#daimonscala 19-2 "Apache access_log(combined) parser"

gistfile1.scala

object Main {

  import java.net._
  import java.util.Date

  case class Access(
    ipAddress: InetAddress,
    ident: String,
    user: String,
    time: Date,
    method: String,
    uri: URI,
    version: String,
    status: Int,
    bytes: Int,
    referrer: String,
    userAgent: String
  )

  import util.parsing.combinator._

  object AccessLogParser extends JavaTokenParsers {

    import java.text.SimpleDateFormat
    val timeFormat = new SimpleDateFormat("dd/MMM/yyyy:HH:mm:ss Z")

    override val whiteSpace = "[ \t]+".r

    def eol = opt('\r') <~ '\n'
    def q = "\""

    def lines = repsep(line | "", eol)

    def lineMatch = {
      ipAddress ~
      id ~
      user ~
      "[" ~ time ~ "]" ~
      q ~ method ~ uri ~ version ~ q ~
      status ~
      bytes ~
      q ~ referrer ~ q ~
      q ~ ua ~ q
    }


    def line = lineMatch ^^ {
      case (
        ip ~ 
        id ~ 
        user ~ 
        "[" ~ time ~ "]" ~ 
        _ ~ method ~ uri ~ ver ~ _ ~ 
        status ~ 
        bytes ~ 
        _ ~ referrer ~ _ ~ 
        _ ~ ua ~ _
      ) => {
        new Access(
          InetAddress.getByName(ip), 
          id, 
          user, 
          timeFormat.parse(time), 
          method, 
          new URI(uri), 
          ver, 
          status.toInt, 
          bytes match { case "-" => 0 case b => b.toInt }, 
          referrer, 
          ua
        )
      }
    }


    def notSpaceAtLeastOne = "[^\\s]+".r
    def notQuoteAtLeastOne = "[^\"]+".r

    def num = "\\d"
    def atLeastOne = "+"

    def ipAddress = ((num + "{1,3}" + "\\.") * 3 + num + "{1,3}").r
    def id = notSpaceAtLeastOne
    def user = notSpaceAtLeastOne

    def month = "[a-zA-Z]{3}".r
    def time = (num + "{2}" + "/" + month + "/" + num + "{4}" + (":" + num + "{2}") * 3 + " \\+" + num + "{4}").r

    def method = "[A-Z]+".r
    def uri = notSpaceAtLeastOne
    def version = "HTTP/1\\.\\d".r
    def status = (num + "{3}").r
    def bytes = "-" | (num + atLeastOne).r
    def referrer = notQuoteAtLeastOne
    def ua = notQuoteAtLeastOne

    def parse(json: String): ParseResult[Any] = parseAll(lines, json)

  }

  def main(args: Array[String]) {
    val accessLog = """66.249.69.220 - - [03/Oct/2011:01:22:54 +0900] "GET /blog/23/ HTTP/1.1" 200 22716 "-" "SAMSUNG-SGH-E250/1.0 Profile/MIDP-2.0 Configuration/CLDC-1.1 UP.Browser/6.2.3.3.c.1.101 (GUI) MMP/2.0 (compatible; Googlebot-Mobile/2.1; +http://www.google.com/bot.html)"
64.233.172.34 - - [16/Oct/2011:04:03:27 +0900] "GET /api/search/?format=atom&q=testtest HTTP/1.1" 200 20 "-" "Rome Client (http://tinyurl.com/64t5n) Ver: UNKNOWN AppEngine-Google; (+http://code.google.com/appengine; appid: xxxxx)"
64.233.172.36 - - [19/Oct/2011:05:18:52 +0900] "GET / HTTP/1.1" 304 - "http://twitter.com/" "Mozilla/5.0 (Windows NT 5.1) AppleWebKit/535.1 (KHTML, like Gecko) Chrome/14.0.835.202 Safari/535.1"
"""
    val result = AccessLogParser.parse(accessLog)
    println(result.get)
  }

}

// vim: set ts=4 sw=4 et: