Dan Foreman-Mackey

Thanks for this, but the xlims are already set a little lower (look for the call to `_set_xlim`) so can you share a demo that doesn't work with the existing...

You can do that using the `fig` argument: ```python fig = corner.corner(dataset1, ..., color="C0", weights=np.ones(len(dataset1))/len(dataset1)) corner.corner(dataset2, ..., color="C1", fig=fig, weights=np.ones(len(dataset2))/len(dataset2)) ``` If you have different numbers of points you'll need...

This isn't really built into corner. Your best bet is probably to use matplotlib's rcParams.

This code doesn't set the tick font size anywhere. Make sure that you're using the correct rc params and this will work. If it doesn't, then it's probably a bug...

This would be great! Do you know if the arviz functions handle weighted samples properly? That would be crucial for this to be included. I'm probably not going to be...

This does sound like the right algorithm to me, and I'd be open to including something like this. We already have a custom `quantile` function https://github.com/dfm/corner.py/blob/c0c62645cba6909d08ff498f4da2bf5b0788dc09/src/corner/core.py#L388 because the numpy `percentile`...

This is not directly supported and I don't have plans for support since most of the code base is actually there to do sensible things with samples. Can you say...

This is an interesting question and it would certainly be possible to refactor the corner backend into a two step procedure: (1) compute the histograms, (2) do the plotting with...

That's not currently supported. That can be changed by editing [this section](https://github.com/dfm/corner.py/blob/master/corner/corner.py#L269) but it might actually be better to re-think the title API a bit if you really want that.

It would be a pretty non-trivial refactor of the code to support something like this and I feel like this might lead to a deeper discussion of how matplotlib itself...