FeRx-NLME/ferx-core: v0.3.1

Résumé fourni par la source

What's Changed test(estimation): don't pin analytic/FD block_sigma OFVs to a shared basin by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/965 fix(estimation): cap analytic L-BFGS first step to clear warfarin FOCEI stall [closes #960] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/967 test(estimation): restore imp_defensive_alpha discrimination via pinned-ISCALE contrast [closes #961] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/966 chore(deps): bump the cargo-minor-patch group across 1 directory with 5 updates by @dependabot[bot] in https://github.com/FeRx-NLME/ferx-core/pull/916 chore(ci): collapse three overlapping test jobs into two non-overlapping ones by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/968 perf(ci): split per-PR jobs onto a ci-fast (lto="off") profile [#969] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/973 perf(sens): bucket the ODE IOV dual-width dispatch ladder [closes #971] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/974 fix(api): arm the non-PD SIR fallback under sir = true [closes #972] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/975 feat(io): read a missing DV as a design point when simulating [closes #957] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/981 feat: mixture models — [mixture] block, MIXNUM index, FOCE/FOCEI estimation [#977] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/980 feat(estimation): mixture SEs/covariance, p(k) form, .fitrx round-trip [#983] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/984 feat(estimation): inter-occasion variability under a mixture model [#985] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/986 feat(estimation): SAEM estimation for mixture models [#985] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/987 feat(estimation): Bayesian estimation for mixture models [#985] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/991 feat(estimation): IMP/IMPMAP estimation + objective evaluation for mixtures [#985] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/992 fix(estimation): recover a runaway inner EBE instead of certifying it [closes #958] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/998 fix(estimation): retry a fit that never left its initial estimates [closes #751] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/988 fix(saem): harden iiv_on_ruv re-centring + growth caps [#903 review] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/982 feat(parser): allow a model with no random effects (empty Omega) on continuous endpoints [closes #989] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1002 feat(estimation): class-aware mu-referencing for mixture models [closes #996] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/999 fix(parser): reject a dose attribute that is applied by the engine and read by the model [closes #993] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1003 docs(claude): downstream repos pin ferx-core transitively via ferx-r's lock, not directly by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1010 fix(estimation): SA-damp the SAEM numerical theta/sigma M-step [closes #1011] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1014 feat(estimation): SAEM warns when an estimated theta carries no ETA by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1012 fix(api): report a missing omega_iov instead of panicking when simulating an IOV model [closes #1019] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1036 fix(estimation): make SIR survive a rank-deficient proposal covariance [closes #1021] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1037 fix(ode): stop clamping negative Form C [scaling] readouts to zero [closes #1020] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1038 fix(ode): gate the adaptive DV monitor assay clamp on the readout [closes #1039] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1043 fix(parser): reject unknown [block] names instead of silently ignoring them [closes #1040] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1041 fix(parser): resolve TIME in a [scaling] Form C readout, and stop undefined names in [scaling] evaluating to 0 [closes #1028] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1042 fix(scaling): put the adaptive readout on the #1028 data clock, and fix the T escape hatch by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1045 feat(api): guard imp/impmap/bayes and gn at n_eta = 0 in ferx check [closes #1007, closes #1006] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1033 fix(parser): build the absorption ODE twin at parse time and decline on failure [closes #1008] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1027 fix(parser): reject a dose attribute mapped by pk(..., f=F) and also read on the analytical prediction path [closes #1004] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1035 fix(estimation): stop reporting trust_region as converged on MaxIters [closes #1000] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1034 fix(parser): reject a single-endpoint [error_model] with mis-ordered sigmas [closes #1001] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1025 feat(dsl): first-class residual weighting weight = , applied in every estimator [closes #1029] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1044 feat(parser): [scaling] intermediates, continuation lines, and 2-arg min/max [closes #1030] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1056 fix(parser): accept a dose attribute read by an [odes] init(...) seed [closes #1046] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1055 fix(parser): reject TIME in init() and publish the init scope rule [closes #994] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1061 feat(parser): sample-size-weighted IOV kappa K ~ γ² weight = N (BTAV) [closes #1031] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1062 refactor(estimation): extract fixed-eta observation-NLL gradients from saem by @dominic-irx in https://github.com/FeRx-NLME/ferx-core/pull/1015 fix(ci): repair the slow-tests push-leg path filter [closes #949] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1067 fix(sens): evaluate a moving dose boundary's two sides on their own covariate snapshots [closes #1060] by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1071 feat(parser): compartment-free $PRED-equivalent structural model [#811] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1063 feat(ode)!: a-priori stiffness probe, ode_method = auto, now the default [#978] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1066 feat(ode): report solver decisions, add a stiff-abort budget, cheapen the probe [#1080] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1087 fix(io): re-source the naproxen MBMA fixture to its primary publication [closes #1085] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1086 feat(api): let a [simulation] design state the covariates of the arms it invents [closes #1083] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1088 docs(claude): require every oracle to exercise every side of the object under test by @TeunP in https://github.com/FeRx-NLME/ferx-core/pull/1074 feat(parser): three-argument clamp(x, lo, hi) in the model DSL [closes #1092] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1093 fix(pk): evaluate the analytic Form C readout per occasion under IOV [closes #1079] by @roninsightrx in https://github.com/FeRx-NLME/ferx-core/pull/1090 docs: document packed parameter space and uncertainty draws by @roninsightrx in https:/