structural changes

01_find_new_entries.R

@@ -4,6 +4,7 @@ library("tidyverse")
 library("xlsx")
 
 source("config/config.R")
+source("functions.R")
 
 #Übernimmt die Liste der zu durchsuchenden Excelfiles aus der Konfiguration
 file_list <- FIND_DISTINCTS_FILES

02_not_in_db.R

@@ -8,7 +8,7 @@ library("tidyverse")
 
 #Konfiguration einbinden
 source("config/config.R")
-source("sparql.R")
+source("functions.R")
 
 #Datenbank einlesen und doppelte entfernen
 inDB <- data.frame(get_current_geoera_lit_db()$citation)

03_find_URLs.R

@@ -4,36 +4,7 @@ library("writexl")
 
 #Konfiguration einbinden
 source("config/config.R")
-
+source("functions.R")
-## nicht meine Funktion - eingebunden von https://rdrr.io/cran/retractcheck/src/R/utils.R
-find_doi <- function (strings) {
-  regex <- '10\\.\\d{4,9}/[-._;()/:A-Z0-9]+'
-  doiLoc <- gregexpr(text = strings, pattern = regex, perl = TRUE, ignore.case = TRUE)
-  
-  i <- 1
-  res <- NULL
-  
-  # for each in the doiLoc list check whether match (!-1)
-  for ( i in 1:length(doiLoc) ) {
-    if ( doiLoc[[i]][1] != -1 ) {
-      for ( j in 1:length(doiLoc[[i]]) ) {
-        res <- c(res,
-                 substring(strings[i], doiLoc[[i]][j], doiLoc[[i]][j] + attr(doiLoc[[i]], 'match.length')[j] - 1))
-      }
-    }
-  }
-  
-  return(res)
-}
-
-
-## nicht meine Funktion - eingebunden von https://stackoverflow.com/questions/52911812/check-if-url-exists-in-r
-valid_url <- function(url_in,t=2){
-  con <- url(url_in)
-  check <- suppressWarnings(try(open.connection(con,open="rt",timeout=t),silent=T)[1])
-  suppressWarnings(try(close.connection(con),silent=T))
-  ifelse(is.null(check),TRUE,FALSE)
-}
 
 
 url_pattern <- "http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\\(\\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+"

04_find_and_replace_uri.R

@@ -8,7 +8,7 @@ library("tictoc")
 library("rlang")
 
 #Funktionen & Konfiguration einbinden
-source("sparql.R")
+source("functions.R")
 source("config/config.R")
 
 

functions.R

@@ -59,3 +59,32 @@ get_current_geoera_lit_db <- function() {
   
 }
 
+## nicht meine Funktion - eingebunden von https://rdrr.io/cran/retractcheck/src/R/utils.R
+find_doi <- function (strings) {
+  regex <- '10\\.\\d{4,9}/[-._;()/:A-Z0-9]+'
+  doiLoc <- gregexpr(text = strings, pattern = regex, perl = TRUE, ignore.case = TRUE)
+  
+  i <- 1
+  res <- NULL
+  
+  # for each in the doiLoc list check whether match (!-1)
+  for ( i in 1:length(doiLoc) ) {
+    if ( doiLoc[[i]][1] != -1 ) {
+      for ( j in 1:length(doiLoc[[i]]) ) {
+        res <- c(res,
+                 substring(strings[i], doiLoc[[i]][j], doiLoc[[i]][j] + attr(doiLoc[[i]], 'match.length')[j] - 1))
+      }
+    }
+  }
+  
+  return(res)
+}
+
+
+## nicht meine Funktion - eingebunden von https://stackoverflow.com/questions/52911812/check-if-url-exists-in-r
+valid_url <- function(url_in,t=2){
+  con <- url(url_in)
+  check <- suppressWarnings(try(open.connection(con,open="rt",timeout=t),silent=T)[1])
+  suppressWarnings(try(close.connection(con),silent=T))
+  ifelse(is.null(check),TRUE,FALSE)
+}