gaitalytics 0.2.2

Installation

Fast install with anaconda:

conda install gaitalytics -c DartLab-LLUI

Or with pip:

pip install gaitalytics

Configuration

Gaitalytics can be used with any marker set, which at least includes three or four hip markers (front left/right, back left/right or sacrum) and four foot markers (left heel/toe, right heel/toe).

Additionally markers can be defined on which standard time-series features such as min max mean etc. will be calculated.

All functionalities in the libraries only take points into account which are configured in as specific yaml file.

Minimal requirements would look like this:

analysis:
  markers:
    - "LHipAngles"
    - "LKneeAngles"
    - "LAnkleAngles"
    - "LPelvisAngles"
    - "LThoraxAngles"
mapping:
  markers:
    # Foot
    l_heel: "LHEE"
    r_heel: "RHEE"
    l_toe: "LTOE"
    r_toe: "RTOE"

    # Hip
    l_ant_hip: "LASI"
    r_ant_hip: "RASI"
    l_post_hip: "LPSI"
    r_post_hip: "RPSI"
    sacrum: "SACR"

Simple Pipeline

from gaitalytics import api
# Load configuration (yaml file from above)
config = api.load_config("./pig_config.yaml")

# Load trial from c3d file
trial = api.load_c3d_trial("./test_small.c3d", config)

# Detect events
events = api.detect_events(trial, config)
try:
    # check events
    api.check_events(events)

    # write events to c3d in the same file
    api.write_events_to_c3d("./test_small.c3d", events, './test.c3d')

    # add events to trial
    trial.events = events

    # segment trial to gait cycles. (Events are already existing in the c3d file)
    trial_segmented = api.segment_trial(trial)

    # calculate features
    features = api.calculate_features(trial_segmented, config)

    # normalise time
    trial_normalized = api.time_normalise_trial(trial_segmented)

    # save features
    features.to_netcdf("features.nc")

    # export segmented trial to netcdf
    api.export_trial(trial_segmented, "output_segments")
    api.export_trial(trial_normalized, "output_norm")

except ValueError as e:
    print(e)