Remove requirement that input samples for simplify have the IS_SAMPLE flag about tskit HOT 5 CLOSED

I'm good with this, although it does seem like if someone tries to simplify to a non-sample, it's most likely an error. Related: previously, we have said that a "sample" must be a node for which we have parents for across the entire genome. Should we no longer say this, and only say that is a focal node in analyses?

from tskit.

I'm good with this, although it does seem like if someone tries to simplify to a non-sample, it's most likely an error.

I think we should mark it as a sample in the output though, regardless of its sample status in the input. Right now, I'm layering up a tree sequence by sequentially adding new stuff. Things that were samples before are not now, and things that weren't samples become samples; I want to simplify WRT to the current samples, and also WRT to everything that is not the current samples. It's a powerful operation, there's a million different things you might want to do with it, and having to edit the tree sequence first just to set the sample flag on the node you're interested in is really annoying.

Anyway, this is a long-winded way of saying, "I don't think it's most likely an error"!

Related: previously, we have said that a "sample" must be a node for which we have parents for across the entire genome. Should we no longer say this, and only say that is a focal node in analyses?

Yes, you're right. I think we should drop this clause --- it's unnecessary and doesn't add anything.

from tskit.

I think we should mark it as a sample in the output though, regardless of its sample status in the input.

Yes.

Anyway, this is a long-winded way of saying, "I don't think it's most likely an error"!

Sounds good. I meant "for everyone else but you, currently, it is most likely an error." =)

from tskit.

Sounds good. I meant "for everyone else but you, currently, it is most likely an error." =)

Haha, true enough. Hopefully more people than me will be doing these sort of tree sequence manipulation soon...

from tskit.

@gtsambos has run into this annoyance, here

from tskit.