munchclient/main.go

package main

import (
	"bufio"
	"encoding/csv"
	"flag"
	"fmt"
	"io"
	"log"
	"net"
	"net/http"
	"os"
	"strings"
	"time"

	"github.com/BurntSushi/toml"
	"github.com/google/gopacket"
	"github.com/google/gopacket/layers"
	"github.com/google/gopacket/pcap"
	"github.com/nats-io/nats"
	"github.com/nats-io/nats/encoders/protobuf"

	"git.scraperwall.com/scw/data"
	"git.scraperwall.com/scw/ip"
)

var (
	doLiveCapture         = flag.Bool("live", false, "Capture data in real time from a given interface")
	iface                 = flag.String("interface", "eth0", "Interface to get packets from")
	snapshotLen           = flag.Int("snapshot-len", 8192, "Snapshot Length in Bytes")
	filter                = flag.String("filter", "tcp", "PCAP filter expression")
	promiscuous           = flag.Bool("promiscuous", false, "Switch interface into promiscuous mode?")
	natsURL               = flag.String("nats-url", "nats://127.0.0.1:4222", "The URL of the NATS server")
	natsQueue             = flag.String("nats-queue", "requests", "The NATS queue name")
	sleepFor              = flag.Duration("sleep", 0, "Sleep this long between sending data (only when replaying a file)")
	requestsFile          = flag.String("requests", "", "CSV file containing requests (IP and URL)")
	useXForwardedAsSource = flag.Bool("use-x-forwarded", false, "Use the IP address in X-Forwarded-For as source")
	configFile            = flag.String("config", "", "The location of the TOML config file")
	beQuiet               = flag.Bool("quiet", true, "Be quiet")
	doVersion             = flag.Bool("version", false, "Show version information")

	natsEC  *nats.EncodedConn
	count   uint64
	timeout = -1 * time.Second
	ipPriv  *ip.IP
	config  Config

	// Version contains the program Version, e.g. 1.0.1
	Version string

	// BuildDate contains the date and time at which the program was compiled
	BuildDate string
)

// Config contains the program configuration
type Config struct {
	Live                  bool
	Interface             string
	SnapshotLen           int
	Filter                string
	Promiscuous           bool
	NatsURL               string
	NatsQueue             string
	SleepFor              duration
	RequestsFile          string
	UseXForwardedAsSource bool
	Quiet                 bool
}

type duration struct {
	time.Duration
}

func (d *duration) UnmarshalText(text []byte) error {
	var err error
	d.Duration, err = time.ParseDuration(string(text))
	return err
}

func (c Config) print() {
	fmt.Printf("Live:                  %t\n", c.Live)
	fmt.Printf("Interface:             %s\n", c.Interface)
	fmt.Printf("SnapshotLen:           %d\n", c.SnapshotLen)
	fmt.Printf("Filter:                %s\n", c.Filter)
	fmt.Printf("Promiscuous:           %t\n", c.Promiscuous)
	fmt.Printf("NatsURL:               %s\n", c.NatsURL)
	fmt.Printf("NatsQueue:             %s\n", c.NatsQueue)
	fmt.Printf("SleepFor:              %s\n", c.SleepFor.String())
	fmt.Printf("RequestsFile:          %s\n", c.RequestsFile)
	fmt.Printf("UseXForwardedAsSource: %t\n", c.UseXForwardedAsSource)
	fmt.Printf("Quiet:                 %t\n", c.Quiet)
}

func init() {
	flag.Parse()

	nats.RegisterEncoder(protobuf.PROTOBUF_ENCODER, &protobuf.ProtobufEncoder{})
}

func main() {
	if *doVersion {
		version()
		os.Exit(0)
	}

	loadConfig()

	// Output how many requests per second were sent
	if !config.Quiet {
		go func(c *uint64) {
			for {
				fmt.Printf("%d requests per second\n", *c)
				*c = 0
				time.Sleep(time.Second)
			}
		}(&count)
	}

	// NATS
	//
	if config.NatsURL == "" {
		log.Fatal("No NATS URL specified (-nats-url)!")
	}

	natsConn, err := nats.Connect(config.NatsURL)
	if err != nil {
		log.Fatal(err)
	}

	natsEC, err = nats.NewEncodedConn(natsConn, protobuf.PROTOBUF_ENCODER)
	if err != nil {
		log.Fatalf("Encoded Connection: %v!\n", err)
	}

	// What should I do?
	if config.RequestsFile != "" {
		replayFile()
	} else if config.Live {
		fmt.Printf("live capture (%s, %s) to %s\n", config.Interface, config.Filter, config.NatsURL)
		liveCapture()
	}
}

func liveCapture() {
	ipPriv = ip.NewIP()

	// PCAP setup
	//
	handle, err := pcap.OpenLive(config.Interface, int32(config.SnapshotLen), config.Promiscuous, timeout)
	if err != nil {
		log.Fatal(err)
	}
	defer handle.Close()

	err = handle.SetBPFFilter(config.Filter)
	if err != nil {
		log.Fatal(err)
	}

	packetSource := gopacket.NewPacketSource(handle, handle.LinkType())

	for packet := range packetSource.Packets() {
		go processPacket(packet)
	}
}

// processPacket receives a raw packet from pcap, builds a Request item from it and sends it to the queue
func processPacket(packet gopacket.Packet) {
	count++

	ipLayer := packet.Layer(layers.LayerTypeIPv4)
	if ipLayer == nil {
		log.Println("No IPv4 Layer!")
		return
	}

	ip, _ := ipLayer.(*layers.IPv4)

	if ip.Protocol != layers.IPProtocolTCP {
		log.Println("No TCP Protocol!")
		return
	}

	ipSrc := ip.SrcIP.String()
	ipDst := ip.DstIP.String()

	tcpLayer := packet.Layer(layers.LayerTypeTCP)
	if tcpLayer == nil {
		return
	}

	tcp, _ := tcpLayer.(*layers.TCP)

	portSrc := tcp.SrcPort
	portDst := tcp.DstPort
	sequence := tcp.Seq

	applicationLayer := packet.ApplicationLayer()
	if applicationLayer == nil {
		return
	}

	data := data.Request{}

	reader := bufio.NewReader(strings.NewReader(string(applicationLayer.Payload())))

	req, err := http.ReadRequest(reader)
	if err != nil {
		return
	}

	data.IpSrc = ipSrc
	data.IpDst = ipDst
	data.PortSrc = uint32(portSrc)
	data.PortDst = uint32(portDst)
	data.TcpSeq = uint32(sequence)
	data.CreatedAt = time.Now().UnixNano()
	data.Url = req.URL.String()
	data.Method = req.Method
	data.Referer = req.Referer()
	data.Host = req.Host
	data.Protocol = req.Proto
	data.Origin = data.Host
	if _, ok := req.Header["Connection"]; ok {
		data.Connection = req.Header["Connection"][0]
	}
	if _, ok := req.Header["X-Forwarded-For"]; ok {
		data.XForwardedFor = req.Header["X-Forwarded-For"][0]
	}
	if _, ok := req.Header["X-Real-IP"]; ok {
		data.XRealIP = req.Header["X-Real-IP"][0]
	}
	if _, ok := req.Header["X-Requested-With"]; ok {
		data.XRequestedWith = req.Header["X-Requested-With"][0]
	}
	if _, ok := req.Header["Accept-Encoding"]; ok {
		data.AcceptEncoding = req.Header["Accept-Encoding"][0]
	}
	if _, ok := req.Header["Accept-Language"]; ok {
		data.AcceptLanguage = req.Header["Accept-Language"][0]
	}
	if _, ok := req.Header["User-Agent"]; ok {
		data.UserAgent = req.Header["User-Agent"][0]
	}
	if _, ok := req.Header["Accept"]; ok {
		data.Accept = req.Header["Accept"][0]
	}
	if _, ok := req.Header["Cookie"]; ok {
		data.Cookie = req.Header["Cookie"][0]
	}
	data.Source = data.IpSrc

	if config.UseXForwardedAsSource && data.XForwardedFor != "" {
		if strings.Contains(data.XForwardedFor, ",") {
			ips := strings.Split(data.XForwardedFor, ",")
			for i := len(ips) - 1; i >= 0; i-- {
				ipRaw := strings.TrimSpace(ips[i])
				ipAddr := net.ParseIP(ipRaw)
				if ipAddr != nil && !ipPriv.IsPrivate(ipAddr) {
					data.Source = ipRaw
					break
				}
			}
		} else {
			ipAddr := net.ParseIP(strings.TrimSpace(data.XForwardedFor))

			if !ipPriv.IsPrivate(ipAddr) {
				data.Source = data.XForwardedFor
			}
		}
	}

	if data.Source == data.IpSrc && data.XRealIP != "" {
		data.Source = data.XRealIP
	}

	natsEC.Publish(config.NatsQueue, &data)
}

// replayFile takes a file containing a list of requests (SourceIP Url) and queues the requests
// e.g.
// 157.55.39.229 /gross-gerau/12012260-beate-anstatt
// 103.232.100.98 /weinsheim-eifel/13729444-plus-warenhandelsges-mbh
func replayFile() {
	var req data.Request
	var startTs time.Time
	var endTs time.Time

	for {
		fh, err := os.Open(config.RequestsFile)
		if err != nil {
			log.Fatalf("Failed to open request file '%s': %s", config.RequestsFile, err)
		}

		c := csv.NewReader(fh)
		c.Comma = ' '

		for {
			if config.SleepFor.Duration > time.Nanosecond {
				startTs = time.Now()
			}

			r, err := c.Read()

			if err == io.EOF {
				break
			}

			if err != nil {
				log.Println(err)
				continue
			}

			req.IpSrc = r[0]
			req.Source = r[0]
			req.Url = r[1]
			req.UserAgent = "Munch/1.0"
			req.Host = "demo.scraperwall.com"
			req.CreatedAt = time.Now().UnixNano()

			natsEC.Publish(config.NatsQueue, &req)

			count++
			if config.SleepFor.Duration >= time.Nanosecond {
				endTs = time.Now()
				if endTs.Before(startTs.Add(config.SleepFor.Duration)) {
					time.Sleep(config.SleepFor.Duration - endTs.Sub(startTs))
				}
			}
		}
	}
}

func loadConfig() {

	// initialize with values from the command line / environment
	config.Live = *doLiveCapture
	config.Interface = *iface
	config.SnapshotLen = *snapshotLen
	config.Filter = *filter
	config.Promiscuous = *promiscuous
	config.NatsURL = *natsURL
	config.NatsQueue = *natsQueue
	config.SleepFor.Duration = *sleepFor
	config.RequestsFile = *requestsFile
	config.UseXForwardedAsSource = *useXForwardedAsSource
	config.Quiet = *beQuiet

	if *configFile == "" {
		return
	}

	_, err := os.Stat(*configFile)
	if err != nil {
		log.Printf("%s: %s\n", *configFile, err)
		return
	}

	if _, err = toml.DecodeFile(*configFile, &config); err != nil {
		log.Printf("%s: %s\n", *configFile, err)
	}

	if !config.Quiet {
		config.print()
	}
}

// version outputs build information
func version() {
	fmt.Printf("munchclient %s, built on %s\n", Version, BuildDate)
}