quantumclaw/clawhdf5
3
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

docs: changelog for float D-scale and N-Bit filter support

Browse Source 
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
This commit is contained in:
osobh
2026-06-03 12:36:50 +00:00
co-authored by Claude Opus 4.8
parent 4a5ab1c584
commit bff039fa29
1 changed files with 11 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
CHANGELOG.md
+11 -5
View File
@@ -3,11 +3,17 @@
## Unreleased ## Unreleased
### New Features ### New Features
- `clawhdf5-format`: decode the **scale-offset filter** (id 6), integer variant - `clawhdf5-format`: decode the **scale-offset filter** (id 6) — both the
(`H5Z_SO_INT`) — previously unsupported. Handles signed/unsigned element integer variant (`H5Z_SO_INT`) and the floating-point **D-scale** variant
sizes, negative minima and fill values; the on-disk format was reverse- (`H5Z_SO_FLOAT_DSCALE`). Handles signed/unsigned int sizes, f32/f64, negative
engineered against HDF5 2.0 and validated end-to-end. Floating-point minima, decimal scale factors and fill values; reverse-engineered against
scale-offset (D-scale/E-scale) remains unsupported. HDF5 2.0 and validated end-to-end. The float E-scale variant remains
unsupported.
- `clawhdf5-format`: decode the **N-Bit filter** (id 5), atomic variant —
previously unsupported. Unsigned reduced-precision datasets read end-to-end;
signed values are restored to canonical bytes (datatype-level sign extension
of reduced-precision integers is a separate, pre-existing concern). Recursive
compound/array N-Bit layouts remain unsupported.
### Bug Fixes ### Bug Fixes
- `clawhdf5-format`: read datasets written by modern HDF5 (1.14+/2.0, i.e. - `clawhdf5-format`: read datasets written by modern HDF5 (1.14+/2.0, i.e.