import numpy as np
import ctypes
import wave
replaygain = ctypes.CDLL('libreplaygain.so')


def string_to_samples(string, channel_count):
    """
    Takes a byte string of raw PCM data and returns a float32 numpy array containing
    audio data in range [-1, 1].
    """
    samples = np.fromstring(string, dtype='int16')
    samples = samples / float(2**15)
    samples.astype(np.float32)
    frame_count = samples.size / channel_count
    return samples.reshape([frame_count, channel_count])


def read_wav(f, start=None, end=None):
    # Opening the file and getting infos
    raw = wave.open(f, 'rb')
    channel_count = raw.getnchannels()
    sample_width = raw.getsampwidth()       # Sample width in byte
    if sample_width != 2: raise ValueError('Wave format not supported')
    frame_rate = raw.getframerate()

    # Calculating start position and end position
    # for reading the data
    if start is None: start = 0
    start_frame = start * frame_rate
    if end is None: end_frame = raw.getnframes()
    else: end_frame = end * frame_rate
    frame_count = end_frame - start_frame

    # Reading only the data between `start` and `end`,
    # putting this data to a numpy array 
    raw.setpos(int(start_frame))
    data = raw.readframes(int(frame_count))
    data = string_to_samples(data, channel_count)
    return data, {'frame_rate': frame_rate, 'channel_count': channel_count}


def calculate_replaygain(samples, frame_rate=44100):
    """
    pass in a generator that produces audio samples
    returns the maximum sample.

    https://github.com/vontrapp/replaygain
    """
    replaygain.gain_init_analysis(frame_rate)
    block_size = 10000
    channel_count = samples.shape[1]
    i = 0
    samples = samples.astype(np.float64)

    while i * block_size < samples.shape[0]:
        channel_left = samples[i*block_size:(i+1)*block_size,0]
        channel_right = samples[i*block_size:(i+1)*block_size,1]

        samples_p_left = channel_left.ctypes.data_as(ctypes.POINTER(ctypes.c_double))
        samples_p_right = channel_right.ctypes.data_as(ctypes.POINTER(ctypes.c_double))

        replaygain.gain_analyze_samples(samples_p_left, samples_p_right, channel_left.shape[0], channel_count)
        i += 1
    
    return replaygain.gain_get_chapter()


if __name__ == '__main__':
    samples, infos = read_wav('directions.wav')
    gain = calculate_replaygain(samples, frame_rate=infos['frame_rate'])
    print "Recommended gain: %f dB" % gain
    gain = calculate_replaygain(np.random.random((441000, 2)) * 2 - 1, frame_rate=44100)
    print "Recommended gain: %f dB" % gain