library(dismo)
library(rJava)
library(sp)
# data prepare
# set working directory that including environmental variables
setwd(choose.dir()) 
rasters<-list.files(pattern=".asc")
stackrasters <- stack(rasters)

# use extent() and crop() functions to get target area
ext<-extent(105,130,15,42)
china <- crop(stackrasters,ext)

#load longitude and latitude data
xy <- read.csv(file.choose(), header = 1) #Metasyhis gotoi occurrence data
newsp <- read.csv(file.choose(), header = 1) #M. varicollaris sp. n.occurrence data


# Keep 20% of the data points as test datasets
occ <- xy
fold <- kfold(occ, k=5)
occtest <- occ[fold == 1, ]
occtrain <- occ[fold != 1, ]

# MAXENT model fitting
me <- maxent(x=china, p=occtrain)
me

# predict to entire dataset
Pre <- predict(me, china) 
plot(Pre)
plot(Pre, xaxt = "n", yaxt="n", ylim=c(15, 41))
points(occ, pch=3, col="red", cex=0.7)
points(newsp, pch=4, col="blue", cex=0.7)
map(resolution =0, fill = T, col = 8, border=NA, add=T)
degAxis(side=1)   
degAxis(side=2)
legend("topleft", legend = c("Metasychis gotoi","Metasychis varicollaris sp. n."),pch = c(3,4), bty="n", text.font=3, text.col="black", col=c("red","blue"), cex=1.0)

# background data
bg <- randomPoints(china, 1000)

# model evaluating
e1 <- evaluate(me, p=occtest, a=bg, x=china)
e1
# plot AUC
plot(e1, "ROC")