zarr.tree() example in docs doesn't work as of 2.4.0

[ryan-williams]

Minimal, reproducible code sample, a copy-pastable example if possible

Taken from the docs:

import h5py
import zarr
import numpy as np
source = h5py.File('data/example.h5', mode='w')
foo = source.create_group('foo')
baz = foo.create_dataset('bar/baz', data=np.arange(100), chunks=(50,))
spam = source.create_dataset('spam', data=np.arange(100, 200), chunks=(30,))
zarr.tree(source)

Problem description

Tree widget renders correctly in Zarr 2.3.2:

Same code throws a ValueError (and renders some text) in 2.4.0:

Version and installation information

2.3.2 works as expected; pip freeze:

appnope==0.1.0
asciitree==0.3.3
backcall==0.1.0
decorator==4.4.1
fasteners==0.15
h5py==2.10.0
ipykernel==5.1.4
ipython==7.12.0
ipython-genutils==0.2.0
jedi==0.16.0
jupyter-client==5.3.4
jupyter-core==4.6.2
monotonic==1.5
numcodecs==0.6.4
numpy==1.18.1
parso==0.6.1
pexpect==4.8.0
pickleshare==0.7.5
prompt-toolkit==3.0.3
ptyprocess==0.6.0
Pygments==2.5.2
python-dateutil==2.8.1
pyzmq==18.1.1
six==1.14.0
tornado==6.0.3
traitlets==4.3.3
wcwidth==0.1.8
zarr==2.3.2

2.4.0 exhibits the failure; pip freeze:

appnope==0.1.0
asciitree==0.3.3
backcall==0.1.0
decorator==4.4.1
fasteners==0.15
h5py==2.10.0
ipykernel==5.1.4
ipython==7.12.0
ipython-genutils==0.2.0
jedi==0.16.0
jupyter-client==5.3.4
jupyter-core==4.6.2
monotonic==1.5
numcodecs==0.6.4
numpy==1.18.1
parso==0.6.1
pexpect==4.8.0
pickleshare==0.7.5
prompt-toolkit==3.0.3
ptyprocess==0.6.0
Pygments==2.5.2
python-dateutil==2.8.1
pyzmq==18.1.1
six==1.14.0
tornado==6.0.3
traitlets==4.3.3
wcwidth==0.1.8
zarr==2.4.0

(the Zarr version is the only difference)

#450 seems relevant (cc @jakirkham); possibly discussion at #259 as well.

[jakirkham]

Are you rendering it in the notebook? If so, you need to install ipytree now. This dependency was added so rendering would work in JupyterLab. Though, once installed, it should work in the notebook too.

Sounds like we could use a clear error message though.

[jakirkham]

@ryan-williams? 🙂

[mzjp2]

Should we add ipytree to our requirements in that case? (or make it an extras_require?)

[jakirkham]

It is listed in the extras_require.

[mzjp2]

Ah, my bad. Thanks. :) Agree with your earlier comment, we could do with a clearer error message.

…

On 22 Feb 2020, at 18:28, jakirkham ***@***.***> wrote:  It is listed in the extras_require. — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or unsubscribe.

[jakirkham]

No worries 😉

Yeah we can make this fallback behavior nicer 🙂