Repository navigation
ACTINV v1.1.1
ACTINV v1.1.1 — release notes
ACTINV 1.1.1 is a data-catalog patch release with one solver hardening fix. Schemas, solver order, public
interfaces and physics are unchanged from 1.1.0. The embedded data catalog advances to v1.1.0 so that
actinv data fetch installs the tendl-2025-patched remediation derivative qualified in P25c (data-v1.1.0
release).
Upgrade the Python package with:
python -m pip install --upgrade actinvRust CLI users can install this exact patch with:
cargo install --locked --force actinv-cli --version 1.1.1What changes
- The embedded catalog becomes
actinv-data-catalog-v1.1.0.json; installed data lives under
actinv-data/v1.1.0/. - The new default bundle
tendl-2025-patched-neutroninstalls the P25c-qualified patched 709-group neutron
activation library — an Avila Labs remediation derivative of TENDL-2025 that zeroes exactly the 44 enumerated
leaked leading ordinates across the 28 confirmed-signature files and is byte-identical to the official archive
everywhere else. It is not an official TENDL release. tendl-2025-patched-neutron-covarianceadds the MF=33 covariance sidecar rebuilt against the patched artifact;
the sidecar index cryptographically pins the patched library's SHA-256.actinv newnow emitscatalog:tendl-2025-patched-neutron-709greferences.- All v1.0.0 artifact IDs resolve to exactly the same bytes: the proton, deuteron, alpha and decay assets are
unchanged, and the supersededtendl-2025-neutronbundles remain fetchable for byte-compatible reproduction of
earlier work. - CRAM linear solves now apply bounded iterative refinement (
Lu::solve_refined) with compensated residual
evaluation; the run certificate'snumerical_floornote no longer claims a total numerical-error bound. Results
are unchanged on previously green cases; the fix removes a trace-inventory precision loss in shifted solves.
Qualification boundary
The patched corpus passed the frozen P25c construction, coverage, surgical-integrity, leak-clearance and
retrospective nonregression gates — retrospective evidence of internal consistency, not blind validation. The
shipped patched artifact holds 1,679 targets / 87,075 rows: the full-corpus strict build ledgered 1,172
source-file exclusions across all defect classes. Five dosimetry-critical targets (Ni-58, Nb-93, Ag-109, In-113,
Au-197) still fail closed under non-signature defect classes, and 16 of the 17 previously defect-bearing IRDFF-II
targets are absent. ACTINV and the data remain research-grade; docs/DATA_LIMITATIONS.md and
docs/QUALIFICATION.md still apply.