Issues

HH_Param_Multi_Layer() accepts NaN parameters, so a trailing delimiter adds a layer of NaN and an empty cell gives NaN curves, without an error

Caltech-geoquake/PySeismoSoil#125 · opened

on Oct 14, 2026

Nonlinear_Simulation accepts MKZ_Param_Multi_Layer, as documented, but run() crashes because the NLHH kernel reads only HH parameters

Caltech-geoquake/PySeismoSoil#124 · opened

on Oct 14, 2026

Material-number error says "last error" instead of "last layer"

Caltech-geoquake/PySeismoSoil#121 · opened

on Oct 11, 2026

1

_calc_vertical_stress() changes the caller's thickness when the last layer is 0 m

Caltech-geoquake/PySeismoSoil#120 · opened

on Oct 11, 2026

One-layer Vs profiles crash HH parameter and damping calibration with IndexError

Caltech-geoquake/PySeismoSoil#119 · opened

on Oct 11, 2026

Tracking: all open issues by type, priority, area and status

Caltech-geoquake/PySeismoSoil#114 · opened

on Oct 11, 2026

_optimization_kernel() says it "further relaxes" the gamma_t lower bound to 0.005%, but raises it when the bound is already 0.001%

Caltech-geoquake/PySeismoSoil#113 · opened

on Oct 11, 2026

Three more tests check less than they say (r instead of R^2, atol=7.0 on damping, a clamped value checked for >= 0)

Caltech-geoquake/PySeismoSoil#112 · opened

on Oct 11, 2026

Slicing a Multiple_Damping_Curves object drops the original file name and sep, so get_all_HH_x_params()/get_all_H4_x_params() can no longer name or delimit their output

Caltech-geoquake/PySeismoSoil#111 · opened

on Oct 11, 2026

Parameter() raises TypeError with its documented default allowable_keys=None

Caltech-geoquake/PySeismoSoil#110 · opened

on Oct 11, 2026

Multiple_GGmax_Curves.get_curve_matrix() and Multiple_Damping_Curves.get_curve_matrix() raise ValueError when a shorter curve starts at strain 0

Caltech-geoquake/PySeismoSoil#109 · opened

on Oct 11, 2026

calc_GGmax_from_stress_strain() and calc_damping_from_stress_strain() do not check that strain and stress have the same length, and silently broadcast a length-1 array

Caltech-geoquake/PySeismoSoil#108 · opened

on Oct 11, 2026

Vs_Profile() input checks: an empty array raises IndexError, inf is reported as NaN, 2-column units are silently ignored, and one error branch is unreachable

Caltech-geoquake/PySeismoSoil#107 · opened

on Oct 11, 2026

Vs_Profile(add_halfspace=True) leaves material number 0 on the old last layer, so the new profile fails the class's own format check

Caltech-geoquake/PySeismoSoil#106 · opened

on Oct 11, 2026

Site_Factors._interpolate() is not trilinear, and its result depends on the order of the 8 corner points

Caltech-geoquake/PySeismoSoil#105 · opened

on Oct 11, 2026

Site_Factors() rejects Vs30 in (400, 450] m/s with z1 > 750 m, although the 450 m/s, 900 m data ships and loads

Caltech-geoquake/PySeismoSoil#104 · opened

on Oct 11, 2026

SVM.repr() rounds Vs30 and z1 to two significant digits, e.g. "Vs30 = 2.6e+02 m/s, z1 = 1e+02 m"

Caltech-geoquake/PySeismoSoil#103 · opened

on Oct 11, 2026

SVM() treats `Vs_cap=np.True_` as a 1 m/s cap, and accepts a `Vs_cap` below the surface Vs, which gives a non-monotonic profile

Caltech-geoquake/PySeismoSoil#102 · opened

on Oct 11, 2026

Batch_Simulation.run() adds an `output_dir` (or `sim_dir`) key to the caller's `options` dict

Caltech-geoquake/PySeismoSoil#101 · opened

on Oct 11, 2026

Four tests cannot fail for the problem they are named after

Caltech-geoquake/PySeismoSoil#100 · opened

on Oct 11, 2026

fit_all_damping_curves(), response_spectra() and Batch_GOF_Scores.run() never close their multiprocessing pools when parallel=True

Caltech-geoquake/PySeismoSoil#99 · opened

on Oct 11, 2026

linear_tf() frequency grid stops one step short of fmax for common inputs, including the defaults (fmax=30, freq_resolution=0.05)

Caltech-geoquake/PySeismoSoil#98 · opened

on Oct 11, 2026

fit_all_damping_curves(save_txt=True) checks func_serialize and txt_filename only after all the fitting is done

Caltech-geoquake/PySeismoSoil#97 · opened

on Oct 11, 2026

_align_two_time_arrays() adds an extra sample for some identical inputs, and always starts at dt instead of the records' start time

Caltech-geoquake/PySeismoSoil#96 · opened

on Oct 11, 2026

Vs_Profile() converts damping and density units inside the caller's array, and its private arrays are views of that array

Caltech-geoquake/PySeismoSoil#95 · opened

on Oct 11, 2026

Navigation Menu

Repository navigation

Issues

HH_Param_Multi_Layer() accepts NaN parameters, so a trailing delimiter adds a layer of NaN and an empty cell gives NaN curves, without an error

Nonlinear_Simulation accepts MKZ_Param_Multi_Layer, as documented, but run() crashes because the NLHH kernel reads only HH parameters

Material-number error says "last error" instead of "last layer"

_calc_vertical_stress() changes the caller's thickness when the last layer is 0 m

One-layer Vs profiles crash HH parameter and damping calibration with IndexError

Tracking: all open issues by type, priority, area and status

_optimization_kernel() says it "further relaxes" the gamma_t lower bound to 0.005%, but raises it when the bound is already 0.001%

Three more tests check less than they say (r instead of R^2, atol=7.0 on damping, a clamped value checked for >= 0)

Slicing a Multiple_Damping_Curves object drops the original file name and sep, so get_all_HH_x_params()/get_all_H4_x_params() can no longer name or delimit their output

Parameter() raises TypeError with its documented default allowable_keys=None

Multiple_GGmax_Curves.get_curve_matrix() and Multiple_Damping_Curves.get_curve_matrix() raise ValueError when a shorter curve starts at strain 0

calc_GGmax_from_stress_strain() and calc_damping_from_stress_strain() do not check that strain and stress have the same length, and silently broadcast a length-1 array

Vs_Profile() input checks: an empty array raises IndexError, inf is reported as NaN, 2-column units are silently ignored, and one error branch is unreachable

Vs_Profile(add_halfspace=True) leaves material number 0 on the old last layer, so the new profile fails the class's own format check

Site_Factors._interpolate() is not trilinear, and its result depends on the order of the 8 corner points

Site_Factors() rejects Vs30 in (400, 450] m/s with z1 > 750 m, although the 450 m/s, 900 m data ships and loads

SVM.repr() rounds Vs30 and z1 to two significant digits, e.g. "Vs30 = 2.6e+02 m/s, z1 = 1e+02 m"

SVM() treats `Vs_cap=np.True_` as a 1 m/s cap, and accepts a `Vs_cap` below the surface Vs, which gives a non-monotonic profile

Batch_Simulation.run() adds an `output_dir` (or `sim_dir`) key to the caller's `options` dict

Four tests cannot fail for the problem they are named after

fit_all_damping_curves(), response_spectra() and Batch_GOF_Scores.run() never close their multiprocessing pools when parallel=True

linear_tf() frequency grid stops one step short of fmax for common inputs, including the defaults (fmax=30, freq_resolution=0.05)

fit_all_damping_curves(save_txt=True) checks func_serialize and txt_filename only after all the fitting is done

_align_two_time_arrays() adds an extra sample for some identical inputs, and always starts at dt instead of the records' start time

Vs_Profile() converts damping and density units inside the caller's array, and its private arrays are views of that array