v0.20.0 Bugfixes and additional methods for SparseFrames

Implements various bugfixes and some new methods to dask implementation of SparseFrame.

Integrate as dask extension

This fixes sparsity's compatibility with dask>0.19.2.

Accept pathlib objects in io module

This resolves problems that appeared after changing drtools' FileSystems behaviour.

Eventually this should be handled more elegantly. Currently there's some duplicated code which is the same as in filesystem module in drtools. Maybe we should make FileSystems a separate package (opensource) and use it both in sparsity and drtools?

v0.18.2 Support for latest pandas version

This adds back support for latest pandas version while still supporting older versions as well.
This also includes some bugfixes:

• loc indexing on duplicated indizes works now
• getitem indexing is closer to pandas behaviour
• fixed a bug which appeared with older scipy version (<1.0.0)

v0.18.1 Drop support for pandas>=0.23.0

Changelist:

• drop support for pandas>=0.23.0
• update dask collection usage

This release temporarily drops support for pandas >= 0.23.0 as the internal API changed and we have to adjust sparsity to reflect those changes. The next release will most probably add support for the latest pandas version.

v0.18.0 distributed procedures

Changes:

• distributed to_npz method (#39)
• distributed set_index method (#36)
• distributed sort_index (#37)
• distributed groupby sum (#35)
• add from_ddf method to initialize distributed sparse frames (#32)
• Support for distributed join (#34)
• Bugfix in SparseFrame.join method (#30)

v0.17.0

• Update to support dask collection interface new in version 0.16.0
• Support listlike label based indexing
• add support for arbritary remote storages
• support empty frames in elementwise operations

v0.16

Added multipart upload to support saving big datasets to s3.

v0.15

Fixes for new pandas versions, one-hot-encoding categorical columns...

v0.9.3

Mimic pandas api hotfix:

- add kwargs to add method