Autogaita

- Fixed an issue with RM ANOVA's input parameters

This version only differs to v1.0.0 w.r.t. some info prints being standardised so I copied **v1.0.0's changelog** below:

- **Major code refactoring**
- **Added AutoGaitA SLEAP**
- **AutoGaitA Simi has been changed to AutoGaitA Universal 3D**
- Including a “Datafile Preparation” part that makes it universal
- **Major Improvements to PCA**
- Cycle bins can be chosen so PCA is only performed over those instead of over full cycle
- Number of PCs can be chosen via percentage explained variance
- Custom Scatterplots can be configured
- Top features going into each PC provided as Excel table
- ID x PC value Excel table provided for recognising IDs in scatterplots
- **Major improvements to Tukey’s & ANOVA**
- 1 way ANOVA if only 1 within factor
- Always running Tukeys irrespective of ANOVA results
- Always printing & storing ANOVA tables
- Saving Multiple Comparison Results as two versions, one sheet for each feature
- Ensuring that simi would work if IDs were just a number, e.g. just 1
- Fixed an issue in dlc gui where the “wrong trial number” error would break the GUI instead of showing a message-window to users
- Added an error message if dlc’s step cycle latency extraction should break because of wrong Annotation Table swing/stance column names
- Fixed not checking for user-missed underscore “_” in camera identifier in dlc gaita
- Fixed an issue in plot_velocities in group where some random group’s number instead of joint name was included in the title
- Fixed assuming metres in universal3D velocity & acceleration plots

- **Major code refactoring**
- **Added AutoGaitA SLEAP**
- **AutoGaitA Simi has been changed to AutoGaitA Universal 3D**
- Including a “Datafile Preparation” part that makes it universal
- **Major Improvements to PCA**
- Cycle bins can be chosen so PCA is only performed over those instead of over full cycle
- Number of PCs can be chosen via percentage explained variance
- Custom Scatterplots can be configured
- Top features going into each PC provided as Excel table
- ID x PC value Excel table provided for recognising IDs in scatterplots
- **Major improvements to Tukey’s & ANOVA**
- 1 way ANOVA if only 1 within factor
- Always running Tukeys irrespective of ANOVA results
- Always printing & storing ANOVA tables
- Saving Multiple Comparison Results as two versions, one sheet for each feature
- Ensuring that simi would work if IDs were just a number, e.g. just 1
- Fixed an issue in dlc gui where the “wrong trial number” error would break the GUI instead of showing a message-window to users
- Added an error message if dlc’s step cycle latency extraction should break because of wrong Annotation Table swing/stance column names
- Fixed not checking for user-missed underscore “_” in camera identifier in dlc gaita
- Fixed an issue in plot_velocities in group where some random group’s number instead of joint name was included in the title
- Fixed assuming metres in universal3D velocity & acceleration plots

- Fixed an **important bug** in dlc where Angle velocities & accelerations were computed twice which translated to group’s feature window
- Fixed issue in dlc where angular velocities/accel. wouldn’t have been computed if primary joints were empty
- Fixed an issue with group gaita GUI where results-dir-check depended on Average (and not Normalised) Stepcycle.xls (couldn’t work on example data in repo!)

- **Major rework of all three GUIs**
- **Note that for now updates to first-level analyses are only provided to the dlc and not the simi toolbox due to low demand.**
- Might change in the future. Results of autogaita_simi are still analysable with autogaita_group.
- Sanity checking user’s anova_design cfg & informing users about which IDs are included in RM ANOVAs.
- Ignoring warnings of pandas2.2.2. deprecated dtypes. Fixing once all tests are implemented to be sure fixes don’t break stuff.
- Users do not need to worry about the decimal precision in their annotation tables (Laura Shoenhals’ contribution)
- DLC: x coordinates can now be standardised to a standardisation (baseline) joint
- Automatically generates XLS files with and without standardisation (also if autogaita_group is run)
- Step-cycle-level plots are based on non-standardised x-coord.s and average-level plots are based on standardised x-coord.s
- DLC: y coordinates can now be standardised to a standardisation (baseline) joint
- Fixed that simi assumed units to be in metres in plots
- Fixed that simi broke if only one leg had no valid step cycles
- Fixed an issue that did not plot y-axis labels of stats-plots if features were velocities or accelerations
- Added cfg entries:
- dlc: standardise_x_coordinates, x_standardisation_joint, standardise_y_to_a_joint, y_standardisation_joint

- Closing all figures using plt.close("all")
- Critical bugfix of simi w.r.t. analyse_average_y implementation

This version's documentation [is archived here!](https://docs.google.com/document/d/11mJd7jUHk7joQ0BdZT98CJRrIANdyosMQMJGFtp6yR4/edit?usp=sharing)