# HAR classification 
# Author: Burak Himmetoglu
# 8/15/2017

import pandas as pd 
import numpy as np
import os

def read_data(data_path, split = "train"):
	""" Read data """
	# Fixed params
	n_steps = 128

	# Paths
	path_ = os.path.join(data_path, split)
	path_signals = os.path.join(path_, "Inertial_Signals")

	# Read labels and one-hot encode
	label_path = os.path.join(path_, "y_" + split + ".txt")
	labels = pd.read_csv(label_path, header = None)

	# Read time-series data
	channel_files = os.listdir(path_signals)
	channel_files.sort()
	n_channels = len(channel_files)
	posix = len(split) + 5

	# Initiate array
	list_of_channels = []
	X = np.zeros((len(labels), n_steps, n_channels))
	i_ch = 0
	for fil_ch in channel_files:
		channel_name = fil_ch[:-posix]
		dat_ = pd.read_csv(os.path.join(path_signals,fil_ch), delim_whitespace = True, header = None)
		# X[:,:,i_ch] = dat_.as_matrix()
		X[:,:,i_ch] = dat_.iloc[:,:].values
		# Record names
		list_of_channels.append(channel_name)

		# Iterate
		i_ch += 1

	return X, labels[0].values, list_of_channels

def standardize(train, test):
	""" Standardize data """
	# Standardize train and test
	X_train = (train - np.mean(train, axis=0)[None,:,:]) / np.std(train, axis=0)[None,:,:]
	X_test = (test - np.mean(test, axis=0)[None,:,:]) / np.std(test, axis=0)[None,:,:]

	return X_train, X_test

def one_hot(labels, n_class = 8):
	""" One-hot encoding """
	expansion = np.eye(n_class)
	y = expansion[:, labels-1].T
	assert y.shape[1] == n_class, "Wrong number of labels!"

	return y

def get_batches(X, y, batch_size = 100):
	""" Return a generator for batches """
	n_batches = len(X) // batch_size
	X, y = X[:n_batches*batch_size], y[:n_batches*batch_size]

	# Loop over batches and yield
	for b in range(0, len(X), batch_size):
		yield X[b:b+batch_size], y[b:b+batch_size]