maui.utils.false_color_spectrogram_prepare_dataset

maui.utils.false_color_spectrogram_prepare_dataset(df, datetime_col, duration_col=None, file_path_col=None, output_dir=None, unit='scale_60', calculate_acoustic_indices=True, **kwargs)

Prepare a dataset for generating false-color spectrograms, segmenting audio files, and calculating acoustic indices.

Return type:

DataFrame

Parameters:

dfpandas.DataFrame

DataFrame containing the audio file paths and timestamps.

datetime_colstr

Column name for the start time of the audio files.

duration_colstr, optional

Column name for the duration of the audio files. If None, it will be calculated from file_path_col.

file_path_colstr, optional

Column name for the audio file paths. Required if duration_col is None.

output_dirstr, optional

Directory where segmented audio files and results will be stored.

unitstr, optional

Time unit for segmentation. Default is ‘scale_60’.

calculate_acoustic_indicesbool, optional

If True, acoustic indices will be calculated for the segmented files.

**kwargsdict

Additional parameters for calculating acoustic indices. The available kwargs are:

• acoustic_indices_methods: list of str

  A list of methods used for calculating acoustic indices.

• pre_calculation_method: callable

  A method to be applied before the calculation of acoustic indices.

• parallel: bool

  Whether to perform the calculation of acoustic indices in parallel.

• chunk_size: int, optional

  Size of the chunks of data to be processed in parallel. Default is 5.

• temp_dir: str, optional

  Path to a temporary directory for storing intermediate results.

Returns:

pandas.DataFrame

A DataFrame containing the segmented audio files and, optionally, the calculated acoustic indices.

Raises:

Exception

If both duration_col and file_path_col are None, or if there are overlaps in the audio files, or if time gaps are detected between segments.

Examples

>>> from maui import samples, utils
>>> df = samples.get_audio_sample(dataset="leec")
>>> df["dt"] = pd.to_datetime(df["timestamp_init"]).dt.date
>>> def pre_calculation_method(s, fs):   
>>>     Sxx_power, tn, fn, ext = maad.sound.spectrogram (s, fs) 
>>>     Sxx_noNoise= maad.sound.median_equalizer(Sxx_power, display=False, extent=ext) 
>>>     Sxx_dB_noNoise = maad.util.power2dB(Sxx_noNoise)
>>> 
>>>     Sxx, tn, fn, ext = maad.sound.spectrogram(s, fs, mode='amplitude')
>>>     
>>>     pre_calc_vars = {'Sxx': Sxx, 'tn':tn , 'fn':fn , 'ext':ext, 'Sxx_dB_noNoise':Sxx_dB_noNoise }
>>>     return pre_calc_vars
>>>         
>>> def get_aci(pre_calc_vars):
>>>     aci_xx, aci_per_bin, aci_sum  = maad.features.acoustic_complexity_index(pre_calc_vars['Sxx'])
>>>     indices = {'aci_xx': aci_xx, 'aci_per_bin':aci_per_bin , 'aci_sum':aci_sum}
>>>     return indices
>>> 
>>> def get_spectral_events(pre_calc_vars):
>>>     EVNspFract_per_bin, EVNspMean_per_bin, EVNspCount_per_bin, EVNsp = maad.features.spectral_events(
>>>                 pre_calc_vars['Sxx_dB_noNoise'],
>>>                 dt=pre_calc_vars['tn'][1] - pre_calc_vars['tn'][0],
>>>                 dB_threshold=6,
>>>                 rejectDuration=0.1,
>>>                 display=False,
>>>                 extent=pre_calc_vars['ext'])  
>>>     
>>>     indices = {'EVNspFract_per_bin': EVNspFract_per_bin, 'EVNspMean_per_bin':EVNspMean_per_bin , 'EVNspCount_per_bin':EVNspCount_per_bin, 'EVNsp':EVNsp}
>>>     return indices
>>> def get_spectral_activity(pre_calc_vars):
>>>     ACTspfract_per_bin, ACTspcount_per_bin, ACTspmean_per_bin = maad.features.spectral_activity(pre_calc_vars['Sxx_dB_noNoise'])
>>>     indices = {'ACTspfract_per_bin': ACTspfract_per_bin, 'ACTspcount_per_bin':ACTspcount_per_bin , 'ACTspmean_per_bin':ACTspmean_per_bin}
>>>     return indices
>>> acoustic_indices_methods = [get_aci, get_spectral_activity, get_spectral_events]
>>> 
>>> df_temp = df.iloc[0:1]
>>> segmented_df = utils.false_color_spectrogram_prepare_dataset(
>>>     df_temp, 
>>>     datetime_col = 'timestamp_init',
>>>     duration_col = 'duration',
>>>     file_path_col = 'file_path',
>>>     indices = ['acoustic_complexity_index', 'spectral_activity', 'spectral_events'], 
>>>     output_dir = './segmented_indices',
>>>     store_audio_segments = True,
>>>     unit = 'scale_02',
>>>     acoustic_indices_methods = acoustic_indices_methods,
>>>     pre_calculation_method = pre_calculation_method,
>>>     temp_dir = os.path.abspath('./temp_ac_files/'),
>>>     parallel = True
>>> )