# ---
# title: "SARS-CoV2 Hospital Surveillance and Control System with Contact Tracing for Patients and Health Care Workers at a large reference hospital in Spain during the first wave. An observational descriptive study"
# author: "Anna Llupią  et al"
# date: "2020-Oct-6"
# output: html_document
# ---
# Published as supplementary material of DOI: 10.1002/HSR2.513
# 
# Figure 1
# ------------

library(ggplot2)
library(lubridate)
library(incidence)
library(epicontacts)
library(igraph)
load("results.RData") #Data not shared for privacy

#The network of cases and cases and contacts are stored in different epicontacts objects
#   longlist.cases is the list of cases in the network 'network.of.cases' of epicontacts
#   network.of.analysis is the network.of.cases without non-cases and a valid date of onset

Sys.setenv(LANGUAGE="en")
Sys.setlocale("LC_TIME", "English")

  daily.cuts<-as.Date(seq(first.day,last.day+1,by=1))
  #Part 1/3 of Figure 2: Weekly new cases by date of reporting
  fig2.cases.daily<-ggplot(data=longlist.cases,aes(Date.of.reporting,fill=Patient))+
    geom_histogram(breaks=daily.cuts)+
    scale_x_date(date_breaks = "1 week", date_minor_breaks = "1 week",
                 date_labels = "%d/%m",
                 limits = c(first.day,last.day))+
    scale_fill_discrete(name="Case Classification",                   breaks=c(TRUE, FALSE),                                                    labels=c("Patient", "Health Care Worker"))+
    xlab('')+ylab('New cases')+ theme(axis.title.x=element_blank(),
                                      axis.text.x=element_blank(),
                                      axis.ticks.x=element_blank())+theme_minimal()
  
  #Part 2/3 of Figure 2: Weekly cases which generated contacts
  fig2.caseswithcontacts.daily<- ggplot(data=subset(longlist,(longlist$case) & (longlist$ID %in% relations$Source.UID)),aes(Date.of.reporting,fill=Patient))+
    geom_histogram(breaks=daily.cuts)+
    scale_x_date(date_breaks = "1 week", date_minor_breaks = "1 week",
                 date_labels = "%d/%m",
                 limits = c(first.day,last.day))+
    scale_fill_discrete(name="Case Classification",
                        breaks=c(TRUE, FALSE),
                        labels=c("Patient", "Health Care Worker"))+
    #xlim(first.day,last.day)+
    xlab('')+ylab('Cases generating \n contacts')+theme(legend.position = "none")+theme_minimal()
  
  #Part 3/3 of Figure 2: Average contacts of new cases by date of reporting
  fig2.R0.contacts.daily<-ggplot(data=longlist.cases,aes(Date.of.reporting,number.of.contacts))+
    geom_point(stat = "summary_bin", fun = mean ,breaks=daily.cuts)+
    scale_x_date(date_breaks = "1 week", date_minor_breaks = "1 week",
                 date_labels = "%d/%m",
                 limits = c(first.day,last.day))+
    ylab("Average contacts \n of new cases")+
    xlab('Date of Reporting')+
    theme_minimal()
  
library(ggpubr)
  fig.daily<-ggarrange(fig2.cases.daily,
                 fig2.caseswithcontacts.daily,
                 fig2.R0.contacts.daily,
                 ncol=1,common.legend = TRUE)+theme_minimal()
fig.daily

# 
# Figure 3
# -------------------

#network.of.analysis is the network with valid dates of onset
network.of.analysis<-thin(subset(network.of.cases,node_attribute = list("included"=TRUE)))

#We compute the clusters using epicontacts
clust <- get_clusters(network.of.analysis)
nodes<-clust$linelist
edges<-clust$contacts
#We compute the positions for plotting nodes
nodes$cluster_position<-1
for(cluster in as.numeric(levels(nodes$cluster_member))){
  nodes$cluster_position[nodes$cluster_member==cluster]<-
    order(nodes$cluster_member[nodes$cluster_member==cluster])
}
cluster.positions<-as.data.frame(table(nodes$cluster_member))
cluster.positions$position<-cumsum(cluster.positions$Freq)
nodes$position.of.the.cluster<-cluster.positions$position[match(nodes$cluster_member,cluster.positions$Var1)]

#Define the positions of nodes and edges in timeline
nodes$x<-nodes$Date.of.onset
nodes$y<-nodes$cluster_position+(nodes$position.of.the.cluster)
edges$y_from<-nodes$y[match(edges$from,nodes$id)]
edges$y_to<-nodes$y[match(edges$to,nodes$id)]
edges$x_from<-nodes$x[match(edges$from,nodes$id)]+0.5
edges$x_to<-nodes$x[match(edges$to,nodes$id)]-0.5

#ggplot code for Figure 3
ggplot(data=nodes,aes(x,y))+geom_point(aes(colour =cluster_member),show.legend = FALSE)+
  geom_curve(data = edges,
               aes(x=x_from, xend=x_to, y=y_from, yend=y_to), size = 1,
               arrow=arrow(length = unit(0.1, "cm"),ends = "last"),
              curvature = 0.0,
               alpha=0.1
  )+  scale_x_date(date_breaks = "1 week", date_minor_breaks = "1 week",
                   date_labels = "%d/%m",limits = c(first.day,last.day))+
  ylab('')+
  scale_y_continuous(breaks= NULL)+
  xlab("Date of Onset")+
  theme(aspect.ratio = 10,legend.position = "none")+
  theme_minimal()