diff --git a/global.R b/global.R
index 544d47e..1986f76 100644
--- a/global.R
+++ b/global.R
@@ -12,12 +12,37 @@ library(quanteda)
 library(pdftools)
 library(readtext)
 library(msgxtractr)
+library(transport)
 
 #Enable bookmarking via url
 #enableBookmarking(store = "server")
 
 #Custom functions
 
+#Jensen-Shannon divergence
+JensenShannon <- function(query, matrix){
+  p <- query[1,]
+  q <- t(matrix) # transpose matrix
+  m = (p + q)/2
+  KLDp = mapply(KLD, p, m)
+  KLDq = mapply(KLD, p, m)
+  jsd = sqrt(0.5*(KLDp + KLDq))
+  jsd[is.nan(jsd)] <- 0
+  return (jsd)
+}
+
+#Kullback Leibler divergence
+KLD <- function(a,b){
+  sum(a * log(a/b))
+}
+
+#Earth Movers distance
+EarthMoversDistance <- function(query, matrix){
+  matrix <- t(matrix)
+  emdismatrix = apply(matrix, 1,function(i) { wasserstein1d(query, i, p = 1, wa = NULL, wb = NULL)})
+  return(emdismatrix)
+}
+
 CiteListPubmed <- function(elinkxml){
   #Function to return data frame of citations for articles in entrez link xml document.
   #Input:
diff --git a/server.R b/server.R
index f0b97e3..458affe 100644
--- a/server.R
+++ b/server.R
@@ -2463,13 +2463,25 @@ if((file.exists(paste0(getwd(),"/ToPMine/topicalPhrases/win_run.bat")) == TRUE)
     topicvec <- c(temp$theta)
     names(topicvec) <- paste("Topic", c(1:length(topicvec)))
     
+    ##Distance Metrics
     #Find the cosine distance between search document and all existing documents by topic proportions
+    # jsdist <- JensenShannon(temp$theta, topicmodel$TopicModel$theta)
     cosdist <- proxy::dist(x = temp$theta, y = topicmodel$TopicModel$theta, method = "cosine")
+    # emdist <- EarthMoversDistance(temp$theta, t(topicmodel$TopicModel$theta))
+    # browser()
     
-    #Get document order from closest match to furthest
+    #Get document order from closest match to furthest using Jensen Shannon
+    # closedocs <- topicmodel$Metadata$PMID[order(jsdist)]
+    # distperc <- jsdist[order(jsdist)]
+    
+    #Get document order from closest match to furthest using Cosine Distance
     closedocs <- topicmodel$Metadata$PMID[order(cosdist)]
     distperc <- cosdist[order(cosdist)]
     
+    #Get document order from closest match to furthest using Earth Movers Distance
+    # closedocs <- topicmodel$Metadata$PMID[order(emdist)]
+    # distperc <- emdist[order(emdist)]
+    
     # #Old cosine distance code that calculates distance between all documents instead of just the required distances to the new document
     # #Find the cosine distance between search document and all existing documents by topic proportions
     # cosdist <- proxy::dist(rbind(temp$theta, topicmodel$TopicModel$theta), method = "cosine")