PERMIT SIR add prescription data

#2 files can be used but only 1 is needed. 
#Required: patient ID, code, date, instruction text field

library(stringr)
library(lubridate)

load("sir.datahfonly.Rdata")
temp = list.files(pattern="*1.csv")
for (i in 1:length(temp)) assign(temp[i], read.csv(temp[i]))
dialysis<-sir.data[sir.data$ReadCode %in% Dialysis1.csv$ReadCode,]
levels(as.factor(dialysis$PatientID)) #48 dialysis patients will be removed

sir.data<-sir.data[!sir.data$PatientID %in% Dialysis1.csv$PatientID,]

codes<-read.csv("drug-codes-READv2.csv",header=TRUE)
codes2<-read.csv("drug-codes-EMIS.csv",header=TRUE)
codesfile<-sir.data[sir.data$ReadCode %in% codes$CODE|sir.data$ReadCode %in% codes2$CODE,c(1,2,3,4,5,6)]

#remove punctuation this is not helpful here, 
#replace it with spaces so the regex
#recognises words still

x <-codesfile 
str_replace_all(x,"[[:punct:]]", " ")

gsub("[[:punct:]]", "", codesfile, except(".", "-") )
head(codesfile)

#cut down to prescriptions in or after 2007
codesfile$event.date<-as.Date(as.character(codesfile$EntryDate),format="%Y%m%d")
codesfile<-codesfile[as.numeric(year(strptime(codesfile$event.date, format="%Y-%m-%d")))>=2007,]
na.omit(codesfile)

write.csv(codesfile[,c(1,2,4:6)],file="codesh.txt", row.names = FALSE)


#PERL COMMANDS
cd research-events-medication-htn
npm install
node index.js -a resources\codesh.txt && sort resources\codesh.txt.done > resources\codesh.txt.done.sorted

#CHECK THE REGEX FILE FOR ANY DESCRIPTIONS THAT DID NOT PARSE, 
#YOU WILL MANUALLY NEED TO TEACH THESE TO THE REGEX
#ESPECIALLY IF YOUR MEDICATIONS ARE NOT ALL SOLID DOSE.