When a companion is selected as the reference lens, Gulls searches backward for and appends the primary. When the primary is selected, its companion-search loop starts at catalog row zero and breaks on the first non-match, instead of starting at the primary/next row. Unless the selected primary happens to sit at the beginning of the catalog, it appears to find no companion at all. In recent simulations, among rows where a stellar companion actually appears in the overloaded Planet_ slot, about 95% have q>1.
The outputted lightcurves are not paired with the wrong simulated parameters. They accurately describe what Gulls actually simulated.
What is inconsistent:
- Lens_Is_Binary=1 still says the selected star belongs to a binary.
- Lens_combined_* metadata describes the combined catalog system.
When a companion is selected as the reference lens, Gulls searches backward for and appends the primary. When the primary is selected, its companion-search loop starts at catalog row zero and breaks on the first non-match, instead of starting at the primary/next row. Unless the selected primary happens to sit at the beginning of the catalog, it appears to find no companion at all. In recent simulations, among rows where a stellar companion actually appears in the overloaded
Planet_slot, about 95% have q>1.The outputted lightcurves are not paired with the wrong simulated parameters. They accurately describe what Gulls actually simulated.
What is inconsistent: