Awesome-Mirrors/awesome-network-analysis
1
0
Fork 0
mirror of https://github.com/briatte/awesome-network-analysis.git synced 2024-10-01 01:06:13 -04:00
Code Issues Packages Projects Releases Wiki Activity

make check.r cumulative

Browse Source 
if check.log is present, checks only the URLs in that file
This commit is contained in:
François Briatte 2017-01-02 03:08:36 +01:00
parent a3155cdfc7
commit db6f343a9a
1 changed files with 51 additions and 34 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

47
check.r
View File

@ -1,24 +1,30 @@
library(httr) library(httr)
library(stringr) library(stringr)
u <- str_c( f <- "check.log"
if (!file.exists(f)) {
u <- str_c(
"https://raw.githubusercontent.com/", "https://raw.githubusercontent.com/",
"briatte/awesome-network-analysis/", "briatte/awesome-network-analysis/",
"master/README.md" "master/README.md"
) )
u <- GET(u) %>% cat("Source:", u, "\n")
u <- GET(u) %>%
content("text") %>% content("text") %>%
str_split("\\n") %>% # so as to find [foo]: bar links str_split("\\n") %>% # so as to find [foo]: bar links
unlist unlist
# total number of links # total number of links
t <- str_count(u, "http") %>% t <- str_count(u, "http") %>%
sum sum
cat(t, "URLs, ") cat(t, "URLs, ")
l <- c( l <- c(
# [foo](bar) # [foo](bar)
str_extract_all(u, "\\(http(.*?)\\)") %>% str_extract_all(u, "\\(http(.*?)\\)") %>%
lapply(str_replace_all, "^\\(|\\)$", "") %>% lapply(str_replace_all, "^\\(|\\)$", "") %>%
@ -27,16 +33,27 @@ l <- c(
str_extract_all(u, "^\\[(.*)\\]: (.*)") %>% str_extract_all(u, "^\\[(.*)\\]: (.*)") %>%
unlist %>% unlist %>%
str_replace("^\\[(.*)\\]: (.*)", "\\2") str_replace("^\\[(.*)\\]: (.*)", "\\2")
) )
stopifnot(length(l) == t) stopifnot(length(l) == t)
} else {
cat("Source:", f, "\n")
l <- readLines(f) %>%
str_subset("^http")
cat(length(l), "URLs, ")
}
l <- unique(l) %>% l <- unique(l) %>%
sort sort
cat(length(l), "unique\n") cat(length(l), "unique\n")
sink("check.log", append = FALSE) sink(f, append = FALSE)
cat(as.character(Sys.time()), ": checking", length(l), "URLs\n\n") cat(as.character(Sys.time()), ": checking", length(l), "URLs\n\n")
sink() sink()
@ -50,7 +67,7 @@ for (i in l) {
cat("X") cat("X")
sink("check.log", append = TRUE) sink(f, append = TRUE)
cat(i, "\nStatus code:", x, "\n\n") cat(i, "\nStatus code:", x, "\n\n")
sink() sink()
@ -58,9 +75,10 @@ for (i in l) {
cat("?") cat("?")
sink("check.log", append = TRUE) sink(f, append = TRUE)
cat(i, "\nFailed to access\n\n") cat(i, "\nFailed to access\n\n")
sink() sink()
Sys.sleep(5)
} else { } else {
@ -73,11 +91,10 @@ for (i in l) {
} }
sink("check.log", append = TRUE) sink(f, append = TRUE)
cat(as.character(Sys.time()), ": done.\n") cat(as.character(Sys.time()), ": done.\n")
sink() sink()
cat("Found", str_count(readLines("check.log"), "^http:") %>% cat("Found", str_count(readLines(f), "^http") %>%
sum, sum,
"problems.\n") "problems.\n")