R/compareSyllables.R

In vrios/songspellR: Automates syllable detection and classification in bioacoustic analyses

#' # compareSyllables <- function(x,y) {
#' #' compares two sylables, returning the dtw distance
#' #' receives two selection tables and compares
#' library(seewave)
#' library(tuneR)
#' library(dtw)
#' songs = findSyllables(plotSyllables  = T,dir(here::here("inst/extdata"), full.names = TRUE,include.dirs = T), samplingRate = 44100)
#' syls = waveCutter(songs, write.specs = F)
#' sylsa=syls
#' # syls <- syls$wave.files[1:33]
#' ####syls2 <- lapply(syls, readWave)
#' syls2<-syls$waves[1:33]
#' envs <- lapply(syls2, env, msmooth = c(50, 0), plot = FALSE)
#'
#'
#' ## create a new entry in the registry with two aliases, to permit use of dtw as a distance metric
#' pr_DB$set_entry(FUN = dtwOmitNA, names = c("dtwOmitNA"))
#' # d<-dist(k, method = "dtwOmitNA")
#'
#' k <- pad.to.compare(envs) # to compare different lengths
#'
#'
#' d <- dist(k, method = "dtwOmitNA")
#' min(d)
#' plot(d)
#'
#' cc <- hclust(d, method = "ward.D2")
#' cc.labels= cutree(cc,k=6)
#' cc.labels
#' x11();plot(cc)
#'
#' library(dbscan)
#'
#' ###classify syllables using hdbscan
#' res2=hdbscan(d, minPts = 3)
#' hc=res2$hc
#' clus=res2$cluster
#' clus
#'
#' syls3=data.frame("syllable"=sylsa$selections$selec[1:33],"cluster"=clus, id = 1:length(syls))
#' plotSyllables(syls = sylsa$selections,labels = clus)
#'
#'