- CHANGELOG.rst entries now with most recent changes at the top.
- In rv.py, ensure n_mugrid_1, mugrid_1, n_mugrid_2, mugrid_2 are set if flux_weighted=False
- In J0113+31.py, use verbose=0, add usecols for reading some data files, and change samples = sampler.chain[:, nruns//2:, :].reshape((-1, ndim)), and skip error on calculation of autocorrelation length
- Added scipy, emcee, corner and matplotlib to requirements in setup.py
- Changed url in setup.py to github page.
- Added printing of spot parameters if verbose >= v_user
- Fixed bug with shape of spots_1 and spots_2 arrays in ellc.f90 and used order='F' to create of spot parameter arrays in lc.py, rv.py, fluxes.py
- Pull request #16 "Update ellc for compatibility with Python 3.12"
- fix to bug in starshape for some compilers due to floating point initialisation of a variable.
- correction to inline docs for definition of domdot
- catch and fix cases of wrong root returned in roche_l1
- added exit-with-best-guess to brent for cases where convergence fails
- added installation hint for mac users
- output from fluxes() now normalized consistently with output from lc()
- fixed bugs in fluxes.py
- improved reliability of bright() function for the case exact_grav=True
- fix for case of exactly overlapping circles.
- Fixed bug in lc() that caused a crash NaN appears in the array of calculated fluxes and verbose=True
- Removed "external" declrations in gauss_legendre.f90.
- Fixed indentation error in fluxes.py
- Made bright() routine more robust for the case radius_1 = radius_2
- Added iteration limits to while loops in starshape()
- Added elegent exit from failed calculation of integration limits.
- Added tolerance to minimum area for numerical integration to avoid numerical errors for near-tangent ellipse intersections.
- Caught error in ell_ell_overlap due to four very close near-tangent points
- Made subroutine bright() more robust when working near the limb.
- Changed default grid size to "default" in rv() and fluxes()
- Fixed bug in calling of ellc.rv with flux_weighted=True
- Added "love" model for star shape from Correia et al., 2014A&A...570L...5C
- Added "power-2" limb darkening law
- Added tolerance within which to set mu values > 1 to mu=1 in stellar.f90
- Fixed problems with MANIFEST.in
- Bug fix for total eclipses with spots introduced in previous version.
- Added note to in-line documentation regarding simplified reflection and geometric albedo.
- Added "mugrid" limb darkening law.
- Updated factor for converting K and P to a.sini to be consistent with nominal solar constants from IAU Resolution B3. (https://www.iau.org/static/resolutions/IAU2015_English.pdf)
- Changed polynomial root polishing algorithm in ell_ell_roots to Maehly's procedure.
- Fixed bug with eclipses of spots on the limb (iiswitch values swapped).
- Updated machine constants to double precision in solve_real_poly.f90
- Limit correction for eclipsed spots to less than eclipsed flux.
Tidy up output of ldy.list_bands() - fixes crash in ellc_emcee Made use of priors in ellc_emcee python3 safe. Added "import setuptools" to setup.py
Added ugriz bands to ldy
Revised coordinate systems used internally. This is needed for a consistent fix to a bug in RM effect in 1.4.3 and a bug in the flux-weighted radial velocity when using detailed reflection.
Added module coords.f90
Added GD448_rv.py to examples/GD448 to test RV+heating calculations.
Added notes/figures on coordinate systems in subdirectory doc/.
Fixed bug initialisation of vsini_2 value in lc.py and rv.py. This changes the output of examples/PhotRM/PhotRM.py
Improved calculation of gravity in the exact_grav case - now done on the Roche equipotential at the same angular coordinate as the ellipsoid surface.
Added option to use an improved calculation of ellipsoid size for Roche potential case so that volume of the Roche equipotential surface is constant in an eccentric orbit and is equal to the volume of a sphere with radius specified by the user. With the star model "roche", the volume is calculated either the volume of the approximating ellipsoid. For synchronous rotation only, the new star shape model "roche_v" uses equation (2.18) from Kopal "Dynamics of Close Binary Systems" (Springer, 1978) to calculate the volume enclosed by the equipotential surface. This makes very little difference unless the star is very distorted and is slower.
Made printing of parameters python3 compatible in J0113+31.py Added bin/ellc_emcee
Corrected bug in calculation of simplified reflection Corrected bug in calculation of flux-weighted radial velocity for stars with detailed reflection. Added check for flux_weighted radial velocity combined with simplified reflection in rv().
Avoided problems with NaNs returned in rare cases where the spot is on the limb but no tangent points are defined for the projected ellipse. Also removed stop for the case ntouch > 4. Corrected first line of doc string for rv() (Errors spotted by Bogumil Pilecki) Correct bug in ellc.rv() - calculation of rv for iobs=0 can cause crash
Avoided stop in ell_ell_intersect when ellipse intersection algorithm fails.
Added speed-up to ellc.rv() for flux_weight=False. Corrected bug in initialisation of star shapes/fluxes for eccentric orbits. Fixed crash in lc.py for some choices of limb darkening law.
Added test to catch rare case in module ellipse, subroutine ell_ell_intersect where test for tangent/intersection point fails because one of the mid-points between the intersection/tangent points is also a tangent point (or nearly so).
Added clean exit if polynomial root finding step fails in ell_ell_intersect, and raised warning for cases where results may be inaccurate due to root-polishing convergence problems.
Changed definition of third-light contribution to be more intuitive.
Version sent to A&A