Read edf and detect sleep
=========================

This script demonstrates how to read an EDF file and predict sleep stages.

Author: Tim Hermans (tim-hermans@hotmail.com).

Link to script: `feature_extraction/read_edf_and_detect_sleep.py <https://gitlab.com/timhermans/nnsa_public/-/blob/main/examples/feature_extraction/read_edf_and_detect_sleep.py>`_



.. code-block:: python

    import nnsa


Settings. 



.. code-block:: python

    # Path to the EDF file.
    edf_filepath = r'C:\data_temp\test.edf'
    
    # Whether to use 4-class sleep staging (True) or 2-class (False).
    four_class = True
    
    # Path to the output file (Excel).
    output_filepath = r'C:\data_temp\test_sleep.xlsx'


Read EEG data from EDF file. 



.. code-block:: python

    with nnsa.EdfReader(edf_filepath) as r:
        # Read (see the EdfReader.read_eeg_dataset() function for explanation of the options).
        # Returns an EegDataset object, which has some useful methods for processing the EEG.
        eeg_ds = r.read_eeg_dataset(discontinuous_mode='fill')


Sleep staging (for neonates >= 36 weeks PMA). 



.. code-block:: python

    # Compute result.
    if four_class:
        result = eeg_ds.sleep_stages_robust()
    else:
        result = eeg_ds.sleep_stages_cnn(num_classes=2)
    # The sleep stages are contained in the attribute `df` as a pandas dataframe:
    df = result.df
    
    # Save the results to Excel (or you could also save to csv. by using df.to_csv()).
    df.to_excel(output_filepath, index=False)
    print(f'Saved results to {output_filepath}')