v0.6.0

What's Changed

• Implements hive partitioning by @alpha-beta-soup in #40
• adds -g/--geo enum for GeoParquet output by @alpha-beta-soup in #42

Full Changelog: v0.5.0...v0.6.0

v0.5.0

What's Changed

• Initial test infrastructure including smoke tests for supported indexing functions by @ndemaio in #33
• Object oriented refactor of DGGS integration scripts by @ndemaio in #34
• Recursion bugfix and tiny documentation update by @ndemaio in #35
• Adds support for the A5 DGGS by @alpha-beta-soup in #36
• Adds support for -b/--band to avoid reading or writing irrelevant bands by @alpha-beta-soup in #37

Full Changelog: v0.4.2...v0.5.0

v0.4.2

Dependabot, documentation, CLI help

• https://github.com/manaakiwhenua/raster2dggs/security/dependabot/30
• Adds compaction to the README
• Doesn't refer to H3 in the compaction help text for non-H3 DGGSs

Full Changelog: v0.4.1...v0.4.2

v0.4.1

What's Changed

• sort before writing for better compression by @alpha-beta-soup in #32

Full Changelog: v0.4.0...v0.4.1

v0.4.0

Compression and compaction

What's Changed

• Made sure the parent resolution is of type int internally by @ndemaio in #26
• Implements DGGS compaction by @alpha-beta-soup in #31
• widens compression options by @alpha-beta-soup in #30

Full Changelog: v0.3.0...v0.4.0

---

Compaction

Sample images for compaction (https://github.com/mommermi/geotiff_sample), rounding all bands down to the nearest 100, and only compacting where all three bands with a common ancestor have the same value. (Images only showing B01, but compaction is across all bands.)

H3

rHEALPix

Geohash

S2

---

Compression

Sample compression results for H3, compacted as above, but at a higher maximum resolution (13); and with the following forms of compression:

Compression	Size
none	43 MB
lz4	5.8 MB
brotli	3.3 MB
snappy	6.3 MB
zstd	4.9 MB

The input image is 5.7 MB at 10 m resolution, and with whatever compression that sample image comes with.

H3 represenation of the image requires 2.091 million cells, vs 1.002 million cells for the raster (so this is 2x over-sampled). But note that sample image values are rounded to the nearest 100 for the demonstration of compaction, so the comparison really isn't 1:1.

Even higher compression and compaction would be expected for true categorical raster data (or vector data).

v0.3.0

What's Changed

• Adds S2 support by @alpha-beta-soup in #24
• adds support for geocodes: Geohash and Maidenhead (QTH/IARU) by @alpha-beta-soup in #22

Full Changelog: v0.2.7...v0.3.0

v0.2.7

What's Changed

• demaion/nics mods by @ndemaio in #20

New Contributors

• @ndemaio made their first contribution in #20

Full Changelog: v0.2.6...v0.2.7

v0.2.6

Full Changelog: v0.2.3...v0.2.6

v0.2.3

Previously, output partitions were labelled sequentially as part.{i}.parquet for i in 0...n, where n was the number of unique H3 parent cells in the dataset, one part per parent.

To make this easier to consume, we now label these as {h3parent}.parquet, which allows users to select a particular partition based on a H3 cell without having to do anything except list the output directory. The output order is the same since the sort order was already based on a lexicographic order of the H3 cells.

v0.2.2

Enhancements:

• #13