Repository navigation
DynamicDL Headless v0.1.1-alpha
Pre-release
Pre-release
DynamicDL 0.1.1-alpha is out!
Changelog
New features
- Added support for ImpliedList object to represent implicit index data.
- Added support for CSV file parsing
- Added support for the Namespace class
- Added support for Partial and Combo types
- Added support for progress bars and descriptions for parsing
- Added support for DataType instantiation
- Added token types
GENERIC_INT,GENERIC_WORD,GENERIC_QUANTITYas well as their corresponding Token classes - Reworked the TXTFile parser to be able to handle more advanced use cases
- Revamped the docs to be much more complete and follow RTD convention
Bugfixes
- Critical: Reworked the merge algorithm. Previously there were issues with merging where redundant types would not merge correctly when presented as keys in the dataset, as a hierarchy issue. This has been resolved and the entire algorithm now has been reworked to accommodate the change as the previous version relied on some finicky casework to function
- Critical: Fixed a pairing bug which had unexpected values and obsolete variables, adjusted the algorithm to run an altered version of the merge algorithm
- Various small bugfixes
Other changes
- Renamed the
Imageclass toImageEntryto avoid possible conflicts with the widely-usedPIL.Imageclass - Removed the incomplete
inferencemethod - Reworked some DynamicDL class methods to accept kwargs instead for parameter
mode - Refactored nearly every class into separate modules for organization