GitLab

Batch reader class with below blueprint would be useful to have. Implementation could largely be adapted from filter_analyzer.py

class BatchReader:
    """ Audio file batch reader.

        Reads audio file(s) and serves them in batches.

        Args:
            input: str
                File name, list of file names, or directory name 
            rate: float
                Sampling rate in Hz
            dt_format: str
                Format for parsing date-time data from file names
            overlap: int
                Size of overlap region (number of samples) used for smoothly joining audio signals 
    """
    def __init__(self, input, rate, dt_format=None, overlap=100):
        self.rate = rate
        self.overlap = overlap
        self.index = 0
        self.times = list()
        load(input=input, dt_format=dt_format)

    def load(input, dt_format=None):
        
        if input is file:
            files = [input]
        elif input is list of files:
            files = input
        else: # input is directory
            # find files
        
        if dt_format is not None:
            # attempt to parse date-time data
        
        # sort chronologically

    def next(int max_size=None):
        # loop over files, read and merge, stop when size > max_size
        # return copy of merged signal and update file counter
        # clear local copy of merged signal
        # return None, if there are no more files to read

    def reset()
        self.index = 0
        self.times = list()

    def log():
        # return pandas dataframe with filenames and start times