}
+    folderinfo["root_dir"] = "tests/test_data/universal3D_data/test_data"
+    folderinfo["sctable_filename"] = "SC Latency Table.xlsx"
+    folderinfo["postname_string"] = ""
+    return folderinfo
+
+
+@pytest.fixture
+def extract_cfg():
+    # note space in end of "Midfoot, left " must be here bc. we don't run our fix and
+    # check cfg function of 1_prep_
+    cfg = {}
+    cfg["sampling_rate"] = 100  # base cfg
+    cfg["dont_show_plots"] = True
+    cfg["y_acceleration"] = True
+    cfg["angular_acceleration"] = True
+    cfg["bin_num"] = 25
+    cfg["plot_SE"] = True
+    cfg["standardise_z_at_SC_level"] = True
+    cfg["standardise_z_to_a_joint"] = True
+    cfg["z_standardisation_joint"] = ["Midfoot, left "]
+    cfg["plot_joint_number"] = 5
+    cfg["color_palette"] = "Set2"
+    cfg["legend_outside"] = True
+    cfg["flip_gait_direction"] = True
+    cfg["analyse_average_y"] = True
+    cfg["standardise_y_coordinates"] = True
+    cfg["y_standardisation_joint"] = ["Midfoot, left "]
+    cfg["coordinate_standardisation_xls"] = ""
+    cfg["joints"] = [
+        "Midfoot",
+        "Ankle",
+        "Knee",
+        "Hip",
+    ]
+    cfg["angles"] = {
+        "name": ["Ankle", "Knee"],
+        "lower_joint": ["Midfoot", "Ankle"],
+        "upper_joint": ["Knee", "Hip"],
+    }
+    cfg["direction_joint"] = "Midfoot, left Y"
+    return cfg
+
+
+# ..................................  tests  .........................................
+
+
+def test_clean_cycles_different_angle_joint_coords(
+    extract_info, extract_folderinfo, extract_cfg
+):
+    # start conditions
+    data, _ = some_prep(extract_info, extract_folderinfo, extract_cfg)
+    all_cycles = [[[111, 222], [333, 444]], [[555, 666], [777, 888]]]
+    # test one: leads to both lists still being non-empty, one SC removed
+    data.loc[123, "Ankle, left Y"] = 5.5
+    data.loc[123, "Ankle, left Z"] = 2.1
+    data.loc[123, "Midfoot, left Y"] = 5.5
+    data.loc[123, "Midfoot, left Z"] = 2.1
+    all_cycles = check_different_angle_joint_coords(
+        all_cycles, data, extract_info, extract_cfg
+    )
+    assert all_cycles == [[[333, 444]], [[555, 666], [777, 888]]]
+    with open(os.path.join(extract_info["results_dir"], "Issues.txt")) as f:
+        content = f.read()
+    assert "Run #1 - SC #1" in content
+    assert "equal LEFT" in content
+    assert "Lower joint: Midfoot" in content
+    # test two: leads to one empty list
+    data.loc[345, "Hip, right Y"] = 10.1
+    data.loc[345, "Hip, right Z"] = 0.333
+    data.loc[345, "Ankle, right Y"] = 10.1
+    data.loc[345, "Ankle, right Z"] = 0.333
+    all_cycles = check_different_angle_joint_coords(
+        all_cycles, data, extract_info, extract_cfg
+    )
+    assert all_cycles == [[], [[555, 666], [777, 888]]]
+    with open(os.path.join(extract_info["results_dir"], "Issues.txt")) as f:
+        content = f.read()
+    assert "RIGHT" in content
+    # test three: leads to all_cycles being None
+    data.loc[[567, 789], "Hip, right Y"] = 22.2
+    data.loc[[567, 789], "Hip, right Z"] = 0.11
+    data.loc[[567, 789], "Knee, right Y"] = 22.2
+    data.loc[[567, 789], "Knee, right Z"] = 0.11
+    assert (
+        check_different_angle_joint_coords(all_cycles, data, extract_info, extract_cfg)
+        is None
+    )