app.py

from flask import Flask, request, render_template, Response, stream_with_context

# AUDIO imports
import pyaudio
import struct
import numpy as np
from scipy.fftpack import fft

app = Flask(__name__)

# AUDIO: constants
CHUNK = 1024 * 2  # samples per frame
FORMAT = pyaudio.paInt16  # audio format (bytes per sample?)
CHANNELS = 1  # single channel for microphone
RATE = 44100  # samples per second
DEVICE_INDEX = 1

p = pyaudio.PyAudio()

# print devices
info = p.get_host_api_info_by_index(0)
numdevices = info.get('deviceCount')
for i in range(0, numdevices):
    if (p.get_device_info_by_host_api_device_index(0, i).get('maxInputChannels')) > 0:
        print("i = " + str(i) + ". Input Device id = ", i, " - ", p.get_device_info_by_host_api_device_index(0, i).get('name'))

print("\n - - -")
print("SELECTED Input Device id ", DEVICE_INDEX, " - ", p.get_device_info_by_host_api_device_index(0, DEVICE_INDEX).get('name'))
print("\n - - -")

# x = np.arange(0, 2 * CHUNK, 1)
# create an evenly divided array for the freq spectrum. 0-44100
xf = np.linspace(0, RATE, CHUNK)

# stream object to get data from microphone
stream = p.open(
    format=FORMAT,
    channels=CHANNELS,
    rate=RATE,
    input=True,
    output=True,
    #input_device_index=DEVICE_INDEX,
    frames_per_buffer=CHUNK
)

print('stream started')

@app.route('/')
def index():
    return render_template('index.html')

@app.route('/wavefeed')
def wavefeed():
    return Response(get_sound_wave(), mimetype="text/plain")

@app.route('/spectrumfeed')
def spectrumfeed():
    return Response(get_sound_freq(), mimetype="text/plain")

def get_sound_wave():
    # get sound data
    # binary data
    data = stream.read(CHUNK)

    # convert data to integers, make np array, then offset it by 127
    data_int = struct.unpack(str(2 * CHUNK) + 'B', data)

    # create np array and offset by 128
    data_np = np.array(data_int, dtype='b')[::2] + 128

    x = data_np.tolist()

    yield str(x)

def get_sound_freq():
    # get sound data
    # binary data
    data = stream.read(CHUNK)

    # convert data to integers, make np array, then offset it by 127
    data_int = struct.unpack(str(2 * CHUNK) + 'B', data)

    # create np array and offset by 128
    #data_np = np.array(data_int, dtype='b')[::2] + 128

    yf = fft(data_int)
    a = yf[0:CHUNK]
    a1 = np.abs(a)
    b = (128 * CHUNK)
    spectrum_data = a1 / b
    result = spectrum_data[::4]

    x = result.tolist()

    yield str(x)