#!/usr/bin/env python
#_*_ coding:utf-8 _*_
import threading
import urllib2
import time
import hashlib
import os



class MultithreadDownload(threading.Thread):
    '''
    url 请求的url
    start_position 多线程下载段落的起始位置
    end_position 多线程下载段落的结束位置
    openfile 打开的文件对象
    buf 一次获取数据的buffer值
    '''

    lock = threading.Lock()

    def __init__(self, url, start_position, end_position, openfile, buf):
        self.url = url
        self.buffer = buf
        self.start_position = start_position
        self.end_position = end_position
        self.fobj = openfile
        self.offset = 0
        threading.Thread.__init__(self)

    def run(self):
        req = urllib2.Request(self.url)
        req.add_header('Range', 'bytes=%s-%s' % (self.start_position, self.end_position))
        #print self.start_position,self.end_position
        request_segment = urllib2.urlopen(req)
        block = request_segment.read(self.buffer)
        self.offset = self.start_position
        while block:
            self.save(block)
            # 当这段所有信息都获取到了跳出循环退出
            if self.offset >= self.end_position +1:
                break
            elif self.offset + self.buffer > self.end_position +1:
                block = request_segment.read(self.end_position - self.offset +1)
                self.save(block)
                break
            else:
                block = request_segment.read(self.buffer)

    def save(self,block):
        #线程锁，先在文件对象中移动至偏移量处，第一次为段落开始位置，然后写数据，偏移量增加。
        MultithreadDownload.lock.acquire()
        self.fobj.seek(self.offset)
        self.fobj.write(block)
        self.offset = self.offset + len(block)
        MultithreadDownload.lock.release()

def download(url,threadnum=5,localfile="/tmp/tmp.txt",buf=4096):
    while True:
        fileCheck(localfile)
        starttime = time.time()  
        req = urllib2.urlopen(url)
        file_size = int(req.info().getheaders('Content-Length')[0])
        md5_remote = req.info().getheaders('ETag')[0].strip("\"")
        #每个线程处理的每段大小以及最后剩余部分大小
        segment_size,remain_size= divmod(file_size, threadnum)
        with open(localfile,'wb') as openfile:
            threadlst = []
            for i in range(threadnum):
                start_position = i*segment_size 
                end_position = start_position + segment_size -1
                if i == threadnum - 1:
                    end_position = end_position + remain_size
                begindowload = MultithreadDownload(url, start_position, end_position, openfile, buf)
                threadlst.append(begindowload)
            for t in threadlst:
                t.start()
            for t in threadlst:
                t.join()

        print 'Download completed!'
        endtime = time.time()
        usedtime =  endtime - starttime
        print usedtime

        md5_local = md5sum(localfile)
        print md5_local,md5_remote
        if md5_local == md5_remote:
            print "md5 check ok"
            break
        else:
            print "download file fail md5 error"
            break

def md5sum(localfile):
    with open(localfile,'rb') as f:
        m = hashlib.md5()
        m.update(f.read())
    return m.hexdigest()

def fileCheck(localfile):
    if os.path.exists(localfile):
        os.remove(localfile)

if __name__ == '__main__':
    url = 'https://artifacts.elastic.co/downloads/elasticsearch/elasticsearch-5.0.1.rpm'
    download(url=url,threadnum=10,localfile='test.rpm',buf=4096)