Coerce mef.get_transform_fxn input to permit None values

[JS3xton]

Coerce mef_values input to mef.get_transform_fxn() to be a numpy array of floats. This coercion converts None values to np.nan, thereby permitting None as a synonym for np.nan. The coercion also eliminated the need for some type checking, as numpy now throws an error if the inner arrays don't have the same length.

Closes #214.

All unit tests pass before (9cd83ef) and after (7be7ffa) the change in Python 3.8 + Anaconda 2020.02 and Python 2.7 + Anaconda 4.4.0.

The following code failed previously at the last line and is fixed by the change:

import numpy as np
import FlowCal as fc
beads_sample = fc.io.FCSData('./examples/FCFiles/Beads006.fcs')
beads_sample = fc.transform.to_rfi(beads_sample, ['FSC','SSC','FL1','FL2','FL3'])
beads_sample = fc.gate.start_end(beads_sample, 250, 100)
beads_sample = fc.gate.high_low(beads_sample, channels=['FSC','SSC'])
beads_sample = fc.gate.density2d(data=beads_sample, channels=['FSC','SSC'], gate_fraction=0.85, xscale='logicle', yscale='logicle', sigma=5.)

mef_values   = [0, 646, 1704, 4827, 15991, 47609, 135896, 273006]
mef_output   = fc.mef.get_transform_fxn(beads_sample, mef_values, mef_channels='FL1', clustering_channels=['FL1', 'FL3'], plot=False)

mef_values   = [0, 646, 1704, 4827, 15991, 47609, 135896, np.nan]
mef_output   = fc.mef.get_transform_fxn(beads_sample, mef_values, mef_channels='FL1', clustering_channels=['FL1', 'FL3'], plot=False)

mef_values   = [0, 646, 1704, 4827, 15991, 47609, 135896, None]
mef_output   = fc.mef.get_transform_fxn(beads_sample, mef_values, mef_channels='FL1', clustering_channels=['FL1', 'FL3'], plot=False)

NOTE: I first tried using pd.isnull in place of np.isnan, but fit_beads_autofluorescence() would fail later in err_fun(). pd.isnull checks for other null types (like None) in addition to np.nan. I think the failure occurred because mef_values=[0,...,None] later yields an object array instead of a numerical array. As such, I decided to just ensure mef_values was a numerical (float) array from the outset.