fix #48

NAMESPACE

@@ -12,6 +12,7 @@ export(dramaTail)
 export(figureStatistics)
 export(figurematrix)
 export(frequencytable)
+export(hamming)
 export(installCollectionData)
 export(installData)
 export(keyness)
@@ -33,6 +34,7 @@ export(qd.colors)
 export(rankFiguresByAppearance)
 export(rankFiguresByDramatisPersonae)
 export(report)
+export(scenicDifference)
 export(setup)
 export(tfidf)
 export(utteranceStatistics)

R/personnelExchange.R

@@ -0,0 +1,49 @@
+#' @name personnelExchange
+#' @title Measuring Personnel Exchange over Boundaries
+#' @aliases hamming
+#' @aliases scenicDifference
+#' @description There are multiple ways to quantify the number of characters that are 
+#' exchanged over a scene or act boundary. 
+#' @param mtext The segmented drama text
+#' @param variant For hamming(), variants are "Trilcke" (default), "NormalizedHamming", and "Hamming"
+#' @param norm For scenicDifference(), specifies the normalization constant
+#' @rdname personnelExchange
+#' @export
+#' @examples 
+#' data(rksp.0)
+#' dist_trilcke  <- hamming(rksp.0$mtext)
+#' dist_hamming  <- hamming(rksp.0$mtext, variant = "Hamming")
+#' dist_nhamming <- hamming(rksp.0$mtext, variant = "NormalizedHamming")
+hamming <- function(mtext, variant=c("Trilcke","Hamming","NormalizedHamming")) {
+  mtext <- na.omit(mtext)
+  numberOfFigures <- length(unique(mtext$Speaker.figure_surface))
+  scenes <- length(unique(mtext$begin.Scene))
+  pm <- configuration(mtext, by="Scene", onlyPresence = TRUE)
+  vec <- vector(mode="integer",length=scenes-1)
+  variant <- match.arg(variant)
+  for (i in 1:(ncol(pm$matrix)-1)) {
+    allFigures <- switch(variant,
+                         Trilcke=sum(pm$matrix[,i] | pm$matrix[,i+1]),
+                         Hamming=1,
+                         NormalizedHamming=numberOfFigures)
+    edits <- sum(xor(pm$matrix[,i],pm$matrix[,i+1]))
+    vec[i] <- edits/allFigures
+  }
+  vec
+}
+
+#' @rdname personnelExchange
+#' @export
+scenicDifference <- function(mtext, norm=length(unique(mtext$Speaker.figure_surface))) {
+  mtext <- na.omit(mtext)
+  numberOfFigures <- length(unique(mtext$Speaker.figure_surface))
+  scenes <- length(unique(mtext$begin.Scene))
+
+  pm <- configuration(mtext, by="Scene", onlyPresence = TRUE)
+  vec <- vector(mode="integer",length=scenes-1)
+  for (i in 1:(ncol(pm$matrix)-1)) {
+    same <- sum(pm$matrix[,i] & pm$matrix[,i+1])
+    vec[i] <- numberOfFigures - same
+  }
+  vec/norm
+}