inverse_transform doesn't work on 1D data

Is it possible to do an inverse transform on a 1D embedding? This is what I see:

from umap import UMAP
import numpy as np

x = np.random.random(size = (100,1000))
umap_1D = UMAP(n_components=1)
transformed = umap_1D.fit_transform(x)
umap_1D_max = transformed.max()
umap_1D_min = transformed.min()
to_be_inverted = np.random.random(size=(1,10))
generated = umap_1D.inverse_transform(to_be_inverted)
generated

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-45-2c8de5c09c47> in <module>()
      8 umap_1D_min = transformed.min()
      9 to_be_inverted = np.random.random(size=(1,10))
---> 10 generated = umap_1D.inverse_transform(to_be_inverted)
     11 generated

/usr/local/lib/python3.6/dist-packages/umap/umap_.py in inverse_transform(self, X)
   2202         # build Delaunay complex (Does this not assume a roughly euclidean output metric)?
   2203         deltri = scipy.spatial.Delaunay(
-> 2204             self.embedding_, incremental=True, qhull_options="QJ"
   2205         )
   2206         neighbors = deltri.simplices[deltri.find_simplex(X)]

qhull.pyx in scipy.spatial.qhull.Delaunay.__init__()

qhull.pyx in scipy.spatial.qhull._Qhull.__init__()

ValueError: Need at least 2-D data

Start at umap/umap_.py in inverse_transform, especially the scipy.spatial.Delaunay call shown in the traceback. Reproduce the supplied 1D embedding example and trace how the one-dimensional embedding reaches that entry point. Done means inverse_transform handles n_components=1 without the reported ValueError.