[asdf] Do not copy array data

[marscher]

Changes

Because asdf does not handle time datatypes (correctly), there was a line of code within the serialization of xarray Variable to convert all ndarray like types from the asdf ndarray wrapper. This effectively prevents the use of the "copy_arrays" flag, since the
data got always copied.

Related Issues

Closes #454, #110

Checks

• updated CHANGELOG.md
• updated tests
• updated doc/
• update example/tutorial notebooks

[pep8speaks]

Hello @marscher! Thanks for updating this PR.

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-08-02 17:06:44 UTC

[codecov]

Codecov Report

Merging #456 (0135a18) into master (282ed4a) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##           master     #456   +/-   ##
=======================================
  Coverage   96.85%   96.85%           
=======================================
  Files          88       88           
  Lines        5595     5599    +4     
=======================================
+ Hits         5419     5423    +4     
  Misses        176      176           

Impacted Files	Coverage Δ
weldx/asdf/tags/weldx/core/common_types.py	100.00% <100.00%> (ø)
weldx/asdf/tags/weldx/core/data_array.py	100.00% <100.00%> (ø)
weldx/visualization/k3d_impl.py	79.68% <0.00%> (+0.08%)	⬆️

---

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 282ed4a...0135a18. Read the comment docs.

[CagtayFabry]

nice, does this work now with copy_arrays?

we should probably dig into the (asdf related) code for more np.array(), np.asarray() etc. calls

[marscher]

Our unit tests switch the copy_arrays flag at least. But I doubt that the underlying functionality of mem mapping is really being tested. We can argue that this is already tested extensively upstream.

[CagtayFabry]

Our unit tests switch the copy_arrays flag at least. But I doubt that the underlying functionality of mem mapping is really being tested. We can argue that this is already tested extensively upstream.

that is certainly not tested, we just use these parameters to catch any errors that might flare up
we would probably have to look into a memory profiler to see the difference

[marscher]

I got my memory profiler booted up anyways for another PR, so eventually I'll take a look.