Merge remote-tracking branch 'origin/master'

.gitignore

@@ -22,3 +22,4 @@ $RECYCLE.BIN/
 # Operating System Files
 # =========================
 .Rproj.user
+README.Rmd

DESCRIPTION

@@ -7,11 +7,9 @@ Description: An R package for estimating affiliate species richness based on pow
 License: What license is it under?
 Encoding: UTF-8
 LazyData: true
-RoxygenNote: 6.1.0
+RoxygenNote: 6.1.1
 Depends: ggplot2,
     EnvStats
-Imports: EnvStats,
-    ggplot2,
-    stats,
+Imports: stats,
     graphics,
     nlstools

NAMESPACE

@@ -1,10 +1,8 @@
 # Generated by roxygen2: do not edit by hand
 
 export(binera)
-export(binera.50)
 export(copredict)
 export(copredict.ci)
-export(curve.50)
 export(curve.df)
 import(EnvStats)
 import(ggplot2)

R/binera.R

@@ -6,15 +6,16 @@
 #' @param assoc.df The raw dataset of associations (pairs of names in two columns, hosts and affiliates)
 #' @param iter Number of iterations required for each subsample
 #' @param plots Do you want plots? Maybe you do. The world is your oyster.
+#' @param subSample Fraction of host species to subsample (default=NULL, use all data)
 #'
 #' @export
 
 
-binera <- function(assoc.df, iter, plots=FALSE) {
+binera <- function(assoc.df, iter, plots=FALSE, subSample=NULL) {
 
   n.host <- n.par <- host <- pred <- 0 
 
-  c <- curve.df(assoc.df, iter)
+  c <- curve.df(assoc.df, iter, subSample=subSample)
   model1 <- stats::nls(n.par~b*n.host^z,start = list(b = 1, z = 0.5),data=c)
 
   if(plots==TRUE) {

R/copredict.R

@@ -1,12 +1,4 @@
-#' 
-#' 
-#' @importFrom stats nls
-#' @importFrom stats sd
-#' @importFrom stats qt
-#' @importFrom stats coef
-#' @importFrom graphics hist
-#' @importFrom graphics par
-#' 
+
 #' @title Extrapolating richness via curve-fitting
 #'
 #' @description
@@ -19,6 +11,13 @@
 #' @param plot (boolean; default is TRUE) plot results?
 #' 
 #' @details 
+#'
+#' @importFrom stats nls
+#' @importFrom stats sd
+#' @importFrom stats qt
+#' @importFrom stats coef
+#' @importFrom graphics hist
+#' @importFrom graphics par
 #' 
 #' @export
 

R/copredict50.R

@@ -20,14 +20,17 @@
 #' @param iter1 Set the number of times to fit a curve
 #' @param iter2 Set the number of points to subsample at each host richness within the generation of each curve
 #' @param plot (boolean; default is TRUE) plot results?
+#' @param subSample Fraction of host species to subsample (default=NULL, use all data)
+#'
 #'
 #' @details 
 #' @export
 
 
-copredict.ci <- function(assoc.df, n.indep, iter, plot=TRUE) {
+copredict.ci <- function(assoc.df, n.indep, iter, 
+  plot=TRUE, subSample=0.5) {
 
-  model <- binera.50(assoc.df, iter)
+  model <- binera(assoc.df, iter, subSample)
   q <- stats::coef(model)
   p.cis <- nlstools::confint2(model)
 

R/curvedf.R

@@ -1,35 +1,63 @@
 #' @title Subsampling of entire network
 #'
 #' @description
-#' This function iteratively samples the entire network and generates a data frame with sub-sample host and affilaite richness tallied together
+#' This function iteratively samples the entire network and generates 
+#' a data frame with sub-sample host and affilaite richness tallied together
 #'
-#' @param assoc.df The raw dataset of associations (pairs of names in two columns, hosts and affiliates)
-#' @param iter Number of iterations required for each subsample
+#' @param assoc.df The raw dataset of associations (pairs of names in 
+#'  two columns, hosts and affiliates)
 #'
+#' @param iter Number of iterations required for each subsample
+#' @param subSample Fraction of host species to subsample for richness estimation
 #' @export
 
+<<<<<<< HEAD
 curve.df <- function(assoc.df, iter, subSample=0){
+=======
+curve.df <- function(assoc.df, iter, subSample=NULL){
+
+  if(!is.null(subSample)){
+    hostlist <- unique(assoc.df[,1])
+    hostCut <- sample(hostlist, round(subSample*hostlist,0), replace = FALSE)
+    assoc.df <- assoc.df[assoc.df[,1] %in% hostCut,]
+  }
+>>>>>>> origin/master
 
   # placeholder: Tad add subsample at this level because you didn't
   u.hosts <- unique(assoc.df[,1])
   n.hosts <- length(u.hosts)
 
-  results.df <- data.frame(n.host=0,n.par=0)
-  results.df <- results.df[-1,]
+  oneSp <- function(x, i=iter){
+    df<- vapply(1:iter, FUN.VALUE=c(double(1), double(1)),
+      function(a){
+        sub.hosts <- sample(u.hosts, x)
+        sub.df <- assoc.df[assoc.df[,1] %in% sub.hosts,]
+        u.par <- unique(sub.df[,2])
+        n.par <- length(u.par)
+        return(c(x, n.par))
+    })
+    return(t(df))
+  }
+
+  results.df <- lapply(1:n.hosts, 
+    oneSp, i=iter)
+  ret <- do.call(rbind, results.df)
+  colnames(ret) <- c('n.host', 'n.par')
+  return(ret)
+}
+
+
+
 
-  for (i in 1:n.hosts) {
-    #for (i in 1:5) {
-    for (j in 1:iter) {
 
-      sub.hosts <- sample(u.hosts,i)
-      sub.df <- assoc.df[assoc.df[,1] %in% sub.hosts,]
-      u.par <- unique(sub.df[,2])
-      n.par <- length(u.par)
 
+<<<<<<< HEAD
       results.df[nrow(results.df)+1,] <- c(i, n.par)
     }
     #print(i)
   }
   results.df <- data.frame(results.df)
   return(results.df)
 }
+=======
+>>>>>>> origin/master

README.md

@@ -8,6 +8,8 @@ Author
 
 Colin J. Carlson ([email protected])
 
+(Contributions by Tad Dallas!)
+
 Installing the package
 ----------------------