Skip to content

Latest commit

 

History

History
190 lines (105 loc) · 7.21 KB

CHANGELOG.md

File metadata and controls

190 lines (105 loc) · 7.21 KB
Raw

Version 0.4.4

Date: 2022-08-01

This release achieves compatibility with recent versions of Dask, Pandas, PyArrow and Shapely.

Enhancements:

  • Avoid inappropriate geometry tests (#90)

Compatibility:

  • Fix Shapely deprecation warnings (#85)
  • Pandas extension array fixes (#88)
  • PyArrow and Dask parquet issues (#92)

Version 0.4.3

Date: 2021-08-07

Enhancements:

  • Expands the optional arguments that can be passed to to_parquet, to_parquet_dask, read_parquet, read_parquet_dask ensuring that storage_options is appropriately passed where needed. (#79)
  • Add typing information to some functions.
  • Update build_sindex to pass through kwargs to underlying HilbertRtree implementation.
  • Change build_sindex methods to consistently return reference to object to allow for method chaining.

Bug fixes:

  • Update internal validate_coerce_filesystem to pass storage_options through. (#78)

Compatibility:

  • Adds the ability to pass storage_options to the to_parquet function for pandas > 1.2, otherwise instantiates an fsspec filesystem with storage_options and passes that.
  • Renames fname parameter to path to align with the pandas convention.

Version 0.4.2

Date: 2021-07-28

This release primarily achieves compatibility with recent releases of Pandas. Many thanks to @Hoxbro for contributing the fixes and @philippjfr for ongoing maintenance of the project.

Compatibility:

  • Compatibility with Pandas>=1.3 (#76)

Version 0.4.1

Date: 2021-06-08

This release primarily achieves compatibility with recent releases of Dask. Many thanks to @jrbourbeau for contributing the fixes and @philippjfr for ongoing maintenance of the project.

Compatibility:

  • Compatibility with Dask>=2021.06.0 (#71)

Version 0.4.0

Date: 2021-05-25

Enhancements:

  • Add HoloViz build infrastructure. (#50)
  • Add --skip-slow and --run-slow options, slow tests are still run by default. (#60)
  • Add some type hints to parquet functions. (#60)
  • Allow using cx indexer without spatial index. (#54)
  • Switch to GitHub Actions. (#55)
  • Updated Geometry class eq method so that if other object is a container, the equality method on the container is called, so now performing an equality check between a geometry object and a geometry array should return correct results, which should be a bool array, whereas previously it would simply return False because the objects were not the same immediate type. (#60)
  • Update GeometryArray class eq method to allow comparison of an individual element to all objects in the array, returning an array of bool. (#60)
  • Add NotImplementedError for contains method. (#60)

Bug fixes:

  • Fix compatibility with latest pandas. (#55)
  • Fix certain tests fail due to hypothesis health check. (#60)
  • Pin numpy and dask on MacOS to pass tests. (#60)

Version 0.3.6

Date: 2020-08-16

Enhancements:

  • More intuitive error when lsuffix == rsuffix on sjoin (#35)

Bug fixes:

  • read_parquet_dask fails to read from s3 glob (#34)
  • Tests failing with ValueError: Cannot mask with a boolean indexer containing NA values (#41)
  • Tests in test_parquet failing with TypeError: argument of type 'PosixPath' is not iterable (#42)
  • Create temp directory for partitions explitictly, fixes failure of test test_pack_partitions_to_parquet

Compatibility:

  • Numba import updated to address deprecation warning (#36)

Version 0.3.5

Date: 2020-07-27

Bug fixes:

  • Fixed GeoDataFrame constructor exception when GeoPandas is not installed.

Version 0.3.4

Date: 2020-02-21

Bug fixes:

  • Support importing GeoPandas geometry series that contain None values.
  • Fixed abstract:// protocol error in pack_partitions_to_parquet when run on local filesystem.
  • Preserve active geometry column when importing GeoPandas GeoDataFrame.
  • Always load index columns when the columns argument is passed to read_parquet.

Compatibility:

  • Added support for pandas 1.0.
  • Added support for pyarrow 0.16. When 0.16 is available, the performance of read_parquet and read_parquet_dask is significantly improved.

Version 0.3.2 / 0.3.3

Date: 2020-01-24

Bug fixes:

  • Various reliability improvements for pack_partitions_to_parquet

Version 0.3.1

Date: 2020-01-12

Bug fixes:

  • Restored categories argument to read_parquet_dask function
  • Retry filesystem operations in pack_partitions_to_parquet using exponential backoff

Version 0.3.0

Date: 2020-01-11

Enhancements:

  • Added partial support for the intersects geometry array method. Currently, it only supports being called on Point arrays, but the elements of the array can be compared to any scaler geometry object (#21).
  • Added partial support for the sjoin spatial join function (#21).
  • Added support for glob path strings, and lists of path strings, to the read_parquet_dask method (#20)
  • Added bounds argument to read_parquet_dask to support filtering the loaded partitions to those that intersect with a bounding box (#20)
  • Added temp_format argument to the pack_partitions_to_parquet method to control the location of temporary files (#22)

Version 0.2.0

Date: 2019-12-28

Enhancements:

  • Added pack_partitions_to_parquet method to DaskGeoDataFrame (#19)
  • Added support for remote filesystems using the fsspec library (#19)

Version 0.1.1

Date: 2019-12-18

Enhancements:

  • Documented dependencies required for the Overview notebook (#18)

Bug fixes:

  • Fixed Ring.to_shapely error (#17)

Version 0.1.0

Date: 2019-12-02

First public release available on PyPI and the pyviz anaconda channel.