WIP: added withMeta parameter to nlapply#501
Conversation
Codecov Report
@@ Coverage Diff @@
## master #501 +/- ##
==========================================
+ Coverage 75.73% 75.75% +0.01%
==========================================
Files 48 48
Lines 5906 5910 +4
==========================================
+ Hits 4473 4477 +4
Misses 1433 1433
Continue to review full report at Codecov.
|
| #' error. The default value (\code{NA}) will result in nlapply stopping with | ||
| #' an error message the moment there is an error. For other values, see | ||
| #' details. | ||
| #' @param withMeta whether or not to include the metadata from the neuronlist |
There was a problem hiding this comment.
I don't think this is 100% clear yet. I think you need to add something like "in the body of each object in the output list"
There was a problem hiding this comment.
Is the intention that the metadata dataframe should be removed from the output? It is. Maybe this should be signalled.
| if (withMeta && ncol(X[,]) > 0) { | ||
| meta_list <- split(X[,], seq(nrow(X[,]))) | ||
| Xlist = mapply(append, X, meta_list, SIMPLIFY = FALSE) | ||
| X = as.neuronlist(lapply(Xlist, as.neuron)) |
There was a problem hiding this comment.
This doesn't work if X contains dotprops rather than neuron objects. I think you need to set the class based on the class of each element of X before appending. I guess you have to use append in case your metadata overwrites one of the existing fields in the elements of X? NB I have tried to avoid assuming that every element of a neuronlist has the same class.
| out2 = nlapply(kcs13, function(x) x, withMeta = T) | ||
| expect_true( | ||
| length(attributes(out1[[2]])$names) < length(attributes(out2[[2]])$names) | ||
| ) |
There was a problem hiding this comment.
I'm not sure this was the intention:
> out1
'neuronlist' containing 3 'dotprops' objects and 'data.frame' with 14 vars [75.9 kB]
> out2
'neuronlist' containing 3 'list' objects and 'data.frame' with 0 vars [94.5 kB]
dokato
changed the title
|
Thanks @jefferis , maybe it requires more thinking on design indeed. The idea was to have access within the I know it's possible with subset, but then you need to stitch neuronlists together, which messes the order etc. I found out a couple of times that it'd be handy. |
2 participants
Details in #500