# Hierarchical partitioning using the “glmm.hp” package in R version 4.4.0

# Analyze the relative importance of different variables for the yield of Ophiopogon japonicus
library(glmm.hp)
mydata <- read.csv("maidong_hp1.csv")
head(mydata)
m1 <- lm(yield~PR+SA+APA+UA+MYSH,mydata)
m2 <- glmm.hp(m1)
m2

# Analyze the relative importance of different variables for the drying  rate of Ophiopogon japonicus
library(glmm.hp)
mydata <- read.csv("maidong_hp2.csv")
head(mydata)
m1 <- lm(DR ~PR+SA+APA+UA+MYSH,mydata)
m2 <- glmm.hp(m1)
m2




# Partial least squares path modeling using the “plspm” package in R version 4.4.0

# Analyze the mechanisms underlying the interactions of different variables for the yield of Ophiopogon japonicus

library(plspm)
library(igraph)
f.tx <- list(
  yield ~ PR + SA + MYSH,
  SA ~ PR + MYSH,
  PR ~ MYSH
)

edge.mat <- do.call(
  rbind,
  lapply(
    f.tx, function(i) {
      from <-  all.vars(i)[-1]
      to <- all.vars(i)[1]
      cbind(from, to)
    }
  )
)
dag <- graph_from_edgelist(
  edge.mat,
  directed = TRUE
)
par(mar = c(0, 0, 0, 0), family = "Source Sans Pro", cex = 1)
igraph::plot.igraph(
  dag,
  layout = layout_in_circle,
  vertex.size = 20,
  vertex.color = "grey100",
  vertex.label.color = "green4",
  vertex.label.family = "Source Sans Pro",
  edge.color = "grey0",
  edge.arrow.mode = 2, # 2 is for forward arrows
  edge.arrow.size = 1,
  edge.arrow.width = 1
)
adj.mat0 <- as_adjacency_matrix(
  dag,
  type = "lower",
  sparse = FALSE
)
adj.mat0
g.tp <- igraph::topo_sort(dag)$name
g.tp
rd <- match(g.tp, rownames(adj.mat0))
cd <- match(g.tp, colnames(adj.mat0))
adj.mat1 <- t(adj.mat0[rd, cd])
adj.mat1
MYSH = c(0,0,0,0)
PR = c(1,0,0,0)
SA = c(1,1,0,0)
yield = c(1,1,1,0)
foot_path = rbind(MYSH,PR,SA,yield)
colnames(foot_path) = rownames(foot_path)
foot_path
adj.mat1 == foot_path
innerplot(foot_path)
foot_blocks=list(7,3,4,1)
foot_mod=rep("A",4)
mydata <- read.csv("maidong_plspm1.csv")
head(mydata)
foot_pls=plspm(mydata,foot_path,foot_blocks,
               scheme = "centroid",modes = foot_mod, 
               plscomp = c(1,1,1,1),tol = 0.0000001)
innerplot(foot_pls, colpos = 'red', colneg = 'blue', show.values = TRUE,
          lcol = 'gray20', box.lwd = 0)
outerplot(foot_pls, what = 'loadings', arr.width = 0.1, colpos = 'red', 
          colneg = 'blue', show.values = TRUE, lcol = 'black')
summary(foot_pls)


# Analyze the mechanisms underlying the interactions of different variables for the drying rate of Ophiopogon japonicus

library(plspm)
library(igraph)
f.tx <- list(
  DR ~ PR + APA + MYSH,
  APA ~ PR + MYSH,
  PR ~ MYSH
)

edge.mat <- do.call(
  rbind,
  lapply(
    f.tx, function(i) {
      from <-  all.vars(i)[-1]
      to <- all.vars(i)[1]
      cbind(from, to)
    }
  )
)
dag <- graph_from_edgelist(
  edge.mat,
  directed = TRUE
)
par(mar = c(0, 0, 0, 0), family = "Source Sans Pro", cex = 1)
igraph::plot.igraph(
  dag,
  layout = layout_in_circle,
  vertex.size = 20,
  vertex.color = "grey100",
  vertex.label.color = "green4",
  vertex.label.family = "Source Sans Pro",
  edge.color = "grey0",
  edge.arrow.mode = 2, # 2 is for forward arrows
  edge.arrow.size = 1,
  edge.arrow.width = 1
)
adj.mat0 <- as_adjacency_matrix(
  dag,
  type = "lower",
  sparse = FALSE
)
adj.mat0
g.tp <- igraph::topo_sort(dag)$name
g.tp
rd <- match(g.tp, rownames(adj.mat0))
cd <- match(g.tp, colnames(adj.mat0))
adj.mat1 <- t(adj.mat0[rd, cd])
adj.mat1
MYSH = c(0,0,0,0)
PR = c(1,0,0,0)
APA = c(1,1,0,0)
DR = c(1,1,1,0)
foot_path = rbind(MYSH,PR,APA,DR)
colnames(foot_path) = rownames(foot_path)
foot_path
adj.mat1 == foot_path
innerplot(foot_path)
foot_blocks=list(7,3,4,1)
foot_mod=rep("A",4)
mydata <- read.csv("maidong_plspm2.csv")
head(mydata)
foot_pls=plspm(mydata,foot_path,foot_blocks,
               scheme = "centroid",modes = foot_mod, 
               plscomp = c(1,1,1,1),tol = 0.0000001)
innerplot(foot_pls, colpos = 'red', colneg = 'blue', show.values = TRUE,
          lcol = 'gray20', box.lwd = 0)
outerplot(foot_pls, what = 'loadings', arr.width = 0.1, colpos = 'red', 
          colneg = 'blue', show.values = TRUE, lcol = 'black')
summary(foot_pls)