dask

What happened:

If a negative value for drop_axis is passed into either map_blocks or map_overlap a non-informative exception is raised.

What you expected to happen:

I would expect this would work as in NumPy for negative axis arguments where axis becomes axis = axis % array.ndim. If it is not intended to work, then it should raise a user-friendly AxisError. This came up

Support Series.between

As far as I know, there's no way to see what options are enabled.

It'd be nice to add either xr.get_options or xr.show_options or xr.options so that users can see what is enabled.

I think it would be nice to have a one-line description under each option

"arithmetic_join": "outer"
    controls DataArray/Dataset alignment in binary operations.

The stumpy.snippets feature is now completed in #283 which follows this work:

• The original published paper
• The supporting website with data examples

We have a rough notebook t

Similar to dask/dask#7800, we can replace our tmpfile context manager

https://github.com/dask/distributed/blob/5e5b9835a5e6929bb97acf94f0a2eecccc0b6a2f/distributed/utils.py#L862-L863

with tempfile.TemporaryFile from the standard library.

Describe the bug
According to the multiscene documentation, the property all_same_area does:

Determine if all contained Scenes have the same ‘area’.

However, I have created a multiscene where all scenes have the same area (they just differ between datasets), yet the property returns Fa

Code Sample, a minimal, complete, and verifiable piece of code

from pyresample.boundary import Boundary
b = Boundary(my_lons, my_lats)
print(b.contour_poly.area())

Problem description

The above code doesn't fail if the provided lons/lats are 2D (not sure on 3D+), but the class and all functions/utilities underneath it assume 1D arrays. The end results are incor

This is annoying because you need to scroll up a lot to look at the actual useful output. There is probably a way either to catch these warnings with pytest or to ignore them at the end of the report.

pytest dask_jobqueue

Part of the output at the end of the run:

dask_jobqueue/tests/test_pbs.py::test_job_script[PBSCluster]
  /home/lesteve/miniconda3/lib/python3.9/site-packag

Problem description

Reading a dataset with eager's read functionality raises a ValueError when providing columns.

Example code (ideally copy-pastable)

import pandas as pd

from tempfile import TemporaryDirectory
from functools import partial
from storefact import get_store_from_url

from kartothek.io.eager import store_dataframes_as_dataset, read_dataset_as_data

The ML implementation is still a bit experimental - we can improve on this:

• SHOW MODELS and DESCRIBE MODEL
• Hyperparameter optimizations, AutoML-like behaviour
• @romainr brought up the idea of exporting models (#191, still missing: onnx - see discussion in the PR by @rajagurunath)
• and some more showcases and examples

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

Describe the solution you'd like

https://labs.quansight.org/blog/2021/08/github-actions-benchmarks/

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've c

Currently all of the metrics computed are independent of a target variable or column, but if lens.summarise took the name of a column as the target variable, the output of some metrics could be more interpretable even if the target variable is not used in any kind of predictive modelling.

A good example of this could be PCA (see #14), which could plot the different categories of the target va

Without thinking I put resampling="bilinear" and got an error when I called .compute()

Traceback (most recent call last):
  File "carajas.py", line 92, in <module>
    band_medianNP = band_median.compute()
  File "/home/ubuntu/anaconda3/envs/richard/lib/python3.8/site-packages/xarray/core/dataarray.py", line 899, in compute
    return new.load(**kwargs)
  File "/home/ubuntu/anaco

As per @krfricke, we should create a single RayCommunication actor class that holds both an asyncio Queue and asyncio Event. This way we only need one remote actor for all communication between training workers and the driver, and we don't have to be dependent on ray.util.Queue