BABAPPAlign v1.1.3
Patch release (no algorithmic changes)
Fixed
- Enforced mandatory explicit specification of the neural scoring model (--model) with no silent fallback.
- Restored correct progressive profile construction, ensuring all input sequences are retained in the final alignment.
- Ensured scoring model validation occurs before any embedding inference for improved correctness and reproducibility.
Notes
- The alignment algorithm, neural scoring model, and numerical behavior are unchanged.
- This release resolves a packaging-related regression introduced during CLI enforcement.
- Version bump issued to ensure immutable source distribution compatibility with downstream package managers (e.g. Bioconda).