Skip to content

pyiron_base-0.4.0
Compare
Choose a tag to compare
@niklassiemer niklassiemer released this 18 Nov 17:22
pyiron_base-0.4.0
9253669
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.

pyiron_base<=0.3.10 has a bug that writes all arrays with dtype=object even numeric ones. In this new release the writing of the arrays is fixed.
Since this affects the content of the hdf file this is a minor version bump.
Although it was intended, such dtype=object arrays are not yet always automatically converted into numeric ones in the reading procedure in this release. This will be fixed asap.

Merged PRs:

  • Convert dtype=object arrays if possible (#518)
  • Do not copy files on restart (#511)
  • Refactor global state (#486)
  • Faster hdf reading (#512)
  • Correctly check for ragged array (#503) - Fixing the major bug
  • Test for type conservation (#508)
  • Worker Job Class (#497)
  • Bugfix in case of None type values (#488)
  • Avoid compressing only if zipped file name has the job name in it (#507)
  • erase lines in hdfio.py (#233)
  • Clean up subprocess and multiprocessing calls (#498)
  • Do not lose track of dtypes in get_array_ragged (#494)
  • Save pyiron tables via DataContainer instead of csv (#463)
  • dependency updates: #500, #515
Assets 2
Loading