lattice/pyerrors
1
0
Fork 0
mirror of https://github.com/fjosw/pyerrors.git synced 2025-05-15 12:03:42 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix: Error handling for fits and root finding with numpy instead of autograd.numpy

Browse source 
improved. Tests added.
This commit is contained in:
Fabian Joswig 2022-03-02 15:10:07 +00:00
parent 7f5989dfb9
commit b14314b424
4 changed files with 53 additions and 10 deletions
Download patch file Download diff file Expand all files Collapse all files

10
pyerrors/fits.py
View file

@ -260,7 +260,10 @@ def total_least_squares(x, y, func, silent=False, **kwargs):
output.chisquare_by_expected_chisquare)
fitp = out.beta
hess_inv = np.linalg.pinv(jacobian(jacobian(odr_chisquare))(np.concatenate((fitp, out.xplus.ravel()))))
try:
hess_inv = np.linalg.pinv(jacobian(jacobian(odr_chisquare))(np.concatenate((fitp, out.xplus.ravel()))))
except TypeError:
raise Exception("It is required to use autograd.numpy instead of numpy within fit functions, see the documentation for details.") from None
def odr_chisquare_compact_x(d):
model = func(d[:n_parms], d[n_parms:n_parms + m].reshape(x_shape))
@ -537,7 +540,10 @@ def _standard_fit(x, y, func, silent=False, **kwargs):
output.chisquare_by_expected_chisquare)
fitp = fit_result.x
hess_inv = np.linalg.pinv(jacobian(jacobian(chisqfunc))(fitp))
try:
hess_inv = np.linalg.pinv(jacobian(jacobian(chisqfunc))(fitp))
except TypeError:
raise Exception("It is required to use autograd.numpy instead of numpy within fit functions, see the documentation for details.") from None
if kwargs.get('correlated_fit') is True:
def chisqfunc_compact(d):

5
pyerrors/roots.py
View file

@ -31,7 +31,10 @@ def find_root(d, func, guess=1.0, **kwargs):
# Error propagation as detailed in arXiv:1809.01289
dx = jacobian(func)(root[0], d.value)
da = jacobian(lambda u, v: func(v, u))(d.value, root[0])
try:
da = jacobian(lambda u, v: func(v, u))(d.value, root[0])
except TypeError:
raise Exception("It is required to use autograd.numpy instead of numpy within root functions, see the documentation for details.") from None
deriv = - da / dx
res = derived_observable(lambda x, **kwargs: (x[0] + np.finfo(np.float64).eps) / (d.value + np.finfo(np.float64).eps) * root[0], [d], man_grad=[deriv])