How to plot an observer on a georeferenced tif file

I am new to the forum, but I thought here would be the best place to ask my question, since I could not find the anser so far anywhere else.

I want to visualize a georeferenced tif and plot an observer as a dot on the canvas.

Here is a link to the tif file:

I have written the following code, based on a notebook I found on geoviews:

import datashader as ds 
import geoviews as gv
import xarray as xr
import pandas as pd
import datashader.geo as dsgeo
from datashader.transfer_functions import shade, stack,dynspread
from datashader.colors import Elevation

# read tif
tif = xr.open_rasterio("dsm_sampled.tif")
# take a band and replace nan values
terrain = tif[0].fillna(0)

# create canvas
width = 1000
height = 1000

cvs = ds.Canvas(plot_width=width, plot_height=height)
illuminated = dsgeo.hillshade(terrain)
terrain_shaded = shade(terrain, cmap=Elevation, alpha=128, how='linear')

# some arbitrary position
x_pos = -243793.842
y_pos = -252500

observer_df = pd.DataFrame({'x': [x_pos],'y': [y_pos]})
observer_agg = cvs.points(observer_df, 'x', 'y')

observer_shaded = dynspread(shade(observer_agg, cmap=['orange']),
                            threshold=1, max_px=4)

# it is working if I leave out observer_shaded 

# stack(shade(illuminated, cmap=['black', 'white'], alpha=128, how='linear'),
#       terrain_shaded)

# This call is not working for some reason
stack(shade(illuminated, cmap=['black', 'white'], alpha=128, how='linear'),
      terrain_shaded, observer_shaded)`

But I am getting the following error and can not find a solution:

TypeError                                 Traceback (most recent call last)
<ipython-input-14-ebf9dfe88990> in <module>
     40 # This call is not working for some reason
     41 stack(shade(illuminated, cmap=['black', 'white'], alpha=128, how='linear'),
---> 42       terrain_shaded, observer_shaded)

~/opt/anaconda3/envs/grass/lib/python3.7/site-packages/datashader/transfer_functions/ in stack(*imgs, **kwargs)
    129     imgs = xr.align(*imgs, copy=False, join='outer')
    130     with np.errstate(divide='ignore', invalid='ignore'):
--> 131         out = tz.reduce(tz.flip(op), [ for i in imgs])
    132     return Image(out, coords=imgs[0].coords, dims=imgs[0].dims, name=name)

~/opt/anaconda3/envs/grass/lib/python3.7/site-packages/toolz/ in __call__(self, *args, **kwargs)
    301     def __call__(self, *args, **kwargs):
    302         try:
--> 303             return self._partial(*args, **kwargs)
    304         except TypeError as exc:
    305             if self._should_curry(args, kwargs, exc):

~/opt/anaconda3/envs/grass/lib/python3.7/site-packages/toolz/ in flip(func, a, b)
    735     [1, 2, 3]
    736     """
--> 737     return func(b, a)

TypeError: ufunc 'over' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''

If I am leaving out the observer point from the stack, everything is plotting fine. But with the observer point the plot breaks.

Sorry for the very late reply here. The problem (which should be detected with a better error message) is that your arrays do not line up with each other, i.e. without specifying an x_range and y_range on the Canvas it will fall back to auto-ranging and your single point will have a very different range than the other cmponents. You can fix this by computing the x_range and y_range from your terrain tif:

x_range = terrain.x.min().item(), terrain.x.max().item()
y_range = terrain.y.min().item(), terrain.y.max().item()

cvs = ds.Canvas(plot_width=width, plot_height=height, x_range=x_range, y_range=y_range)