The changelog includes:
Bug fixes (6)
- ORF boundary errors on negative strands — Fixed inversion logic for CDS coordinates during protein maximization.
- Miniprot empty file crashes — Added explicit null guards to prevent
AttributeError: 'NoneType'when Miniprot fails mapping. - Recursive object overwriting in run_liftoff.py — Fixed a variable scoping issue that overwrote the parent Lifton_GENE when traversing nested features.
- Fasta target initialization faults — Added explicit checks to gracefully handle missing or inaccessible Fasta genomes before running the pipeline.
- Unhandled out-of-bounds array indices — Secured DNA sequence extraction with bounds checking to prevent crashes near chromosome edges.
- Non-string integer crash — Hardened structural serialization logic so numeric attributes don't trigger exception traces during final output.
Improvements (4)
- Enhanced pipeline preprocessing hardening — Native Python wrappers now catch runtime errors cleanly and print graceful degradation context instead of unstructured tracebacks.
- Comprehensive unit testing architecture — Added over 60 new test modules verifying database initialization, extraction bounds, recursion logic, and the protein-maximization algorithm.
- Subprocess native stream handling — External tools (Liftoff/Miniprot) correctly capture standard error and provide highly readable validation contexts on failure.
- Package version modernization — Bumped definition to v1.0.8 across setup.py, configuration YAMLs, shell scripts, and documentation.
Documentation:
The link to the documentation: http://ccb.jhu.edu/lifton/
PyPi release
LiftOn v1.0.8 is available on PyPi: https://pypi.org/project/lifton/