import DDC_python as ddp


# helper function
def __index(x, vec, default):
    for i in range(0, len(vec)):
        if x == vec[i]:
            return i + 1
    return default


def DDC(X, DDCpars = {}):

    if DDCpars == None:
        DDCpars = []

    if not isinstance(DDCpars, dict):
        raise TypeError('DDCpars must be a dictionary')

    if not 'numDiscrete' in DDCpars.keys():
        DDCpars['numDiscrete'] = 3

    if not 'precScale' in DDCpars.keys():
        DDCpars['precScale'] = 1e-12

    if not 'cleanNAfirst' in DDCpars.keys():
        DDCpars['cleanNAfirst'] = 'automatic'

    if not 'tolProb' in DDCpars.keys():
        DDCpars['tolProb'] = 0.99

    if not 'corrlim' in DDCpars.keys():
        DDCpars['corrlim'] = 0.5

    if not 'combinRule' in DDCpars.keys():
        DDCpars['combinRule'] = 'wmean'

    if not 'returnBigXimp' in DDCpars.keys():
        DDCpars['returnBigXimp'] = False

    if not 'silent' in DDCpars.keys():
        DDCpars['silent'] = False

    if not 'nLocScale' in  DDCpars.keys():
        DDCpars['nLocScale'] = 25000

    if not 'fastDDC' in DDCpars.keys():
        if X.shape[2] < 750:
            fastDDC = 0
            DDCpars['fastDDC'] = False
        else:
            fastDDC = 1
            DDCpars['fastDDC'] = True
    else:
        fastDDC = DDCpars['fastDDC'] + 1

    if DDCpars['fastDDC']:
        if X.shape[2] < 500:
            fastDDC = 0
    else:
        if X.shape[2] < 2000:
            print("Consider using the option 'fastDDC == TRUE' which runs much faster on datasets with many variables.")

    if not 'standType' in DDCpars.keys():
        DDCpars['standType'] = '1stepM'

    if not 'corrType' in DDCpars.keys():
        DDCpars['corrType'] = 'gkwls'

    if not 'nbngbrs' in DDCpars.keys():
        DDCpars['nbngbrs'] = 100

    # other parameters
    if not 'coreOnly' in DDCpars.keys():
        DDCpars['coreOnly'] = False

    if not 'tolProbCell' in DDCpars.keys():
        DDCpars['tolProbCell'] = DDCpars['tolProb']

    if not 'tolProbRow' in DDCpars.keys():
        DDCpars['tolProbRow'] = DDCpars['tolProb']

    if not 'tolProbReg' in DDCpars.keys():
        DDCpars['tolProbReg'] = DDCpars['tolProb']

    if not 'tolProbCorr' in DDCpars.keys():
        DDCpars['tolProbCorr'] = DDCpars['tolProb']

    if not 'includeSelf' in DDCpars.keys():
        DDCpars['includeSelf'] = 1

    if not 'numiter' in DDCpars.keys():
        DDCpars['numiter'] = 1

    if not 'qdim' in DDCpars.keys():
        DDCpars['qdim'] = 30

    if not 'nCorr' in DDCpars.keys():
        DDCpars['nCorr'] = 100

    transFun  = __index(DDCpars['transFun'], ['Huber', 'wrap', 'rank'], 2)
    standType = __index(DDCpars['standType'], ['1stepM','hubhub','wrap','mcd', 'rawmcd', 'wrapmedmad'], 1) - 1
    corrType  = __index(DDCpars['corrType'], ['wrap', 'rank', 'gkwls'], 3)
    combiRule = __index(DDCpars['combiRule'], ['wmean','wmedian','mean','median'], 1)

    fracNA = DDCpars['fracNA']
    numDiscrete = DDCpars['numDiscrete']
    preScale = DDCpars['preScale']
    returnBigXimp = DDCpars['returnBigXimp']
    silent = DDCpars['silent']
    cleanNAfirst = DDCpars['cleanNAfirst']

    if DDCpars['coreOnly'] == False: