You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The spline evaluation in shapes._curve_F99_method uses scipy.interpolate.splrep and splev, where the splrep documentation states
The default boundary condition is ‘not-a-knot’, i.e. the first and second segment at a curve end are the same polynomial. More boundary conditions are available in CubicSpline.
This is inconsistent with Fitzpatrick's IDL routine which uses CSPLINE, which uses SPL_INIT with default YP0 and YP1 arguments, which ultimately results in a spline with 'natural' boundary conditions (second derivative is 0 at ends) rather than 'not-a-knot' boundary conditions.
Severity of issue
The _curve_F99_method is called in the F99 and F04 dust laws.
The differences due to the boundary conditions is on the order of millimags, but varies with RV and wavelength as shown in the attached figure.
Issue
The spline evaluation in
shapes._curve_F99_method
usesscipy.interpolate.splrep
andsplev
, where the splrep documentation statesThis is inconsistent with Fitzpatrick's IDL routine which uses CSPLINE, which uses SPL_INIT with default YP0 and YP1 arguments, which ultimately results in a spline with 'natural' boundary conditions (second derivative is 0 at ends) rather than 'not-a-knot' boundary conditions.
Severity of issue
The
_curve_F99_method
is called in the F99 and F04 dust laws.The differences due to the boundary conditions is on the order of millimags, but varies with RV and wavelength as shown in the attached figure.
Proposed change
This can be fixed by replacing
with
The text was updated successfully, but these errors were encountered: