Duplicate entry in results table if max_ll is the same across two model runs

[grst]

I have a case where a certain gene appears twice in the results table. This is annoying, because in that case I can't merge it back into an AnnData object.

	FSV	M	g	l	max_delta	max_ll	max_mu_hat	max_s2_t_hat	model	n	s2_FSV	s2_logdelta	time	BIC	max_ll_null	LLR	pval	qval
311	2.06039e-09	4	ENSG00000117090	54	4.85165e+08	1719.84	0.0110949	2.31245e-11	SE	2068	0.0197839	3.37435e+15	0.00366592	-3409.13	1719.84	-0.000104498	1	1
312	2.04339e-09	4	ENSG00000117090	181.915	4.85165e+08	1719.84	0.0110949	2.31245e-11	SE	2068	0.0194589	3.37435e+15	0.00116491	-3409.13	1719.84	-0.000104498	1	1

I think I tracked it down to

SpatialDE/Python-module/SpatialDE/base.py

Line 312 in 77f9fa5

model_results = model_results[model_results.groupby(['g'])['max_ll'].transform(max) == model_results['max_ll']]

where the result from the model run with the max value for max_ll is chosen. In this case, the max_ll value is identical across two model runs, resulting in two values being chosen.

I'm unsure what the best solution is here. Just pick the first one?
The entries seem almost the same anyway, except for FSV and I values.