Last active
December 16, 2016 09:21
-
-
Save prise6/b873f4934083dd9e70e4977da711be0d to your computer and use it in GitHub Desktop.
IPV4, IPV6, ASN from RIRs by date
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# ------------------------------------------------ | |
# | |
# GET IP RANGE BY DATE | |
# with RIR informations | |
# | |
# ------------------------------------------------ | |
library(data.table) | |
library(iptools) | |
library(plyr) | |
read.dirs = function(url, output.tmp = "Datas/tmp", compression = NULL, drop = NULL, skip = 0) { | |
colClasses = c("character", "character", "character", "character", "integer", "integer", "character") | |
col.names = c("REGISTRE", "CC", "TYPE", "START", "VALUE", "DATE", "STATUS") | |
if(!is.null(drop)) | |
colClasses = c(colClasses, "character") | |
if(!is.null(compression)){ | |
download.file( | |
url = url, | |
destfile = output.tmp, | |
mode = "wb", | |
quiet = TRUE | |
) | |
if (compression == "bz") | |
zz = bzfile(output.tmp) | |
else if (compression == "gz") | |
zz = gzfile(output.tmp) | |
else | |
stop("Compression non valide") | |
d = data.table(read.table( | |
file = zz, | |
sep = "|", | |
skip = skip, | |
encoding = "UTF-8", | |
stringsAsFactors = F, | |
colClasses = colClasses, | |
col.names = col.names, | |
na.strings = c("", "NA", "NULL"), | |
header = F | |
)) | |
unlink(output.tmp) | |
} else { | |
d = data.table::fread( | |
input = url, | |
sep = "|", | |
header = F, | |
skip = skip, | |
encoding = "UTF-8", | |
stringsAsFactors = F, | |
drop = drop, | |
colClasses = colClasses, | |
col.names = col.names, | |
na.strings = c("", "NA", "NULL"), | |
showProgress = F | |
) | |
} | |
return(unique(d)) | |
} | |
get.ranges = function(rir, date = NULL) { | |
url = getUrlFromRIR(rir, date = date) | |
options = switch( | |
rir, | |
ARIN = list(compression = NULL, drop = 8, skip = 4), | |
RIPE = list(compression = "bz", drop = NULL, skip = 4), | |
AFRINIC = list(compression = NULL, drop = NULL, skip = 4), | |
APNIC = list(compression = "gz", drop = NULL, skip = 31), | |
LACNIC = list(compression = NULL, drop = NULL, skip = 4)) | |
d = read.dirs(url, output.tmp = "Datas/tmp", options$compression, options$drop, options$skip) | |
d = d[TYPE == "ipv4"] | |
d[, `:=` ( | |
TYPE = NULL, | |
RANGE = paste0(START, "/", 32-log2(VALUE)), | |
START = NULL, | |
VALUE = NULL, | |
STATUS = NULL, | |
DATE = as.IDate(as.character(DATE), format = "%Y%m%d") | |
)] | |
} | |
getUrlFromRIR = function(rir, date = NULL) { | |
if(date == "today" || is.null(date)) | |
date = Sys.Date() | |
date = as.IDate(date) | |
actual = FALSE | |
if(format(date, "%Y") == format(Sys.Date(), "%Y")) | |
actual = TRUE | |
if(rir == "ARIN") { | |
if(isTRUE(actual)) | |
url = paste0("ftp://ftp.arin.net/pub/stats/arin/", "delegated-arin-extended-", format(date, "%Y%m%d")) | |
else | |
url = paste0("ftp://ftp.arin.net/pub/stats/arin/archive/", format(date, "%Y"), "/delegated-arin-extended-", format(date, "%Y%m%d")) | |
} else if ( rir == "RIPE") { | |
url = paste0("ftp://ftp.ripe.net/pub/stats/ripencc/", format(date, "%Y"), "/delegated-ripencc-", format(date, "%Y%m%d"), ".bz2") | |
} else if ( rir == "AFRINIC" ) { | |
url = paste0("ftp://ftp.afrinic.net/pub/stats/afrinic/", format(date, "%Y"), "/delegated-afrinic-", format(date, "%Y%m%d")) | |
} else if ( rir == "APNIC" ) { | |
url = paste0("ftp://ftp.apnic.net/pub/stats/apnic/", format(date, "%Y"), "/delegated-apnic-", format(date, "%Y%m%d"), ".gz") | |
} else if ( rir == "LACNIC" ) { | |
url = paste0("ftp://ftp.lacnic.net/pub/stats/lacnic/", "delegated-lacnic-", format(date, "%Y%m%d")) | |
} else | |
stop(sprintf("%s RIR doesn't exit", rir)) | |
return(url) | |
} | |
formatIPV4 = function(ipv4) { | |
ipv4[, c("MIN_NUMERIC", "MAX_NUMERIC") := as.list(iptools::range_boundaries(RANGE)[, 3:4])][ | |
, `:=` (MIN_NUMERIC = bit64::as.integer64(MIN_NUMERIC), MAX_NUMERIC = bit64::as.integer64(MAX_NUMERIC))] | |
setnames(ipv4, old = colnames(ipv4), new = paste0("IPV4_", colnames(ipv4))) | |
setkeyv(ipv4, c("IPV4_DATE_ALIM", "IPV4_RANGE", "IPV4_CC", "IPV4_DATE")) | |
setorder(ipv4, IPV4_RANGE, -IPV4_DATE) | |
ipv4 | |
} | |
findIPCountyCode = function(x, y, x.field, y.start, y.end) { | |
tmp.x = x[, list(IP = unique(get(x.field)))][!is.na(IP)][, X := bit64::as.integer64(iptools::ip_to_numeric(IP))][, dummy := X] | |
setnames(tmp.x, old = "IP", new = x.field) | |
setkeyv(tmp.x, c("X", "dummy")) | |
setkeyv(y, c(y.start, y.end)) | |
d = data.table::foverlaps(tmp.x, y, nomatch = 0L, mult = "first")[, `:=` (dummy = NULL, X = NULL, IPV4_REGISTRE = NULL, | |
IPV4_DATE = NULL, IPV4_RANGE = NULL, | |
IPV4_DATE_ALIM = NULL, IPV4_MIN_NUMERIC = NULL, IPV4_MAX_NUMERIC = NULL)] | |
d[, IPV4_CC := factor(IPV4_CC)] | |
d = merge(x, d, all.x = T, by = x.field) | |
return(d) | |
} | |
# utilisation | |
system.time({ | |
arin = get.ranges(rir = "ARIN", date = "2016-11-01") | |
ripe = get.ranges(rir = "RIPE", date = "2016-11-01") | |
afrinic = get.ranges(rir = "AFRINIC", date = "2015-01-02") | |
apnic = get.ranges(rir = "APNIC", date = "2015-01-02") | |
lacnic = get.ranges(rir = "LACNIC", date = "2015-01-02") | |
}) | |
system.time({ | |
dates = c("2014-01-10", "2015-01-10", "2016-01-10") | |
ipv4 = rbindlist(lapply(dates, function(date){ | |
d = rbindlist(lapply(list("ARIN", "RIPE", "AFRINIC", "APNIC", "LACNIC"), get.ranges, date = date)) | |
d[, DATE_ALIM := date] | |
})) | |
ipv4 = formatIPV4(ipv4) | |
}) | |
connexions = findIPCountyCode(connexions, ipv4, x.field = "ADR_IP", y.start = "IPV4_MIN_NUMERIC", y.end = "IPV4_MAX_NUMERIC") | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment