src.data_reader package

src.data_reader.imu module

This module contains the IMU class.

class src.data_reader.imu.IMU(data)

Bases: object

An IMU object holds all the data recorded by an actual IMU and offers getter and conversion methods.

acc_to_g()

Conversion method for acceleration data from m/s**2 to g.

acc_to_meter_per_square_sec()

Conversion method for acceleration data from g to m/s**2.

accel(i=None)

Getter method for acceleration.

Depending on if i is specified, this returns either data from timestep i or the whole timeseries data.

Parameters

i (int) – Optional - index of the data

Returns

data at the specified timestep i or all data

Return type

float or np.array

accel_variance()

Calculate acceleration variance.

Returns

variance of acceleration

Return type

float

check_sampling()

Calculate mean and standard deviation of the sampling rate.

Returns

mean and std. deviation of sampling rate

Return type

tuple(float, float)

crop(min_time, max_time, inplace=False)

Crops data to timeframe between min_time and max_time. Attention: all other data is dropped.

Parameters

• min_time (float) – start time of cropping

• max_time (float) – end time of cropping

• inplace (bool) – flag if cropping should be performed inplace or on a copy

Returns

If not inplace, a cropped copy of the IMU object is returned.

Return type

None or IMU

gyro(i=None)

Getter method for gyroscope data.

Depending on if i is specified, this returns either data from timestep i or the whole timeseries data.

Parameters

i (int) – Optional - index of the data

Returns

data at the specified timestep i or all data

Return type

float or np.array

gyro_to_degree()

Conversion method for gyroscope data from rad to degree.

gyro_to_rad()

Conversion method for gyroscope data from degree to rad.

resample(f)

Resample a copy of IMU data to given frequency.

Parameters

f (float) – target frequency

Returns

resampled IMU copy

Return type

IMU

time(i=None)

Getter method for timestamps.

Depending on if i is specified, this returns either data from timestep i or the whole timeseries data.

Parameters

i (int) – Optional - index of the data

Returns

data at the specified timestep i or all data

Return type

float or np.array

time_shift(time_offset)

Shift time by a given offset.

Parameters

time_offset (float) – offset used to shift time

Returns

None

zero_base_time()

Let timestamps start at zero.

src.data_reader.opto_gait_xml_reader module

This module contains functionality that reads data from XML files as they can be exported from the OptoGait software. This is only used to initially convert the XML files to CSV files, since the Excal XML format is rather cumbersome.

src.data_reader.opto_gait_xml_reader.read_opto_gait_raw_xml_simple(input_file)

Read OptoGait raw data from the excel xml file. The main information is stored in the spreadsheet “simple”.

Parameters

input_file (str) – xml input file name.

Returns

DataFrame with information from the “simple” spreadsheet.

Return type

(DataFrame)

src.data_reader.zebris_json_reader module

This module contains the ZebrisJsonReader.

class src.data_reader.zebris_json_reader.ZebrisJsonReader(raw_data_file, step_data_file)

Bases: object

The ZebrisJsonReader can be used to read and load preprocessed Zebris files. Besides the functionality for extracting gait parameters from the data, this class provides functionality for creating images and videos from the data.

get_raw_data()

Get raw data. Lazy loading with caching.

Returns

raw data

Return type

dict

get_step_data()

Get step (aggregated) data. Lazy loading with caching.

Returns

step data

Return type

dict

read_treadmill_velocity()

Read treadmill velocity.

Returns

Treadmill velocity for each sample.

Return type

list[float]

read_zebris_raw_json_heel_positions(generate_images=False, generate_heel_image=False)

Zebris does not provide stride length out of the box. Therefore, the positions of heel strikes are measured in order to obtain stride length. This function can also be used to create a series of images that could be later merged into a video or used to inspect the raw data. Also an image of the performed heel detection can be exported.

Parameters

• generate_images (bool) – flag, if a series of images should be created

• generate_heel_image (bool) – flag, if an image of the heel detection should be created

Returns

Positions of heelstrike for the left and right foot.

Return type

(tuple[list[float], ..])

read_zebris_raw_json_ic_fo()

Read the initial contact and foot-off timestamps. Zebris provides this data as part of the aggregated data.

Returns

DataFrames of IC and FO timestamps for the left and right foot.

Return type

tuple[DataFrame, ..]

read_zebris_raw_json_initial_contact()

The aggregated data from Zebris does not contain the first step. However, the raw data does. Therefore, the initial contact timestamp needs to be extracted from the raw data and is not the start of the first “detected step”.

Returns

IC timestamp

Return type

float

read_zebris_raw_json_to_video(out_file)

Creates an mp4 video of the raw sensor data. The framerate will be the actual sensor reading rate (some players might have problems with unusual framerates). One pixel is one sensor cell.

Parameters

out_file (str) – mp4 file name

Returns

None