ValueError when Displaying GeoTIFFs with geoviews.util.from_xarray(rioxarray.open_rasterio(filename))

[venuswku]

Hello GeoViews team! Thanks for taking the time to continuously maintain and improve this library. I'm having some difficulty trying to display TIFF files with the recent release (1.10.0), and I was hoping you could help me out.

ALL software version info

Geoviews 1.10.0 (Build pyhd8ed1ab_0, Channel conda-forge)
Bokeh 3.1.1 (Build pyhd8ed1ab_0, Channel conda-forge)
Rioxarray 0.14.1 (Build pyhd8ed1ab_0, Channel conda-forge)
Python 3.10.11 (Build h4de0772_0_cpython, Channel conda-forge)
JupyterLab 4.0.0 (Build pyhd8ed1ab_1, Channel conda-forge)
OS: Windows 11 Home (Version 22H2, Build 22621.1702)
Browser: Google Chrome (Version 113.0.5672.127, Official Build, 64-bit)

Description of expected behavior and the observed behavior

I was able to display GeoTIFFs with geoviews.util.load_tiff() before it got deprecated.

With the new release of GeoViews, I tried using the suggested geoviews.util.from_xarray(rioxarray.open_rasterio(filename)) but I was not able to get the same Image element as before. I got a Dataset element instead, so I cannot apply any image options to it (see stack traceback below).

Complete, minimal, self-contained example code that reproduces the issue

import geoviews as gv
import rioxarray as rxr
import cartopy.crs as ccrs
gv.extension("bokeh")

geotiff_img = gv.util.from_xarray(
    da = rxr.open_rasterio(filename = "https://github.com/mapbox/rasterio/raw/1.2.1/tests/data/RGB.byte.tif"),
    vdims = "Some Value",
    crs = ccrs.epsg(32618)
).opts(width = 500, height = 500, xaxis = None, yaxis = None)
geotiff_img

Stack traceback and/or browser JavaScript console output

ValueError                                Traceback (most recent call last)
Cell In[1], line 6
      3 import cartopy.crs as ccrs
      4 gv.extension("bokeh")
----> 6 geotiff_img = gv.util.from_xarray(
      7     da = rxr.open_rasterio(filename = "https://github.com/mapbox/rasterio/raw/1.2.1/tests/data/RGB.byte.tif"),
      8     vdims = "Some Value",
      9     crs = ccrs.epsg(32618)
     10 ).opts(width = 500, height = 500, xaxis = None, yaxis = None)
     11 geotiff_img

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\accessors.py:43, in AccessorPipelineMeta.pipelined.<locals>.pipelined_call(*args, **kwargs)
     40     inst._obj._in_method = True
     42 try:
---> 43     result = __call__(*args, **kwargs)
     45     if not in_method:
     46         init_op = factory.instance(
     47             output_type=type(inst),
     48             kwargs={'mode': getattr(inst, 'mode', None)},
     49         )

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\accessors.py:568, in Opts.__call__(self, *args, **kwargs)
    562         msg = ("Calling the .opts method with options broken down by options "
    563                "group (i.e. separate plot, style and norm groups) is deprecated. "
    564                "Use the .options method converting to the simplified format "
    565                "instead or use hv.opts.apply_groups for backward compatibility.")
    566         param.main.param.warning(msg)
--> 568 return self._dispatch_opts( *args, **kwargs)

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\accessors.py:572, in Opts._dispatch_opts(self, *args, **kwargs)
    570 def _dispatch_opts(self, *args, **kwargs):
    571     if self._mode is None:
--> 572         return self._base_opts(*args, **kwargs)
    573     elif self._mode == 'holomap':
    574         return self._holomap_opts(*args, **kwargs)

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\accessors.py:651, in Opts._base_opts(self, *args, **kwargs)
    648     return opts.apply_groups(self._obj, **dict(kwargs, **new_kwargs))
    650 kwargs['clone'] = False if clone is None else clone
--> 651 return self._obj.options(*new_args, **kwargs)

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\data\__init__.py:199, in PipelineMeta.pipelined.<locals>.pipelined_fn(*args, **kwargs)
    196     inst._in_method = True
    198 try:
--> 199     result = method_fn(*args, **kwargs)
    200     if PipelineMeta.disable:
    201         return result

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\data\__init__.py:1203, in Dataset.options(self, *args, **kwargs)
   1201 @wraps(Dimensioned.options)
   1202 def options(self, *args, **kwargs):
-> 1203     return super().options(*args, **kwargs)

File ~\anaconda3\envs\test\lib\site-packages\holoviews\core\dimension.py:1276, in Dimensioned.options(self, clone, *args, **kwargs)
   1274     expanded_backends = opts._expand_by_backend(options, backend)
   1275 else:
-> 1276     expanded_backends = [(backend, opts._expand_options(options, backend))]
   1278 obj = self
   1279 for backend, expanded in expanded_backends:

File ~\anaconda3\envs\test\lib\site-packages\holoviews\util\__init__.py:347, in opts._expand_options(cls, options, backend)
    345 objtype = objspec.split('.')[0]
    346 if objtype not in backend_options:
--> 347     raise ValueError(f'{objtype} type not found, could not apply options.')
    348 obj_options = backend_options[objtype]
    349 expanded[objspec] = {g: {} for g in obj_options.groups}

ValueError: Dataset type not found, could not apply options.

Screenshots or screencasts of the bug in action

I ran the example code in a Jupyter notebook and nothing except for the error gets outputted.

[venuswku]

@hoxbro Thanks for fixing my issue! Am I able to test your changes in a dev release? Or do I have to wait for a future release?

[hoxbro]

I will make a dev release for you.

[hoxbro]

The dev release is now out