Examples

Below you will find some code examples for the various modules in the package. These will help you get started and understand what the main functions of each module does and how to use them.

Cometa

How to load the cometa package and start loading data:

from epilepsy_tools import cometa
import pandas as pd

# you can (and probably should) use a pathlib.Path object
data: pd.DataFrame = cometa.load_data("path/to/cometa_data.c3d")

# keep only the acceleration data (similar for emg data)
acceleration_data: pd.DataFrame = cometa.extract_acceleration_data(data)

How to get the recordings metadata with the cometa.RecordingInfo object:

# get information about a recording
recording_info = cometa.RecordingInfo.from_file("path/to/cometa_data.c3d")

# and then print out some information
print(recording_info.fs)  # original sampling frequency in Hz
print(recording_info.channels)  # list of channel labels
print(recording_info.start_time, recording_info.end_time)  # datetime objects

You can also easily create plots of the data, for quick visualization:

import matplotlib.pyplot as plt

data: pd.DataFrame = cometa.load_data("path/to/cometa_data.c3d")

# plot acceleration data from all channels (similar for emg data)
fig = cometa.plot_acceleration(data)
plt.show()  # or fig.savefig("figure_name.png") if you want to save

Hexoskin

How to load the hexoskin package and start loading data:

from epilepsy_tools import hexoskin
import pandas as pd

# you can (and probably should) use a pathlib.Path object
data: pd.DataFrame = hexoskin.load_data("path/to/hexoskin_data.edf")

The data in a DataFrame format will contain NaNs. To work around that you have two choices:

# Fill the DatFrame with values, removing the NaNs but creating repetition
data: pd.DataFrame = data.ffill()

# OR

# load the data initially as a dict of Series, which will all be of different length
data: dict[str, pd.Series] = hexoskin.load_data("path/to/hexoskin_data.edf", as_dataframe=False)

How to get the recordings metadata with the hexoskin.RecordingInfo object:

# get information about a recording
recording_info = hexoskin.RecordingInfo.from_file("path/to/hexoskin_data.edf")

# and then print out some information
print(recording_info.patient_name)
print(recording_info. start_time)
for signal in recording_info.signals:
    # print information on the SignalHeader objects
    print(signal.label, signal.sample_rate)

EpiDataVault

First, you will need to have the annotations file, and the patient logs accessible. One recommended way is to store the paths to the files in a configuration file, such as config.py:

# config.py

annotations: str = r"path/to/annotations.xlsx"
log_18: str = r"path/to/patient_log_18.xlsx
log_23: str = r"path/to/patient_log_23.xlsx
password: str = "supersecretpassword"

You can then load this file and the module in a script and load the information:

from epilepsy_tools.epidatavault import (
    load_patient_log,
    load_annotation_file,
    generate_patient_numbers_list,
    build_patient_datavault,
    build_seizure_datavault,
)
import pandas as pd

# config.py in the same directory
import config

annotations = load_annotation_file(config.annotations)
patient_numbers = generate_patient_numbers_list(pd.ExcelFile(config.annotations))
seizure_types = ["FBTCS", "GTCS"]

patient_datavault: pd.DataFrame = build_patient_datavault(
    annotations=annotations,
    patient_numbers=patient_numbers,
    seizure_types=seizure_types,
    log18=load_patient_log(config.log_18, "log18", config.password),
    log23=load_patient_log(config.log_23, "log23"),
    save_path="path/where/to/save/FBTCS_patient_datavault.parquet",
)

seizure_datavault: pd.DataFrame = build_seizure_datavault(
    annotations=annotations,
    patient_numbers=patient_numbers,
    seizure_types=seizure_types,
    save_path="path/where/to/save/FBTCS_seizure_datavault.parquet",
)