autogaita 1.0.4__tar.gz → 1.0.5__tar.gz

{autogaita-1.0.4 → autogaita-1.0.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogaita
-Version: 1.0.4
+Version: 1.0.5
 Summary: Automatic Gait Analysis in Python. A toolbox to streamline and standardise the analysis of kinematics across species after ML-based body posture tracking. Despite being optimised for gait analyses, AutoGaitA has the potential to be used for any kind of kinematic analysis.
 Home-page: https://github.com/mahan-hosseini/AutoGaitA/
 Author: Mahan Hosseini

autogaita-1.0.5/autogaita/batchrun_scripts/group_dlcrun.py

@@ -0,0 +1,112 @@
+import autogaita
+
+
+def group_dlcrun():
+    """
+    Batchrun script to run AutoGaitA Group for Results obtained with AutoGaitA DLC.
+    folderinfo & cfg dictionaries must be configured as explained in our documentation. See the "AutoGaitA without the GUI" section of our documentation for references to in-depth explanations to all dictionary keys (note that each key of dicts corresponds to some object in the AutoGaitA Group GUI)
+    """
+    # folderinfo
+    # => Note that length of folderinfo's group_names & group_dirs lists determines how #    many groups are compared.
+    # => Also note that indices must correspond (i.e., idx #    1's name will be used #    for dataset stored in group_dir's idx 1)
+    folderinfo = {}
+    folderinfo["group_names"] = [
+        # RM Dataset
+        "5mm",
+        "12mm",
+        "25mm",
+        # Interaction Issue Dataset
+        # "Control",
+        # "Silenced",
+        # Testing Dataset
+        # "one",
+        # "two",
+    ]
+    folderinfo["group_dirs"] = [
+        # --------------------------
+        # Full Dataset (39 subjects)
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Sanity Test Interaction Graziana/Result 10 joints/Control Full",
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Sanity Test Interaction Graziana/Result 10 joints/Silenced Full/",
+        # --------------------------
+        # Reduced Dataset (5 subjects)
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Sanity Test Interaction Graziana/Result 10 joints/Control",
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Sanity Test Interaction Graziana/Result 10 joints/Silenced/",
+        # --------------------------
+        # Testing Dataset
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Testing/Group 1/",
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Testing/Group 2/",
+        # --------------------------
+        # RM Dataset (example data in repo)
+        "/Users/mahan/sciebo/PythonCode/gaita_repo/example data/5mm/Results/",
+        "/Users/mahan/sciebo/PythonCode/gaita_repo/example data/12mm/Results/",
+        "/Users/mahan/sciebo/PythonCode/gaita_repo/example data/25mm/Results/",
+    ]
+    folderinfo["results_dir"] = (
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/example_data_results5/"
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Sanity Test Interaction Graziana/Result 10 joints/Mahan Results/"
+        # "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Testing/GroupResults/"
+    )
+    folderinfo["load_dir"] = (
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/example_data_results/"
+        # ""
+    )
+    # cfg
+    cfg = {}
+    cfg["do_permtest"] = True
+    cfg["do_anova"] = True
+    cfg["permutation_number"] = 100
+    cfg["PCA_n_components"] = 3
+    # cfg["PCA_n_components"] = 0.33
+    # cfg["PCA_custom_scatter_PCs"] = "4,5,6;4,5;2,4,6"
+    cfg["PCA_custom_scatter_PCs"] = ""
+    cfg["PCA_save_3D_video"] = False  # True
+    cfg["PCA_bins"] = ""  # "0-10,24,50-75"
+    cfg["stats_threshold"] = 0.05
+    cfg["plot_SE"] = False
+    cfg["color_palette"] = "viridis"
+    cfg["dont_show_plots"] = True
+    cfg["legend_outside"] = True
+    cfg["which_leg"] = "left"
+    cfg["anova_design"] = "RM ANOVA"
+    # cfg["anova_design"] = "Mixed ANOVA"
+    cfg["permutation_number"] = 100
+    cfg["PCA_variables"] = [
+        "Hind paw tao y",
+        "Ankle y",
+        "Knee y",
+        "Ankle Angle",
+        "Knee Angle",
+        "Nose x",
+        "Knee y",
+        "Knee x",
+        "Knee Velocity",
+        "Knee Acceleration",
+        "Knee Angle",
+        "Knee Angle Velocity",
+        "Knee Angle Acceleration",
+        "Hip Angle",
+    ]
+    cfg["stats_variables"] = [
+        "Hind paw tao y",
+        # "Ankle y",
+        # "Knee y",
+        # "Ankle Angle",
+        # "Knee Angle",
+        # "Nose x",
+        # "Knee y",
+        # "Knee x",
+        # "Knee Velocity",
+        # "Knee Acceleration",
+        # "Knee Angle",
+        # "Knee Angle Velocity",
+        # "Knee Angle Acceleration",
+        # "Hip Angle",
+        # "Elbow Angle"
+    ]
+    # run
+    autogaita.group(folderinfo, cfg)
+
+
+# %% what happens if we just hit run
+if __name__ == "__main__":
+    group_dlcrun()

autogaita-1.0.5/autogaita/batchrun_scripts/group_dlcrun_forpaper.py

@@ -0,0 +1,151 @@
+import autogaita
+
+
+def group_dlcrun():
+    """
+    Batchrun script to run AutoGaitA Group for Results obtained with AutoGaitA DLC.
+    folderinfo & cfg dictionaries must be configured as explained in our documentation. See the "AutoGaitA without the GUI" section of our documentation for references to in-depth explanations to all dictionary keys (note that each key of dicts corresponds to some object in the AutoGaitA Group GUI)
+    """
+
+    # ----------------------------- 3 Months 3 Beams --------------------------------
+    # folderinfo = {}
+    # folderinfo["group_names"] = [
+    #     # RM Dataset
+    #     "3m_5mm",
+    #     "3m_12mm",
+    #     "3m_25mm",
+    # ]
+    # folderinfo["group_dirs"] = [
+    #     "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/5mm/9w/",
+    #     "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/12mm/9w/",
+    #     "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/25mm/9w/",
+    # ]
+    # folderinfo["results_dir"] = (
+    #     "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/3Months3BeamsForPaper"
+
+    # ----------------------------- Old Young 3 Beams --------------------------------
+    folderinfo = {}
+    folderinfo["group_names"] = [
+        # RM Dataset
+        "3m_5mm",
+        "3m_12mm",
+        "3m_25mm",
+        "24m_5mm",
+        "24m_12mm",
+        "24m_25mm",
+    ]
+    folderinfo["group_dirs"] = [
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/5mm/9w/",
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/12mm/9w/",
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/25mm/9w/",
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/5mm/104w/",
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/12mm/104w/",
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/255mm/104w/",
+    ]
+    folderinfo["results_dir"] = (
+        "/Users/mahan/sciebo/Research/AutoGaitA/Mouse/Full Dataset and Results Nov 2024/YoungOld3BeamsForPaper_v2"
+    )
+    folderinfo["load_dir"] = ""
+
+    # cfg
+    cfg = {}
+    cfg["do_permtest"] = False
+    cfg["do_anova"] = True
+    cfg["permutation_number"] = 10000
+    cfg["PCA_n_components"] = 3
+    # cfg["PCA_n_components"] = 0.33
+    # cfg["PCA_custom_scatter_PCs"] = "4,5,6;4,5;2,4,6"
+    cfg["PCA_custom_scatter_PCs"] = ""
+    cfg["PCA_save_3D_video"] = False  # True
+    cfg["PCA_bins"] = ""  # "0-10,24,50-75"
+    cfg["stats_threshold"] = 0.05
+    cfg["plot_SE"] = True
+    cfg["color_palette"] = "Set2"
+    cfg["dont_show_plots"] = True
+    cfg["legend_outside"] = True
+    cfg["which_leg"] = "left"
+    cfg["anova_design"] = "RM ANOVA"
+    cfg["PCA_variables"] = [
+        "Nose x",
+        "Nose y",
+        "Ear base x",
+        "Ear base y",
+        "Front paw tao x",
+        "Front paw tao y",
+        "Wrist x",
+        "Wrist y",
+        "Elbow x",
+        "Elbow y",
+        "Lower Shoulder x",
+        "Lower Shoulder y",
+        "Upper Shoulder x",
+        "Upper Shoulder y",
+        "Iliac Crest x",
+        "Iliac Crest y",
+        "Hip x",
+        "Hip y",
+        "Knee x",
+        "Knee y",
+        "Ankle x",
+        "Ankle y",
+        "Hind paw tao x",
+        "Hind paw tao y",
+        "Tail base y",
+        "Tail center x",
+        "Tail center y",
+        "Tail tip x",
+        "Tail tip y",
+        "Hind paw tao Velocity",
+        "Hind paw tao Acceleration",
+        "Ankle Velocity",
+        "Ankle Acceleration",
+        "Knee Velocity",
+        "Knee Acceleration",
+        "Hip Velocity",
+        "Hip Acceleration",
+        "Iliac Crest Velocity",
+        "Iliac Crest Acceleration",
+        "Ankle Angle",
+        "Knee Angle",
+        "Hip Angle",
+        "Wrist Angle",
+        "Elbow Angle",
+        "Lower Shoulder Angle",
+        "Iliac Crest Angle",
+        "Ankle Angle Velocity",
+        "Ankle Angle Acceleration",
+        "Knee Angle Acceleration",
+        "Hip Angle Velocity",
+        "Hip Angle Acceleration",
+        "Wrist Angle Velocity",
+        "Wrist Angle Acceleration",
+        "Elbow Angle Velocity",
+        "Elbow Angle Acceleration",
+        "Lower Shoulder Angle Velocity",
+        "Lower Shoulder Angle Acceleration",
+        "Iliac Crest Angle Velocity",
+        "Iliac Crest Angle Acceleration",
+    ]
+    cfg["stats_variables"] = [
+        "Hip x",
+        "Hip y",
+        "Knee x",
+        "Knee y",
+        "Ankle x",
+        "Ankle y",
+        "Hind paw tao x",
+        "Hind paw tao y",
+        "Hind paw tao Velocity",
+        "Ankle Velocity",
+        "Knee Velocity",
+        "Ankle Angle",
+        "Knee Angle",
+        "Hip Angle",
+    ]
+    # run
+    autogaita.group(folderinfo, cfg)
+
+
+# %% what happens if we just hit run
+if __name__ == "__main__":
+    group_dlcrun()

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/batchrun_scripts/group_universal3Drun.py

@@ -23,6 +23,7 @@ def group_universal3Drun():
             + cfg["which_leg"]
             + " leg/"
         )
+        folderinfo["load_dir"] = ""
         # cfg
         cfg["do_permtest"] = True
         cfg["do_anova"] = True

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/group/group_1_preparation.py

@@ -49,43 +49,59 @@ def some_prep(folderinfo, cfg):
         if os.path.exists(info_file_path):
             os.remove(info_file_path)
 
-    # extracted_cfg_vars: save_to_xls, PCA stuff & dont show plots
-    cfg = extract_cfg_vars(folderinfo, cfg)
-
-    # see if there's a config json file and add to cfg dict
-    for g_idx, group_dir in enumerate(group_dirs):
+    # *********** IMPORTANT ***********
+    # if load_dir, we have already saved a group config.json (see below before return)
+    # => use this and just return the cfg
+    # => make sure to write folderinfo["contrast"] manually above as is and then return
+    #    that folderinfo plus the cfg you load from the file
+    if folderinfo["load_dir"]:
         with open(
-            os.path.join(group_dir, CONFIG_JSON_FILENAME), "r"
+            os.path.join(folderinfo["load_dir"], CONFIG_JSON_FILENAME), "r"
         ) as config_json_file:
-            config_vars_from_json = json.load(config_json_file)
-            for key in config_vars_from_json.keys():
-                # assigning like this ensure all keys are in all jsons across groups
-                if g_idx == 0:
-                    cfg[key] = config_vars_from_json[key]
-                else:
-                    # sanity check for group-differences in cfg variables!
-                    if (key not in cfg.keys()) | (
-                        cfg[key] != config_vars_from_json[key]
-                    ):
-                        error_message = (
-                            "config.json variables differ between groups!"
-                            + "\nPlease make sure that all cfg variables between "
-                            + "groups match & try again!"
-                        )
-                        raise ValueError(error_message)
-                    else:
+            cfg = json.load(config_json_file)
+            cfg["loaded"] = True  # used in a unit test in test_group_unit.py
+
+    else:  # if not, do the below things based on group dirs' configs
+        # 1. extracted_cfg_vars: save_to_xls, PCA stuff & dont show plots
+        cfg = extract_cfg_vars(folderinfo, cfg)
+
+        # 2. ennsure each key's across-group-equivalence and then add to cfg dict
+        for g_idx, group_dir in enumerate(group_dirs):
+            with open(
+                os.path.join(group_dir, CONFIG_JSON_FILENAME), "r"
+            ) as config_json_file:
+                config_vars_from_json = json.load(config_json_file)
+                for key in config_vars_from_json.keys():
+                    # assigning like this ensure all keys are in all jsons across groups
+                    if g_idx == 0:
                         cfg[key] = config_vars_from_json[key]
+                    else:
+                        # sanity check for group-differences in cfg variables!
+                        if (key not in cfg.keys()) | (
+                            cfg[key] != config_vars_from_json[key]
+                        ):
+                            error_message = (
+                                "config.json variables differ between groups!"
+                                + "\nPlease make sure that all cfg variables between "
+                                + "groups match & try again!"
+                            )
+                            raise ValueError(error_message)
+                        else:
+                            cfg[key] = config_vars_from_json[key]
 
-    # after having checked cfg keys for equivalence, we have to make sure that
-    # hind_joints is renamed to joints if DLC
-    # => note that cfg will not be updated (and overwritten from input if Universal 3D so it's)
-    #    okay that we do this
-    if cfg["tracking_software"] == "DLC":
-        cfg["joints"] = cfg["hind_joints"]
-    joints = cfg["joints"]
-    angles = cfg["angles"]
+        # 3. rename hind_joints is to joints if DLC or SLEAP
+        if "hind_joints" in cfg.keys():
+            cfg["joints"] = cfg["hind_joints"]
 
-    # prepare some plotting color stuff
+    # ******** IMPORTANT *******
+    # => Do the following two things regardless of load_dir:
+    # 1. save cfg to json file in results_dir for load_dir capability
+    config_json_path = os.path.join(results_dir, CONFIG_JSON_FILENAME)
+    if os.path.exists(config_json_path):  # overwrite if exists
+        os.remove(config_json_path)
+    with open(config_json_path, "w") as config_json_file:
+        json.dump(cfg, config_json_file)
+    # 2. create this plot stuff manually (cycler objects cannot be written to json)
     cfg["group_color_cycler"] = plt.cycler(
         "color", sns.color_palette(cfg["color_palette"], len(group_names))
     )
@@ -93,10 +109,10 @@ def some_prep(folderinfo, cfg):
         zip(group_names, cfg["group_color_cycler"].by_key()["color"])
     )
     cfg["joint_color_cycler"] = plt.cycler(
-        "color", sns.color_palette(cfg["color_palette"], len(joints))
+        "color", sns.color_palette(cfg["color_palette"], len(cfg["joints"]))
     )
     cfg["angle_color_cycler"] = plt.cycler(
-        "color", sns.color_palette(cfg["color_palette"], len(angles["name"]))
+        "color", sns.color_palette(cfg["color_palette"], len(cfg["angles"]["name"]))
     )
 
     return folderinfo, cfg

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/group/group_2_data_processing.py

@@ -190,7 +190,7 @@ def bin_num_error_helper_function(folderinfo, dfs, g, group_names, sc_breaks, b)
     raise ValueError(message)
 
 
-def test_PCA_and_stats_variables(df, group_name, name, folderinfo, cfg):
+def check_PCA_and_stats_variables(df, group_name, name, folderinfo, cfg):
     """Tests if PCA & stats variables are present in all groups' datasets"""
     PCA_variables = cfg["PCA_variables"]
     stats_variables = cfg["stats_variables"]
@@ -275,7 +275,7 @@ def import_and_combine_dfs(
     else:
         df_copy = df.copy()
         # test: are our PCA & stats variables present in this ID's dataset?
-        test_PCA_and_stats_variables(df_copy, group_name, name, folderinfo, cfg)
+        check_PCA_and_stats_variables(df_copy, group_name, name, folderinfo, cfg)
         # add some final info & append to group_df
         if tracking_software == "DLC":
             # add this run's info to df (& prepare Stepcycle column)
@@ -542,7 +542,6 @@ def idxs_list_to_list_of_SC_lists(idxs):
 
 def grand_avg_and_std(avg_dfs, folderinfo, cfg):
     """Compute the grand averages & standard deviations for all columns of df"""
-
     # unpack
     group_names = folderinfo["group_names"]
     results_dir = folderinfo["results_dir"]
@@ -592,3 +591,73 @@ def grand_avg_and_std(avg_dfs, folderinfo, cfg):
         )
         save_results_sheet(g_std_dfs[g], save_to_xls, g_std_filepath)
     return g_avg_dfs, g_std_dfs
+
+
+# %% .................  local functions #4 - load previous dataframes  .................
+def load_previous_runs_dataframes(folderinfo, cfg):
+    """If user asked for it load previous runs dataframes instead of generating them (i.e., avg_dfs, g_avg_dfs, g_std_dfs)"""
+    avg_dfs = [[]] * len(folderinfo["group_names"])
+    g_avg_dfs = [[]] * len(folderinfo["group_names"])
+    g_std_dfs = [[]] * len(folderinfo["group_names"])
+    for g, group_name in enumerate((folderinfo["group_names"])):
+        try:
+            avg_dfs[g] = pd.read_excel(
+                os.path.join(
+                    folderinfo["load_dir"],
+                    group_name + " - " + AVG_GROUP_SHEET_NAME + ".xlsx",
+                )
+            )
+            g_avg_dfs[g] = pd.read_excel(
+                os.path.join(
+                    folderinfo["load_dir"],
+                    group_name + " - " + G_AVG_GROUP_SHEET_NAME + ".xlsx",
+                )
+            )
+            g_std_dfs[g] = pd.read_excel(
+                os.path.join(
+                    folderinfo["load_dir"],
+                    group_name + " - " + G_STD_GROUP_SHEET_NAME + ".xlsx",
+                )
+            )
+        except FileNotFoundError:
+            error_msg = (
+                "\n******************\n! CRITICAL ERROR !\n******************\n"
+                + f"Unable to load the data of group '{group_name}' from \n"
+                + f"{folderinfo['load_dir']}"
+                + "\n\nTry again!"
+            )
+            print(error_msg)
+            write_issues_to_textfile(error_msg, folderinfo)
+            raise FileNotFoundError
+    # re-index avg_dfs based on unique SC Percentages
+    array_of_idxs = np.arange(len(avg_dfs[0][SC_PERCENTAGE_COL].unique()))
+    for g in range(len(avg_dfs)):
+        repeat_index_this_often = len(avg_dfs[g]) // len(array_of_idxs)
+        new_idx = np.tile(array_of_idxs, repeat_index_this_often)
+        avg_dfs[g] = avg_dfs[g].set_index(new_idx)
+    # check if PCA & stats variables are present in all dataframes
+    for g, group_name in enumerate(folderinfo["group_names"]):
+        check_PCA_and_stats_variables(
+            avg_dfs[g], group_name, "Average", folderinfo, cfg
+        )
+        check_PCA_and_stats_variables(
+            g_avg_dfs[g], group_name, "Grand Average", folderinfo, cfg
+        )
+        check_PCA_and_stats_variables(
+            g_std_dfs[g], group_name, "Grand Standard Deviation", folderinfo, cfg
+        )
+    # since import_data writes bin_num (running a sanity check before) we have to do
+    # it here too. No need to run the sanity check again since that was done previously
+    cfg["bin_num"] = len(np.unique(avg_dfs[0][SC_PERCENTAGE_COL]))
+    if cfg["bin_num"] * len(np.unique(avg_dfs[0][ID_COL])) != len(avg_dfs[0]):
+        error_msg = (
+            "\n******************\n! CRITICAL ERROR !\n******************\n"
+            + "Something is wrong with your IDs and SC Percentage columns."
+            + "\nThere seem to be duplicates. Make sure that this is not the case."
+            + "\nOtherwise, run it again without loading your previous run's results."
+        )
+        print(error_msg)
+        write_issues_to_textfile(error_msg, folderinfo)
+        raise ValueError(error_msg)
+    else:
+        return avg_dfs, g_avg_dfs, g_std_dfs, cfg

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/group/group_4_stats.py

@@ -446,7 +446,7 @@ def plot_permutation_test_results(
                 if tracking_software in ["DLC", "SLEAP"]:
                     unit = "x in pixels"
                 elif tracking_software == "Universal 3D":
-                    unit = "Y in (your units)"
+                    unit = "Y"
             f.supylabel(
                 ylabel_velocity_and_acceleration(feature, unit, sampling_rate),
                 fontsize=STATS_PLOTS_SUPLABEL_SIZE,
@@ -896,7 +896,7 @@ def plot_multcomp_results(
                 if tracking_software in ["DLC", "SLEAP"]:
                     unit = "x in pixels"
                 elif tracking_software == "Universal 3D":
-                    unit = "Y in (your units)"
+                    unit = "Y"
             f.supylabel(
                 ylabel_velocity_and_acceleration(feature, unit, sampling_rate),
                 fontsize=STATS_PLOTS_SUPLABEL_SIZE,

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/group/group_main.py

@@ -4,6 +4,7 @@ from autogaita.group.group_2_data_processing import (
     import_data,
     avg_and_std,
     grand_avg_and_std,
+    load_previous_runs_dataframes,
 )
 from autogaita.group.group_3_PCA import PCA_main
 from autogaita.group.group_4_stats import (
@@ -18,14 +19,6 @@ from autogaita.resources.utils import print_finish, PlotPanel
 import matplotlib
 import matplotlib.pyplot as plt
 
-# %% A note on cross species functionality
-# => This function supports cross species analyses, however the data must be obtained
-#    from the same tracking_software (=> I.e., resulting from autogaita_dlc or _universal3D)
-# => Also users have to ensure that the comparison makes sense across species w.r.t.
-#    features
-# => Adding "both" leg functionality in compute_avg & g_avg_dfs at some point.
-
-
 # .................................  constants  ........................................
 # SET PLT BACKEND
 # Agg is a non-interactive backend for plotting that can only write to files
@@ -62,23 +55,34 @@ def group(folderinfo, cfg):
         # build loading screen
         plot_panel_instance.build_plot_panel_loading_screen()
 
-    # ..................................  unpack  ......................................
+    # ................................  preparation  ...................................
+    # => either creates and sanity-checks folderinfo & cfg or loads it from a previous
+    #    run's config.json file (if load_dir)
     folderinfo, cfg = some_prep(folderinfo, cfg)
 
     # ..............................  print start  ....................................
     # => print start after some_prep since we do some stuff to cfg["PCA_bins"] there
     print_start(folderinfo, cfg)
 
-    # ...................................  import  .....................................
-    # => DLC ONLY: dfs is x-standardised automatically if 1st-level standardised x
-    #   -- As a result all average & std dfs are x-standardised as well
-    dfs, raw_dfs, cfg = import_data(folderinfo, cfg)
-
-    # .................................  avgs & stds  ..................................
-    avg_dfs, std_dfs = avg_and_std(dfs, folderinfo, cfg)
-
-    # ..............................  grand avgs & stds  ...............................
-    g_avg_dfs, g_std_dfs = grand_avg_and_std(avg_dfs, folderinfo, cfg)
+    # .....................  import & transform (a) or load (b)  .......................
+
+    # approach a - import & transform (i.e. no previous results to load from)
+    if not folderinfo["load_dir"]:
+        # in dlc/sleap, dfs is x-standardised automatically if 1st-level standardised x
+        # => As a result all average & std dfs are x-standardised as well
+        dfs, raw_dfs, cfg = import_data(folderinfo, cfg)
+        avg_dfs, std_dfs = avg_and_std(dfs, folderinfo, cfg)
+        g_avg_dfs, g_std_dfs = grand_avg_and_std(avg_dfs, folderinfo, cfg)
+
+    # approach b - load previous run's dfs (avg, g_avg & g_std_dfs)
+    else:
+        try:
+            # creating cfg["bin_num"] here too (since done in import_data in (a)!)
+            avg_dfs, g_avg_dfs, g_std_dfs, cfg = load_previous_runs_dataframes(
+                folderinfo, cfg
+            )
+        except (FileNotFoundError, ValueError):
+            return
 
     # ...................................  PCA  ........................................
     if cfg["PCA_variables"]:  # empty lists are falsey!

{autogaita-1.0.4 → autogaita-1.0.5}/autogaita/group/group_utils.py

@@ -30,6 +30,10 @@ def print_start(folderinfo, cfg):
     for group_name in folderinfo["group_names"]:
         start_string += "\n" + group_name
 
+    # load dir
+    if folderinfo["load_dir"]:
+        start_string += "\n\nLoad Directory\n--------------\n" + folderinfo["load_dir"]
+
     # pca
     PCA_string = "P R I N C I P A L | C O M P O N E N T | A N A L Y S I S"
     space_add = " " * ((INFO_TEXT_WIDTH - len(PCA_string)) // 2)