Fix bucket resamplers trying to print non-existent name

[BENR0]

This fixes bucket resampler AttributeError from log message in compute methods.

• Closes Bucket resamplers AttributeError in logging message #1647
• Tests added

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (main@cb3e197). Click here to learn what that means.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1648   +/-   ##
=======================================
  Coverage        ?   92.30%           
=======================================
  Files           ?      258           
  Lines           ?    38088           
  Branches        ?        0           
=======================================
  Hits            ?    35157           
  Misses          ?     2931           
  Partials        ?        0           

Flag	Coverage Δ
behaviourtests	4.84% <0.00%> (?)
unittests	92.84% <100.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
satpy/resample.py	80.24% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cb3e197...1619173. Read the comment docs.

[djhoese]

Do you have an example piece of code that shows you running in to this error? As far as I can tell Satpy should always be passing a DataArray to .compute. 🤔 unless the resampler in pyresample does something different.

[BENR0]

I just use scn.resample(area_def, resampler="bucket_avg"). In Line 1112 the np.array is extracted from the DataArray and then handed to the compute method of the resampler:

satpy/satpy/resample.py

Line 1112 in 9257378

data_arr = data.data

At least for the bucket_avg resampler the bug was introduced in #1539 I think.

[djhoese]

@pnuu do you know if the compute method is supposed to be expecting a dask array or a DataArray? The referenced pull request adds the log message by @ameraner but the DataArray -> dask change was there before.

[pnuu]

According to the docstring resampler.get_average() expects to receive either dask or Numpy array. I think @BENR0 is correct and the resamplers are always passed a plain dask array.

[djhoese]

Ok so dask arrays should have a .name if I remember correctly, but I don't think it is useful as it is usually based on the last task in the dask graph and includes a generated UUID most of the time (not super human-readable). If we are trying to get the name of the DataArray then we should really be doing data_arr.attrs['name']...but if we are trying to print out the DataArray's identifier then we should really be printing out its DataID. @mraspaud is the proper way to get the DataID for a DataArray to check for .attrs['_satpy_id']?

[mraspaud]

Yes