用luasocket读取双色球中奖号码

其实是蛮简单的一个功能,就是读取这个页面:http://61.129.89.226/fcopen/cp_kjgg_dfw.jsp?lottery_type=ssq&lottery_issue=2012138, 把某一个期的中奖号码抽取出来。

【luarocks安装library】

首先,如果是想用某个库,真没必要自己wget再make, install啥的,太烦。lua社区提供了一个非常好用的工具,用法和apt(advanced package tools)很类似:

apt-cache search xxx
sudo apt-get install xxx

luarocks search xxx
luarocks install xxx

这里需要luasocket,很简单:

$ luarocks search socket

Search results:
===============

Rockspecs and source rocks:
---------------------------

luasocket
2.0.2-5 (rockspec) - http://luarocks.org/repositories/rocks
2.0.2-5 (src) - http://luarocks.org/repositories/rocks
2.0.2-4 (rockspec) - http://luarocks.org/repositories/rocks
。。。。。。

 

$ luarocks install luasocket
Installing http://luarocks.org/repositories/rocks/luasocket-2.0.2-5.src.rock...
Archive: /tmp/luarocks_luarocks-rock-luasocket-2.0.2-5-2574/luasocket-2.0.2-5.src.rock
。。。。。。
cd src; cp mime.so.1.0.2 /home/baiyanh/.luarocks/lib/luarocks/rocks/luasocket/2.0.2-5/lib/mime/core.so
Updating manifest for /home/baiyanh/.luarocks/lib/luarocks/rocks

 

【读取页面】

然后用luasocket发送http get request,来获取web页面:

issuenum = arg[1]
if not issuenum then
        error "please provide the lottery issue num!"
end


socket = require "socket"

host = '61.129.89.226'
port = 80
fileformat = '/fcopen/cp_kjgg_dfw.jsp?lottery_type=ssq&lottery_issue=%s'
starting = '开奖结果'
numpattern = '>%d%d<'

function getlotterywinner(issuenum)
        local client = assert(socket.connect(host, port))
        client:send('GET ' .. string.format(fileformat, issuenum)  .. " HTTP/1.0\r\n\r\n")
        local line = client:receive('*l')
        local start = false
        local winner = {}
        while line do
                if line:find(starting) then
                        start = true
                end

                if start then
                        local num = line:match(numpattern)
                        if num then
                                table.insert(winner, num:sub(2, -2))
                        end
                end
                if #winner == 7 then break end
                line = client:receive('*l')
        end
        client:close()
        return winner
end

--http://61.129.89.226/fcopen/cp_kjgg_dfw.jsp?lottery_type=ssq&lottery_issue=2012138
winner = getlotterywinner(issuenum)
for _, v in ipairs(winner) do
        io.write(v .. "  ")
end
io.write("\n")

 

但遗憾的是,上述代码无法正确输出结果,事实上,上面“开奖结果”根本无法匹配 - 为什么?

【解决编码问题】

查看双色球网页的source,可以看到这么一行:

<meta http-equiv="Content-Type" content="text/html; charset=gb2312">

该网页的编码是gb2312, 而我写在lua代码中的“开奖结果”,是和这个lua脚本采用同样的编码:utf8,所以无法匹配(他们在内存中的表示方式不同,匹配自然失败) - 知道了原因就好办了,只要把他们转成同样的编码即可:

  • 转码的方向 - 自然是把“开奖结果”转成gb2312最省钱
  • 用什么库 - lua提供了对iconv库的wrapper,叫lua-iconv

好,下载:

$ luarocks install lua-iconv
Installing http://luarocks.org/repositories/rocks/lua-iconv-7-1.src.rock...
Archive: /tmp/luarocks_luarocks-rock-lua-iconv-7-1-2223/lua-iconv-7-1.src.rock
inflating: lua-iconv-7-1.rockspec
extracting: lua-iconv-7.tar.gz
gcc -O2 -I/usr/include/lua5.1 -c luaiconv.c -o luaiconv.o -I/usr/include
gcc -shared -o iconv.so -L/usr/local/lib luaiconv.o -L/usr/lib
Updating manifest for /home/baiyanh/.luarocks/lib/luarocks/rocks

要熟悉iconv很简单,因为linux下自带了一个iconv的命令行工具,查看支持的编码:

$ iconv --list | grep "^GB"
GB//
GB2312//
GB13000//
GB18030//
GBK//
GB_1988-80//
GB_198880//

修改代码:

iconv = require "iconv"
cd = iconv.new('gb2312', 'utf8')
starting = cd:iconv('开奖结果')

运行:

$ lua getlotterywinner.lua 2012138
01 07 16 17 19 21 14

 

【玩点酷的】

下面用这个小程序做点有意思的事:查看03年以来蓝色球各个数字的概率,或许对指导我买彩票有一定的帮助(篮球中了,最少也有5块奖金了:))

历年每年的期数都不相同,但所有的期数都列在这个网页中了:http://61.129.89.226/fcopen/cp_kjgg_dfw.jsp?lottery_type=ssq&lottery_issue=2012138,所有只要拿到所有的期数,然后调用上面的函数得到中奖号码,统计一下就ok了:

function allissues()
        local client = assert(socket.connect(host, port))
        client:send('GET ' .. string.format(fileformat, 2012138)  .. " HTTP/1.0\r\n\r\n")
        local line = client:receive('*l')
        local issuepattern = '>%d%d%d%d%d%d%d<'
        return function ()
                while line do
                        local issuenum = line:match(issuepattern)
                        line = client:receive('*l')
                        if issuenum then
                                return issuenum:sub(2, -2)
                        end
                end
                return nil
        end

end

local data = {}
local count = 0
for issuenum in allissues() do
        count = count + 1
        print ("getting " .. issuenum .. "...")
        local winner = getlotterywinner(issuenum)
        local blue = tonumber(winner[7])
        if not data[blue] then
                data[blue] = 0
        end
        data[blue] = data[blue] + 1
end

for k, v in pairs(data) do
        print(string.format("%2d: %3d/%d = %0.4f", k,  v, count, v/count))
end

最终结果:

1: 90/1423 = 0.0632
2: 88/1423 = 0.0618
3: 90/1423 = 0.0632
4: 78/1423 = 0.0548
5: 96/1423 = 0.0675
6: 85/1423 = 0.0597
7: 77/1423 = 0.0541
8: 76/1423 = 0.0534
9: 100/1423 = 0.0703
10: 86/1423 = 0.0604
11: 101/1423 = 0.0710
12: 90/1423 = 0.0632
13: 91/1423 = 0.0639
14: 91/1423 = 0.0639
15: 91/1423 = 0.0639
16: 93/1423 = 0.0654

看来挺平均的,11和9稍微高一点,但也就胜出零点几个百分点而已。。。

 

【下一步】

做个web service封装以下这个很hardcode的实现,这样如果人家网页的格式变了,或者编码编了,或者我想换个新的实现,都不会影响最终使用者 - 及时更新web service即可

posted @ 2012-11-25 19:03  lzprgmr  阅读(3254)  评论(0编辑  收藏  举报

黄将军