#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Tue May 21 17:36:58 2024

@author: jdaniel
"""

import numpy as np
from matplotlib import pyplot as plt

def AltMarket(num_samples,nummarkets,varmarket,variances):
    num_vals = num_samples
    var = varmarket
    nreg = len(variances[:,0])
    ndim = nummarkets
    a = np.zeros([ndim,ndim])
    rans = np.random.randn(num_vals,ndim)
    c = np.copy(rans)
    c = np.zeros((num_vals,nreg,ndim))
    for i in range(0,ndim):
        for j in range(0,ndim):
            a[i,j] = -var/(float(ndim)-1)
            a[i,i] = var*1.00001
    b = np.linalg.cholesky(a)
    for i in range(0,num_vals):
        for j in range(0,nreg):
            c[i,j,:] = np.matmul(b,rans[i,:])
    return(c)

def AltMarket_FractionMethod(num_samples,nummarkets,varmarket,variances):
    num_vals = num_samples
    var = varmarket
    ndim = nummarkets
    nreg = len(variances[:,0])
    ndimactual = len(variances[0,:])
    rans = np.random.randn(num_vals,nreg,ndim)
    rans = np.random.randn(num_vals,ndim)
    c = np.zeros((num_vals,nreg,ndim))
    for i in range(0,num_vals):
        for j in range(0,nreg):
            if (ndim == ndimactual):
                #c[i,j,:] = rans[i,:] * .4*variances[j,k]
                for k in range(0,ndim):
                #    c[i,j,k] = rans[i,k] * .5*variances[j,k]
                    c[i,j,k] = rans[i,k] * np.max([.2*variances[j,k],.05])
            else:
                for k in range(3,ndim-1):
                    c[i,j,k] = rans[i,k] * np.max([.2*variances[j,k],.05])
    return(c)

def marketmc(ranArr,marketArr):
    nsample = len(ranArr[:,0,0])
    nran = len(ranArr[0,0,:])
    nregion = len(marketArr[:,0])
    nmarket = len(marketArr[0,:])
    marketOut = np.zeros([nsample,nregion,nmarket])
    marketOut[0,:,:] = marketArr[:,:]
    for i in range(1,nsample):
        for j in range(0,nregion):
            marketOut[i,j,:] = marketArr[j,:] + ranArr[i,j,:]
            kk = np.where(marketOut[i,j,:] > 1)
            ii = np.where(marketOut[i,j,:] < 0)
            jj = np.where(marketOut[i,j,:] >= 0)
            ishape = np.shape(ii)
            if (ishape[1] != 0):
                marketOut[i,j,ii] = 0. 
                sumval = np.sum(marketOut[i,j,:])
                marketOut[i,j,jj] = marketOut[i,j,jj] / sumval
    return(marketOut)