Releases: ArtifactDB/ritsuko
Releases · ArtifactDB/ritsuko
v0.6.1
v0.6.0
- Implement
ritsuko::hdf5::vlsutilities to handle custom variable-length string arrays. This uses a dataset of pointers that reference a (possibly compressed) heap to efficiently store and retrieve variable length string arrays. - Modify
open_and_load_*_missing_placeholder_attribute()functions to return astd::optional<>instance, which is more idiomatic. - Split
check_missing_placeholder_attribute()function into numeric and string versions to separate the logic.
v0.5.2
- Added utilities for block-wise iteration over a N-dimensional array.
- Migrated utility to validate N-dimensional string datasets from takane.
- Added utility to quickly get dataset dimensions.
- Bugfix for variable-length string streaming.
- Migrated alabaster.base functions to incrementally scan for special values.
v0.5.1
v0.5.0
- Added utilities to quickly load scalar, 1D numeric datasets.
- Extend the as_numeric_datatype to support non-fixed types.
- Avoid copying the PredTypes when we can just reference them.
- Check for string datatypes that use UTF-8-compatible encodings.
- Check for UTF-8 encoding in the miscellaneous attribute loaders.
v0.4.0
- Added utility to choose a HDF5 datatype from a C++ type.
- Added some small helpers for safely opening groups, files.
- Added wrapper function to load 1D numeric datasets/attributes.
- Added helper function to quickly load the missing placeholder.
- Added a class to stream numeric data from a 1D dataset.
- Checks for scalar datasets must now be performed separately.
- Stripped out all back-compatibility aliases.
- Replaced the string dataset loader with the Stream1dStringDataset.
- Minor bugfixes to headers and includes.
- Added more methods for validating and loading string datasets.
- Enable convenient extraction of attribute name via get_name.