# Copyright (C) 2018 The Regents of the University of California, Michael Ludkovski and Aditya Maheshwari
# All Rights Reserved
# This code is published under the GNU Lesser General Public License (GNU LGPL)

from __future__ import division
import numpy as np
import math
import matplotlib.pyplot as plt
import microGrid.parameters as bv
import microGrid.parameters as bp


def residualDemand(param, nstep, nbsimulOpt, T, X0, typeRMC, sampleType="insample"):


	lambd = param.lambd
	sigma = param.sigma
	A = param.ampl
	mu = param.mu

	dt = T/nstep

	if sampleType=="insample":
		np.random.seed(seed=5)
	elif sampleType=="outsample":
		np.random.seed(seed=96)


	if typeRMC == 'regress now 2D':

		loadMatrix = np.zeros((nbsimulOpt,nstep+1))
		loadMatrix[:,0] = X0

		for i in range(1,nstep+1):

                        
			dW = np.random.normal(0,1,nbsimulOpt)*np.sqrt(dt)

			loadMatrix[:,i] = loadMatrix[:,i-1] + lambd*(mu - loadMatrix[:,i-1])*dt + sigma*dW #sigma*loadMatrix[:,i-1]*dW

		loadMatrix = np.where(loadMatrix>param.DemandUB,param.DemandUB,loadMatrix)


		inventory = np.random.uniform(param.I_minMax[0],param.I_minMax[1],nbsimulOpt)

		return loadMatrix, inventory

	if typeRMC == 'gd':


		loadMatrix = np.zeros((nbsimulOpt,nstep+1))
		loadMatrix[:,0] = X0

		for i in range(1,nstep+1):
                        
			dW = np.random.normal(0,1,nbsimulOpt)*np.sqrt(dt)

			loadMatrix[:,i] = loadMatrix[:,i-1] + lambd*(mu - loadMatrix[:,i-1])*dt + sigma*dW #sigma*loadMatrix[:,i-1]*dW

		inventory = np.linspace(param.I_minMax[0],param.I_minMax[1],param.meshI+1)

		loadMatrix = np.where(loadMatrix>param.DemandUB,param.DemandUB,loadMatrix)

		return loadMatrix, inventory


if __name__ == '__main__':
    

	param = bp.basicVariables()
	
	nbsimulOpt =5
	X0 = np.linspace(-param.ampl,param.ampl,nbsimulOpt)
	X0[:]=0
	# print len(X0)
	
	demand, inventory = residualDemand(param, param.nstep,  nbsimulOpt, param.maturity, X0, 'gd')
	print(len(inventory))
	# plt.figure(2)
	for simcnt in range(nbsimulOpt):
		plt.plot(demand[simcnt,:])
	plt.grid(True)
	plt.show()





	# param = bv.basicVariables()
	# print param.B_minMax, param.I_minMax[0],param.I_minMax[1], param.d_minMax, param.K

	# X0 = np.arange(-10,10,20/param.nbsimulOpt)
	# # matrix of prices with rows as (number of simulations) and columns as (number of time steps+1)
	# demand, inventory, control = residualDemand(param, param.nstep,  param.nbsimulOpt, param.maturity, X0, 'regress now 2D')

	# for i in range(0,100):
	# 	plt.plot(demand[i,:])
	# plt.show()