# 	Reads data from SPI.
#	Generates filter (coefficients) and applies them to data (signal)
#
#	The setup:
#	- Read data through MISO GPIO pin.
#	- Other GPIO pins for SPI are grounded.
#	- A voltage regulator is needed to convert from Arduinos 5V pins to Pi's 3.3V pins


from scipy import signal
import numpy as np
import spidev
import matplotlib.pyplot as plt


spi = spidev.SpiDev()

spi.open(0, 0)

ADC_data = []

#t_data = []

#	def tid(data):
#		for i in range(len()):
#			t.append =

''' Write ADC data and sampletime to datafile.txt '''
def datafile_write(input_data, runtime):
	print("Generating output file: datafile.txt")
	print("{0:>15}{1:>10.1f} samples".format("Samples:", len(input_data)))
	print("{0:>15}{1:>10.1f} us".format("Runtime:", runtime))
	interval = runtime / len(input_data)
	times = []
	current_time = 0

# Generates CSV file with [ADC_data][sampletime]
with open('datafile.txt', 'w') as datafile:
	for value in input_data:
		datafile.write(str(ADC_data * step))
		datafile.write("\t")
		datafile.write("{0:.3f}".format(current_time))
		datafile.write("\n")
		times.append(current_time)
		current_time += interval
	print("Datafile written")

def datafile_read():
	values = []
	times = []
	with open('datafile.txt', 'r') as datafile:
		for line in datafile:
			read_data = datafile.readline().split("\t")
			values.append(read_data[0])
			times.append(read_data[1])
	return values, times

''' Calculate filter coefficients '''
def filter_get_coefficients(N, fs, fc, filter_type):
	if filter_type == 'butter':
		b, a = signal.butter(N, Wn=fc / (fs / 2), btype='low')
	else:
		print("Incorrect filter type")

''' Apply filter '''
def datafile_apply_filter_to_data(data, b, a):
	data = np.asarray(data, dtype=np.float64)
	y = signal.lfilter(b, a, data)
	return y

''' Graph for filtered signal '''
def filtered_graph_generate(v, y, t):
	plt.clf()
	v = np.asarray(v, dtype=np.float64)
	plt.plot(t, v, 'b-', linewidth=1)
	plt.plot(t, y, 'g-', linewidth=0.8)
	plt.xlabel('Time [usec]')
	plt.grid()
	#plt.show()
	plt.savefig('filtered_data.png', dpi=500)

''' Plot filter response '''
def filter_plot_response(b, a, fs, fc, title, x_scale):
	plt.clf()
	w, h = signal.freqz(b, a, worN=8000)
	plt.plot(0.5 * fs * w / np.pi, np.abs(h), 'b')
	plt.axvline(fc, color='k')
	plt.xlim(0, x_scale * 0.5 * fs)
	plt.title(title)
	plt.xlabel('Frequency [Hz]')
	plt.grid()
	#plt.show()
	plt.savefig('filter.png', dpi=500)

def four_type_filter(fs, fc):
	b, a = signal.butter(N=5, Wn=fc / (fs / 2), btype='low')
	filter_plot_response(b, a, fc, fs, title="Butter LP Filter", x_scale=0.2)
	#plt.show()
	plt.clf()

while true:
	ADC_data.append = spi.xfer2([int(SPIsend)])

	''' fs = 30000, fc = 10000, order = 7, filter = Butterworth '''
	b, a = filter_get_coefficients(N=7, fs=30000, fc=10000, filter_type='butter')

	filter_plot_response(b, a, 30000, 10000, 'Butterworth LP Filter', 1)
	#print("b: {}".format(b))
	#print("a: {}".format(a))

	datafile_write(ADC_data, 10)
	v, t = datafile_read()
	y = datafile_apply_filter_to_data(v, b, a)
	filtered_graph_generate(v[1000:30000], y[1000:30000], t[100:10000])

spi.close()