Closed mikejiang closed 10 years ago
By installing both versions of fda
, stepping into warpSet
and comparing the output from each fda
function call, here we found the source of problem.
The fda2.4.0::fdPar
returns a single-column
matrix
Browse[1]> WfdPar$fd$coefs
[,1]
bspl4.1 0
bspl4.2 0
bspl4.3 0
bspl4.4 0
bspl4.5 0
bspl4.6 0
bspl4.7 0
bspl4.8 0
bspl4.9 0
bspl4.10 0
bspl4.11 0
bspl4.12 0
bspl4.13 0
yet the fda2.4.3::fdPar
returns a n by n
squared matrix
WfdPar$fd$coefs
[,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11] [,12] [,13]
bspl4.1 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.2 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.3 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.4 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.5 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.6 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.7 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.8 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.9 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.10 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.11 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.12 0 0 0 0 0 0 0 0 0 0 0 0 0
bspl4.13 0 0 0 0 0 0 0 0 0 0 0 0 0
This breaks fda::landmarkreg
call later on.
diff
on the code base of fda package
further confirms this:
diff ../otherPkgs/fda_2.4.3/R/fdPar.R ../otherPkgs/fda_2.4.0/R/fdPar.R
62c62
<
---
> coefs <- matrix(0,nbasis-length(dropind),1)
For now we will simply inline
the fdPar
function from fda2.4.0
before its author addresses this issue.
flowStats
failed to build onBioC
lately due to the upgrade offda
package. Here is the call stack: