gistfile1.go

package main

import (
    "fmt"
    "strings"
    "strconv"
    "regexp"
    "io/ioutil"
    "net/http"
)

// define web web address and pages.
const (
    web_address string = "http://kfxx.info/mm.php"
    root_addr string = "http://kfxx.info/"
    save_location string = "your storage pic address"
)

func generate_url(index int) (url string) {
    url = web_address + "?page=" + strconv.Itoa(index)
    return url
}

func findAllPics(content string) (pics []string){
    re, _ := regexp.Compile(`assets/mm/(.*).jpg`)
    matchs := re.FindAllStringSubmatch(content, -1)
    for _, item := range matchs {
        pics = append(pics, string(item[0]));
    }
    return pics
}

func savePic(pic string){
    url := root_addr + pic
    resp, err := http.Get(url)
    if err != nil {
        return
    }
    body, _ := ioutil.ReadAll(resp.Body)
    defer resp.Body.Close()
    file_name := save_location + getFileName(pic)
    fmt.Println(file_name)
    ioutil.WriteFile(file_name, [] byte(string(body)), 0644)
}

func getFileName(pic string) string{
    tmp := strings.Split(pic, "/")
    fmt.Println(tmp)
    return string(tmp[len(tmp) - 1])
}

func fetchPage() {
    for i := 1; i <= 5; i++ {
        url := generate_url(i)
        resp, err := http.Get(url)
        if err != nil {
            fmt.Println("can not fetch the web address", url)
            return
        }
        body, _ := ioutil.ReadAll(resp.Body)
        defer resp.Body.Close()
        // do something here.
        fmt.Printf("url:%s can access\n", url)
        pics := findAllPics(string(body))
        for _, item := range pics {
            go savePic(item)
        }
    }
}


func main() {
    fmt.Println("Starting fetch web from: ", web_address)
    fetchPage()
}

用土办法写的一个爬妹纸的图
尝试用Golang写东西，才发现变量定义和字符串的引号问题是一个大问题