Issues

Multiple shooting drops the per-measurement binding table when it splits data into segments, so a per-measurement model under job_type = ms is refused with the wrong cause

lanl/PyBNF#1049 · opened

on Nov 5, 2026

A .prop constraint at a time past the experiment's t_end is reported as satisfied and adds no penalty

lanl/PyBNF#1048 · opened

on Nov 5, 2026

An lwf_continue run on another numpy version or CPU cannot write what the uninterrupted run wrote, and the state file's fingerprint records neither

lanl/PyBNF#1046 · opened

on Nov 5, 2026

Without petab installed, objective = expression fails with a message about an 'expression observableFormula'

lanl/PyBNF#1045 · opened

on Nov 5, 2026

continue_run applies adaptive_files/MLE_params.txt by position, with no check of its length or of which parameter each value belongs to

lanl/PyBNF#1044 · opened

on Nov 5, 2026

convert.py silences petab 0.10's whole parameter-scale warning, including its note that converted parameter-scale priors change the MAP estimate

lanl/PyBNF#1043 · opened

on Nov 5, 2026

uv.lock still resolves bngsim 0.15.0, below the bngsim>=0.17.0 floor in pyproject.toml

lanl/PyBNF#1042 · opened

on Nov 5, 2026

The overwrite and continue prompts crash with 'an unknown error occurred: EOFError' when stdin is not a terminal

lanl/PyBNF#1041 · opened

on Nov 5, 2026

On the gradient path a .net model ends the whole fit on any simulation failure, where an SBML model scores the same failure as a failed point

lanl/PyBNF#1040 · opened

on Nov 5, 2026

A gradient fit whose difference-quotient fallback crosses a moving branch fails every point and reports success at objective inf

lanl/PyBNF#1039 · opened

on Nov 5, 2026

job_type = ms silently drops an atol=>/rtol=> that is not a plain number, and no path refuses a negative or non-finite atol

lanl/PyBNF#1038 · opened

on Nov 5, 2026

parameter_scan and bifurcate on the bngsim backend ignore the action's atol=> and rtol=>

lanl/PyBNF#1037 · opened

on Nov 5, 2026

petab 0.9's petab1to2 drops a pre-equilibration whose condition changes nothing, and PyBNF then fits an experiment that starts from the initial conditions

lanl/PyBNF#1036 · opened

on Nov 5, 2026

A malformed noise_model line is reported as "noise_model is not a valid configuration key", and every config parse error reports the line before the right one

lanl/PyBNF#1033 · opened

on Nov 5, 2026

The Docker image leaves out bngsim for a reason that no longer holds, so it cannot run 89 of the 97 tutorial configs

lanl/PyBNF#1032 · opened

on Nov 5, 2026

Tests that importorskip('petab') skip when petab is installed but broken, so a broken petab passes CI as skipped tests

lanl/PyBNF#1031 · opened

on Nov 5, 2026

The publish workflow uploads a wheel that no job has installed: install it and run a small fit on Linux and macOS before the PyPI upload

lanl/PyBNF#1027 · opened

on Nov 4, 2026

PyBNF lists Python 3.14 as supported, but bngsim has no 3.14 wheels, so pip install pybnf there compiles bngsim from source, and no CI job tests that install

lanl/PyBNF#1022 · opened

on Nov 4, 2026

bootstrap never resamples an objective = expression fit's data: every replicate refits the full data set

lanl/PyBNF#1015 · opened

on Nov 4, 2026

start_point.txt labels values set by starting_params as 'sampled'

lanl/PyBNF#1014 · opened

on Nov 4, 2026

starting_params is not checked at load: an out-of-box value leaves am stuck outside the box for the whole run, extra values are dropped, too few crash with IndexError

lanl/PyBNF#1013 · opened

on Nov 4, 2026

A continue=>1 simulate into a new suffix makes BNG2.pl write its .gdat without a header, and PyBNF reads the first data row as the header, so every evaluation on bngl_backend = bionetgen fails

lanl/PyBNF#1012 · opened

on Oct 16, 2026

A BNGL simulate() or parameter_scan() with no method runs as an ODE on the bngsim backend, but BNG2.pl refuses it, so the default backend fits a model that bngl_backend = bionetgen cannot run, and neither refuses it at load

lanl/PyBNF#1008 · opened

on Oct 15, 2026

A parameter_scan whose parameter is not in the model is accepted at load and, on the bngsim network-free path, runs every scan point at the model's own value with no message, so a dose-response is fitted to a flat curve

lanl/PyBNF#1007 · opened

on Oct 15, 2026

A network-free simulate() with t_start above 0 labels its rows t_start…t_end on the bngsim backend but 0…(t_end − t_start) on BNG2.pl, so the same data are scored against different simulated times on the two backends and nothing warns

lanl/PyBNF#1006 · opened

on Oct 15, 2026

Navigation Menu

Repository navigation

Issues

Multiple shooting drops the per-measurement binding table when it splits data into segments, so a per-measurement model under job_type = ms is refused with the wrong cause

A .prop constraint at a time past the experiment's t_end is reported as satisfied and adds no penalty

An lwf_continue run on another numpy version or CPU cannot write what the uninterrupted run wrote, and the state file's fingerprint records neither

Without petab installed, objective = expression fails with a message about an 'expression observableFormula'

continue_run applies adaptive_files/MLE_params.txt by position, with no check of its length or of which parameter each value belongs to

convert.py silences petab 0.10's whole parameter-scale warning, including its note that converted parameter-scale priors change the MAP estimate

uv.lock still resolves bngsim 0.15.0, below the bngsim>=0.17.0 floor in pyproject.toml

The overwrite and continue prompts crash with 'an unknown error occurred: EOFError' when stdin is not a terminal

On the gradient path a .net model ends the whole fit on any simulation failure, where an SBML model scores the same failure as a failed point

A gradient fit whose difference-quotient fallback crosses a moving branch fails every point and reports success at objective inf

job_type = ms silently drops an atol=>/rtol=> that is not a plain number, and no path refuses a negative or non-finite atol

parameter_scan and bifurcate on the bngsim backend ignore the action's atol=> and rtol=>

petab 0.9's petab1to2 drops a pre-equilibration whose condition changes nothing, and PyBNF then fits an experiment that starts from the initial conditions

A malformed noise_model line is reported as "noise_model is not a valid configuration key", and every config parse error reports the line before the right one

The Docker image leaves out bngsim for a reason that no longer holds, so it cannot run 89 of the 97 tutorial configs

Tests that importorskip('petab') skip when petab is installed but broken, so a broken petab passes CI as skipped tests

The publish workflow uploads a wheel that no job has installed: install it and run a small fit on Linux and macOS before the PyPI upload

PyBNF lists Python 3.14 as supported, but bngsim has no 3.14 wheels, so pip install pybnf there compiles bngsim from source, and no CI job tests that install

bootstrap never resamples an objective = expression fit's data: every replicate refits the full data set

start_point.txt labels values set by starting_params as 'sampled'

starting_params is not checked at load: an out-of-box value leaves am stuck outside the box for the whole run, extra values are dropped, too few crash with IndexError

A continue=>1 simulate into a new suffix makes BNG2.pl write its .gdat without a header, and PyBNF reads the first data row as the header, so every evaluation on bngl_backend = bionetgen fails

A BNGL simulate() or parameter_scan() with no method runs as an ODE on the bngsim backend, but BNG2.pl refuses it, so the default backend fits a model that bngl_backend = bionetgen cannot run, and neither refuses it at load

A parameter_scan whose parameter is not in the model is accepted at load and, on the bngsim network-free path, runs every scan point at the model's own value with no message, so a dose-response is fitted to a flat curve

A network-free simulate() with t_start above 0 labels its rows t_start…t_end on the bngsim backend but 0…(t_end − t_start) on BNG2.pl, so the same data are scored against different simulated times on the two backends and nothing warns