Release Notes¶

This is a list of changes to stac-geoparquet.

Unreleased¶

Drop Python 3.9 (stac-utils/stac-geoparquet!133)
Fixed resolution of scheme-prefixed output_paths (e.g., s3://bucket/item.parquet) (stac-utils/stac-geoparquet!143)
Fixed duplication of STAC Item properties with top-level keys (e.g., "collection") (stac-utils/stac-geoparquet!144)
Allow users to provide their own PgSTAC connection factory (stac-utils/stac-geoparquet!145)

Make deltalake an optional dependency (stac-utils/stac-geoparquet!106)
Updated pgstac_reader and removed Planetary Computer specific code (cli.py and pc_runner.py); added an example notebook for dumping pgstac partitions (stac-utils/stac-geoparquet!101)
Spec: use collections (plural) for storing collection metadata (stac-utils/stac-geoparquet!108)
Spill chunks to disk when writing parquet to merge schemas, and add chunking options to the pgstac readers (stac-utils/stac-geoparquet!109)
Removed the in-repo specification documents in favor of the upstream spec (stac-utils/stac-geoparquet!120)
Fixed stac_table_to_ndjson to always insert a type field if one isn't already present (stac-utils/stac-geoparquet!105)
Fixed a case where conversion to JSON created assets values as None (stac-utils/stac-geoparquet!111)
Fixed pgstac_to_arrow to pass through tmpdir (stac-utils/stac-geoparquet!118)
Fixed pyarrow.dataset import in arrow._api (stac-utils/stac-geoparquet!119)
Hydrate links in PgstacRowFactory (stac-utils/stac-geoparquet!121)
Convert Path to str for remote filesystem compatibility (stac-utils/stac-geoparquet!123)

Updated stac-geoparquet Parquet File metadata fields (stac-utils/stac-geoparquet!98)
All fields are placed under the stac-geoparquet key in the parquet file metadata.
Added a version field.
Added json-schema for the metadata parquet file metadata.
Fixed ValueError when converting some strings to dates (stac-utils/stac-geoparquet#79)
Removed upper bound on pyarrow dependency (stac-utils/stac-geoparquet!102)