#read in a dynamic edgelist
edges <- read.csv(file.choose(),stringsAsFactors = F)

edges$V1 <- as.character(edges$V1)
edges$V2 <- as.character(edges$V2)

Net.model <- function(seed.val=round(runif(1,1,10000)),beta=.5,inc=3){  
   set.seed(seed.val)
   #t <- seq.Date(from=min(e$Date),to=max(e$Date),by=1)
   t <- seq(from=min(edges$Date),max(edges$Date),by=1)
   t <- t[1:56]
   
   allindivs <- unique(c(edges$V1,edges$V2))
   herds <- data.frame(name=as.character(allindivs),state=0)
  
   herds$state <- sample(c(rep(0,nrow(herds)-1),1),nrow(herds))
   herds$time.I <- NA
   herds$time.I <- ifelse(herds$state==1,0,NA)  #for recording time infected.  NA for all except initially infected farm
   
   #13-Create a dataframe called track that will keep track of the time that herds are infected
   #file for keeping track of S and I numbers
   track <- data.frame(time=min(t),S=c(nrow(herds)-1),I=c(1))
   
   #This (above) is making a data frame (matrix) that records the time and number of susceptible and infected herds by time. 
   S <- nrow(herds)-1   #object to keep track of how numbers are changing each timestep
   I <- 1                #object to keep track of how numbers are changing each timestep-THIS NEED TO BE CHANGED TO INCORPORATE THE IDENTIFIED (Ic) #AND UNIDENTIFIED INDIVIDUALS (Is)      
   
   rbinom.v <- function(x){rbinom(1,1,x)}#This function vectorizes the rbinom function so that you can run it on columns of data.  a= # of draws (can be set to one), b= probability.  Returns the sum number of 1s drawn in sample of size a
   #rbinom(observations, size, probability); x=prob that farm becomes infected; observations= ;size=
   timestep <- length(t)
   for(i in 1:timestep){
      #create new row for next timestep with same values as previous (will be updated later if transmission occurs)
      row <- data.frame(time=t[i],S=S,I=I)
      track <- rbind(track,row)
      print(i)
      
      
      ###MOVEMENTS
      M.o <- nrow(edges[edges$Date==t[i],]) #observed number of movements on time t
      
      if(M.o>0){  #skips to below if no movements occur
         edges.t <- subset(edges,Date==t[i]) #get list of movements for this timestep
         
         #bring in state of senders and receivers; State 1 is origin and State 2 is destination
         edges.t$state1 <- herds$state[match(edges.t$V1,herds$name)]#a column named state 1 that matches origin and farm name
         edges.t$state2 <- herds$state[match(edges.t$V2,herds$name)] # a column that matched the destination of edge list with farm name
         
         edges.t$time.I1 <- herds$time.I[match(edges.t$V1,herds$name)]#a column named state 1 that matches origin and farm name
         edges.t$time.I2 <- herds$time.I[match(edges.t$V2,herds$name)] # a column that matched the destination of edge list with farm name
         edges.t <- subset(edges.t, state1 != state2)
         
         #now, reorganize so that infected herd is always listed first
         if(nrow(edges.t)>0){
            for(j in 1:nrow(edges.t)){
            if(edges.t$state1[j]==0){
               row <- data.frame(edges.t$V2[j],edges.t$V1[j],edges.t$Date[j],edges.t$state2[j],edges.t$state1[j],edges.t$time.I2[j],edges.t$time.I1[j])
               names(row)<-names(edges.t)
               edges.t$V1[j] <- as.character(row$V1)
               edges.t$V2[j] <- as.character(row$V2)
               edges.t[j,4:7]<- row[4:7]
            }
         }
         edges.t <- subset(edges.t , time.I1 < i - inc)
         }  
         if(sum(edges.t$state1,na.rm=T)>0){  #skip to below if no infected herds moved any animals
            #filter edgelist so only includes movements from I --> S herds
            edges.f <- subset(edges.t,state1==1 & state2==0)
            
            if(nrow(edges.f)>0){ #skip to below if no I-->S movements
               
               
               #calculate state transitions
               edges.f$PROB.I <- 1-beta
               trans <- data.frame(PROB.I.cum =tapply(edges.f$PROB.I,as.character(edges.f$V2),prod))#multiply all prob together. Probability that none of animals from any of the connected herds are I
               trans$name <- row.names(trans)
               trans$m <- 1-trans$PROB.I.cum #Probability that at least one E or I was moved
               #get transition
               trans$trans <- sapply(trans$m,rbinom.v)
               
               #if transmission occurs, update state in herds
               trans.names <- trans$name[trans$trans==1] #newly infected IDs
               herds$state <- ifelse(herds$name %in% trans.names,1,herds$state)
         
               herds$time.I <- ifelse(herds$name %in% trans.names,i,herds$time.I)
               
               if(nrow(subset(trans,trans>0))>0){#update farm counts
                  S <- S - sum(trans$trans) #decrease S count by one farm
                  I <- I + sum(trans$trans) #increase I count by one farm
               } 
            }}
      }
         
      #Update counts of tracker   
      track$S[i+1] <- S
      track$I[i+1] <- I   
   }
   
   return(list(herds,track))
}