function for rerooting neuron

[dokato]

getting foot in the door.

TODO:

• [x] "And it should give the option to provide either indices or ids. See nat::distal_to for ideas."
• [x] Should have a generic of the form reroot(x,...)
• [x] and methods for neuron and neuronlist with arguments idx and pointno for both
• [x] neuronlist method needs to check argument length carefully
• [x] neuron method needs to check that the ids/indices reference a valid node
• [x] would also suggest that the neuron method returns the original object
• [x] finally we will need a specialised catmaidneuron method that ensures that additional metadata (synaptic connectors etc) is copied over when the neuron is remade
• [x] also add an entry in _pkgdown.yml
• [x] update news.md

[dokato]

and pointno for both

Here I decided to be more flexible and rather than point belonging to neuron I provide an option to reroot to a node that is closest to point

[dokato]

@jefferis I'd appreciate advice on how to go around this point:

finally we will need a specialised catmaidneuron method that ensures that additional metadata

It was tempting to just use a function from catmaid:

#' @export
reroot.catmaidneuron<-function(object, ...) {
  n = NextMethod()
  copy_tags_connectors(n, object)
}

But this should go to catmaid then? Here it'd introduce a dependency loop.

[jefferis]

Thanks @dokato! Yes, that was exactly what I was expecting the method should look like;reroot.catmaidneuron should be in catmaid. In future we should move some of the support for neurons with connectivity from catmaid into nat. But that will be a medium-sized change, so we should probably plan that separately.

[dokato]

Thanks @jefferis for useful comments, I added support for matrices, node ids and removed pkgdown build. PTAL now.

[dokato]

BTW while on that I gave it a go and added reroot.catmaidneuron to https://github.com/natverse/rcatmaid

But for some reason I'm getting such error:

> conn=catmaid_login(server="https://l1em.catmaid.virtualflybrain.org")
> orns=read.neurons.catmaid("name:ORN right", .progress='text')
> class(orns[[1]])
[1] "catmaidneuron" "neuron"
> reroot.catmaidneuron(orns[[1]])
Error in NextMethod() : generic function not specified

I'm using nat version with the reroot changes ofkz. Maybe I miss something about generics, but I believe that I found some relevant example where the "next" method came from a different package - similar to our case. Any ideas what's wrong here?

[jefferis]

Maybe you need to import the generic from nat?

Sent from my iPhone

On 3 Jun 2021, at 12:15, dokato @.***> wrote:

 BTW while on that I gave it a go and added reroot.catmaidneuron to https://github.com/natverse/rcatmaid

But for some reason I'm getting such error:

conn=catmaid_login(server="https://l1em.catmaid.virtualflybrain.org") orns=read.neurons.catmaid("name:ORN right", .progress='text') class(orns[[1]]) [1] "catmaidneuron" "neuron" reroot.catmaidneuron(orns[[1]]) Error in NextMethod() : generic function not specified I'm using nat version with the reroot changes ofkz. Maybe I miss something about generics, but I believe that I found some relevant example where the "next" method came from a different package - similar to our case. Any ideas what's wrong here?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[jefferis]

Oh also you shouldn’t call the method directly. Use the generic ie plain reroot.

Sent from my iPhone

On 3 Jun 2021, at 12:15, dokato @.***> wrote:

 BTW while on that I gave it a go and added reroot.catmaidneuron to https://github.com/natverse/rcatmaid

But for some reason I'm getting such error:

conn=catmaid_login(server="https://l1em.catmaid.virtualflybrain.org") orns=read.neurons.catmaid("name:ORN right", .progress='text') class(orns[[1]]) [1] "catmaidneuron" "neuron" reroot.catmaidneuron(orns[[1]]) Error in NextMethod() : generic function not specified I'm using nat version with the reroot changes ofkz. Maybe I miss something about generics, but I believe that I found some relevant example where the "next" method came from a different package - similar to our case. Any ideas what's wrong here?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[dokato]

ah damn, of course, that was it :P I'll send PR soon.

[jefferis]

Actually @dokato the generic should have signature (x, ...) and the method should have ... in addition to the named arguments. Generics should pretty much always have ... so that you can handle any subsequent methods that need other arguments.

[coveralls]

Coverage decreased (-0.06%) to 81.383% when pulling ca00daaf02e7cb6bb663e329ef7eefd2e6142c7a on dokato:reroot into 760f19e2b52b03e286dbb0ba47ba63561c4014c7 on natverse:master.

[jefferis]

@dokato I added some details to the docs. I'll merge when the checks complete.