numpy.isposinf

Here are the examples of the python api numpy.isposinf taken from open source projects. By voting up you can indicate which examples are most useful and appropriate.

3 Examples 7

Example 1

Project: scikit-rf Source File: mathFunctions.py
def inf_to_num(x):
    '''
    converts inf and -inf's to large numbers

    Parameters
    ------------
    x : array-like or number
        the input array or number
    Returns
    -------
    '''
    #TODO: make this valid for complex arrays
    try:
        x[npy.isposinf(x)] = INF
        x[npy.isneginf(x)] = -1*INF

    except(TypeError):
        x = npy.array(x)
        x[npy.isposinf(x)] = INF
        x[npy.isneginf(x)] = -1*INF

Example 2

Project: geostatsmodels Source File: zscoretrans.py
def to_norm( data ):
    '''
    Input  (data) 1D NumPy array of observational data
    Output (z)    1D NumPy array of z-score transformed data
           (inv)  inverse mapping to retrieve original distribution
    '''
    # look at the dimensions of the data
    dims = data.shape
    # if there is more than one dimension..
    if len( dims ) > 1:
        # take the third column of the second dimension
        z = data[:,2]
    # otherwise just use data as is
    else:
        z = data
    # grab the number of data points
    N = len( z )
    # grab the cuemulative distribution function
    f, inv = cdf( z )
    # h will return the cdf of z
    # by interpolating the mapping f
    h = fit( f )
    # ppf will return the inverse cdf
    # of the standard normal distribution
    ppf = scipy.stats.norm(0,1).ppf
    # for each data point..
    for i in range( N ):
        # h takes z to a value in [0,1]
        p = h( z[i] )
        # ppf takes p (in [0,1]) to a z-score
        z[i] = ppf( p )
    # convert positive infinite values
    posinf = np.isposinf( z )
    z = np.where( posinf, np.nan, z )
    z = np.where( np.isnan( z ), np.nanmax( z ), z )
    # convert negative infinite values
    neginf = np.isneginf( z )
    z = np.where( neginf, np.nan, z )
    z = np.where( np.isnan( z ), np.nanmin( z ), z )
    # if the whole data set was passed, then add the
    # transformed variable and recombine with data
    if len( dims ) > 1:
        z = np.vstack(( data[:,:2].T, z )).T
    return z, inv

Example 3

Project: scikit-aero Source File: test_isentropic.py
def test_area_ratio_no_zero_division_error():
    fl = isentropic.IsentropicFlow()
    assert np.isposinf(fl.A_Astar(0))