package main

import (
	"bufio"
	"io"
	"os"
	"fmt"
	"regexp"
	"strings"
)

var (
	filePath string
	logReg = `TCP_HIT/([0-9][0-9][0-9]) [0-9]+ [A-Z]+ http://c13.adrise.tv/04C0BF/v2/sources/content-owners/[0-9a-z]+/([0-9]+)/`
	logMap = make(map[string]int)
)

func init() {
	filePath =os.Args[1]
}

func processLog(contents string) {
	re := regexp.MustCompile(logReg)
	matches := re.FindAllStringSubmatch(contents, -1)
	for _, m := range matches {
		key := m[1]+ "_" +m[2]
		count, ok := logMap [ key ]
		if (ok) {
			logMap [ key ] = count + 1
		} else {
			logMap [ key ] = 1
		}
	}
}

func readLine(fileName string, handler func(string)) error {
	f, err := os.Open(fileName)
	if err != nil {
		return err
	}
	buf := bufio.NewReader(f)
	for {
		line, err := buf.ReadString('\n')
                //TrimSpace 将删除 s 首尾连续的的空白字符
		line = strings.TrimSpace(line)
		handler(line)
		if err != nil {
			if err == io.EOF {
				return nil
			}
			return err
		}
	}
	return nil
}

func main() {
	//filePath = "/data/access.log"
	if err := readLine(filePath, processLog); err != nil {
		panic(err)
	}
	for log:= range logMap {
		fmt.Println(log, "appear count: ", logMap [log])
	}
}