def soft_threshold_colormap(cmap, threshold, vmax, subthreshold=None, symmetric=None):
"""Soft-threshold a colormap to make small values transparent
Parameters
----------
cmap : str
Base colormap.
threshold : scalar
Value at which to threshold the colormap (i.e., the value at which to
start the colormap).
vmax : scalar
Intended largest value of the colormap (used to infer the location of
the ``threshold``).
subthreshold : matplotlib color
Color of sub-threshold values (the default is the end or middle of
the colormap, depending on whether it is symmetric).
symmetric : bool
Whether the ``cmap`` is symmetric (ranging from ``-vmax`` to ``vmax``)
or not (ranging from ``0`` to ``vmax``). The default is ``True`` for
known symmetric colormaps and ``False`` otherwise.
Returns
-------
thresholded_cmap : matplotlib ListedColormap
Soft-thresholded colormap.
"""
assert vmax > threshold >= 0
cmap = mpl.cm.get_cmap(cmap)
if symmetric is None:
symmetric = cmap.name in symmetric_cmaps
colors = cmap(np.linspace(0., 1., cmap.N))
if subthreshold is None:
subthreshold_color = cmap(0.5) if symmetric else cmap(0)
else:
subthreshold_color = to_rgba(subthreshold)
n = int(round(vmax / ((vmax - threshold) / cmap.N)))
out_colors = np.empty((n, 4))
if symmetric:
i_threshold = int(ceil(cmap.N / 2))
out_colors[:i_threshold] = colors[:i_threshold]
out_colors[i_threshold:-i_threshold] = subthreshold_color
out_colors[-i_threshold:] = colors[-i_threshold:]
else:
out_colors[:-cmap.N] = subthreshold_color
out_colors[-cmap.N:] = colors
out = ListedColormap(out_colors, cmap.name)
out._eelbrain_meta = {
'symmetric': symmetric,
'vmin': -vmax if symmetric else 0,
'vmax': vmax,
}
return out
