short_path.R

#!/usr/bin/env Rscript
require(argparse)
psr <- ArgumentParser(description="coauthor glue baseline")
psr$add_argument("ipt", help="input", nargs="+")
psr$add_argument("-o", dest="opt", help="output")
psr$add_argument("--id", help="id pairs order input")
args <- psr$parse_args()

library(igraph)
library(dplyr)
library(jsonlite)
library(stringr)
require(plyr)
library(rhdf5)
library(tidyr)

author<-read.csv(args$ipt) %>% select(-X)
names(author) <- c('name','org','id','seq')
fname <- str_replace_all(basename(args$ipt),pattern='.csv',replacement = '')
auname <- author %>% group_by(name) %>% dplyr::summarise(count = n()) %>% arrange(desc(count))
auname <- auname$name[1]

idp <- h5read(args$id, "id_pairs")

# 将所有的article形成一个矩阵
node_au <- as.character(unique(author$id))
author_1 <- author %>% group_by(name) %>% dplyr::summarise(count = n()) %>% arrange(desc(count)) %>% filter(count>1 & count < 4)
node_coau <- unique(author_1$name)[unique(author_1$name)!=auname]
node_all <- c(node_au,node_coau)

adjacency_co <- merge(node_coau,node_coau,ALL=T)

names(adjacency_co) <- c('node1','node2')
adjacency_co <- adjacency_co %>% mutate(node1 = as.character(node1),
                                  node2 = as.character(node2)) %>% 
        filter(node1<node2) %>% mutate(merge=0)

adjacency_au <- merge(node_au,node_coau,ALL=T)
names(adjacency_au) <- c('node1','node2')
adjacency_au$merge <- 0
adjacency <- rbind(adjacency_au,adjacency_co)
        
N <- length(unique(node_au))

for(j in 1:N){
        name_1 <- author$name[author$id==node_au[j]]
        name_2 <- name_1[name_1!=auname]
        adjacency$merge[adjacency$node1 %in% c(node_au[j],name_2) & adjacency$node2 %in% name_2] <- 1
        cat(j,'\n')
}

adjacency <- adjacency %>% mutate(node1=as.character(node1),node2=as.character(node2))


#提取出有效连边
adjacency_1 <- adjacency[adjacency$merge==1,]
e <- data.frame(from = adjacency_1$node1,
                to = adjacency_1$node2)
node <- unique(c(adjacency_1$node1,adjacency_1$node2,node_au))


# node_au[!node_au %in% node]

# 生成图
net <- graph_from_data_frame(e, directed=F, vertices=node)

# 变形成为similarity 
dist_final <- merge(node_au,node_au,all=TRUE) 
names(dist_final) <- c('node1','node2')
dist_final  <- dist_final %>% mutate(node1 = as.character(node1), node2 = as.character(node2)) %>% filter(node1 < node2)
node_0<-as.factor(node)
dist <- distances(net,v=as.numeric(node_0[node_0%in%dist_final$node1]),to=as.numeric(node_0[node_0%in%dist_final$node2]))

dist_final <- data.frame(dist)
dist_final <- dist_final[rownames(dist_final)%in% node_au,]
dist_final$node1 <- rownames(dist_final)

dist_final <- dist_final %>% gather(node2, value, -node1) %>% mutate(node2 = str_remove(node2,pattern = 'X')) %>%
    filter(node2 %in% node_au) %>% filter(node1 < node2)

names(dist_final) <- c('id1', 'id2', 'dist')
rst <- merge(idp, dist_final)
rst$coau_dist <- 2/rst$dist
rst$coau_dummy <- as.numeric(rst$coau_dist==1)
rst$dist <- NULL
rst$id1 <- NULL
rst$id2 <- NULL

# 输出数据
file.remove(args$opt)
h5createFile(args$opt)
h5write(rst, file=args$opt, "shortpath")