docs: changelog for float D-scale and N-Bit filter support
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
This commit is contained in:
+11
-5
@@ -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.
|
||||||
|
|||||||
Reference in New Issue
Block a user