Potential bug in breakpoint_stretch

[millercommamatt]

goes2go/src/goes2go/accessors.py

Line 494 in 3917c5e

def breakpoint_stretch(C, breakpoint):

goes2go/src/goes2go/rgb.py

Line 439 in 3917c5e

def breakpoint_stretch(C, breakpoint):

These identical breakpoint_stretch functions take a breakpoint value as the second positional input, but it's not used. Subsequent calls to this function as part of NatrualColor provide different breakpoint values.

Is this intentional? I couldn't track down the source for the steps of the NatrualColor processing other than the creation of the pseudo green channel.

I'm guessing it's supposed to look something like this:

def breakpoint_stretch(C, breakpoint):
    """
    Contrast stretching by break point (number provided by Rick Kohrs)
    """
    lower = normalize(C, 0, breakpoint)  # Low end
    upper = normalize(C, breakpoint, 255)  # High end

    # Combine the two datasets
    # This works because if upper=1 and lower==.7, then
    # that means the upper value was out of range and the
    # value for the lower pass was used instead.
    combined = np.minimum(lower, upper)

    return combined