抓妹子图平台的实现，你们就是喜欢这种东西，对吧？（之三）抓图流程的实现

福利，三俗，喜闻乐见

为了后面能够接上web界面，需要一个可随时提供查询和推送信息的结构。一开始，逻辑不用做得太复杂，也不用过多考虑性能问题，只要保证了这样的结构，以后再修改也是可以的。

要实现结构就要选取异步io操作的方式，这个问题让我老头疼了，本人用python编程的时间也不长，知道的路数不多，multiprocessing在windows下是用不了的，gevent安装又太麻烦，需要多少第三方的东西，还需要编译器，还是硬着头皮选择用thread。心里有个架构，可以保证前台的流畅，但实现比较复杂，肯定暂时不会考虑用在这个例子上，不过前面说了，只要结构做出来，以后如果有性能上的需求，放到linux上，到时候再修改也不迟。

python3的新版本有个Future包，封装了异步相关的操作，也算方便，而且就有现成的多线程下载例子

http://docs.python.org/3.3/library/concurrent.futures.html?highlight=future#concurrent.futures.ThreadPoolExecutor

真是想人之所想啊，知道你们都喜欢这种;-)

对每一个网页，给一个任务编号。page_x

对每一个图片，给一个任务编号。pic_x

此外，每一个网页有一个图片列表。和网页原地址

此外，每一个图片有这几个数据项。

状态:0,未开始,1,排队待下,2,下载完毕

所属地址id

已下载bytes

总长bytes

如果这两个值都为0，表示未下载，

如果总长有值，表示下载中，已下载bytes表示它的进度

图片原地址

经过一些尝试，找到了下载的方法，下面是代码，代码是阻塞式实现的，先把过程调通，下一步再改成非阻塞式。实际上现在这个版本已经可以用了

# -*- coding: utf8 -*-
import concurrent.futures
import urllib.request
import re
import json
import ast


visitedURL = {}		
		
maxpageID = [0]
pageDict = {}

def createPageID(url):
	pID = maxpageID.pop()+1
	maxpageID.append(pID)
	pID = 'page_'+str(pID)
	visitedURL[url] = pID
	return pID


maxpicID = [0]
picDict = {}
def createPicID(url):
	pID = maxpicID.pop()+1
	maxpicID.append(pID)
	pID = 'pic_'+str(pID)
	visitedURL[url] = pID
	return pID



stoppedQueue = []
waitingQueue = []
downloadingQueue = []
savedDict = dict()

#for page downloading
pageTpe = concurrent.futures.ThreadPoolExecutor(max_workers=8)
#for picture downloading
picTpe = concurrent.futures.ThreadPoolExecutor(max_workers=4)


def runMachine():
    #add at least 4 tasks to download
    while waitingQueue:
        if len(downloadingQueue)<4:
            picID = waitingQueue.pop(0)
            processload(picID)

def processload(picID):
    downloadingQueue.append(picID)
    #open conn,loading a picture
    picInfo = picDict[picID]
    url = picInfo['url']
    filename = url.split('/')[-1]
    conn = urllib.request.urlopen(url,timeout=10)
    picInfo['total'] = int(conn.info().get('Content-Length').strip())
    outputfile = open('pics/'+filename,'wb')
    picInfo['progress'] = 0
    updateStatus(picInfo)
    while True:
        chunk = conn.read(4096)
        picInfo['progress']+=len(chunk)
        updateStatus(picInfo)
        if not chunk:
            picInfo['state'] = 2
            downloadingQueue.remove(picID)
            savedDict[picID] = True
            updateStatus(picInfo)
            outputfile.close()
            conn.close()
            break
        outputfile.write(chunk)
        #reportProgress(url,progress,total)
    #report

def updateStatus(picInfo):
    url = picInfo['url']
    if picInfo['state']==2:
        print(url,'finished!')
    elif  picInfo['total'] and picInfo['progress']:
        print('{} progress: {:.2%}'.format(url,(picInfo['progress']/picInfo['total'])))
        
    pass

def log(*args):
	f = open('t.txt','ba')
	f.write((','.join(map(str,args))+'\n').encode('utf-8'))
	f.close()

def load_pic(url,pageID):
    if url in visitedURL:
        return
    picID = createPicID(url)
    #状态:0,未开始,1,排队待下,2,下载完毕
    picDict[picID] = {'url':url,'pageID':pageID,'total':0,'progress':0,'state':1}
    waitingQueue.append(picID)
    

def load_page(url):
    if url in visitedURL:
            return
    pID = createPageID(url)
    pageDict[pID] = {'url':url,'links':None}
    conn = urllib.request.urlopen(url)
    text = conn.readall().decode('GBK').encode('utf-8').decode('utf-8')
    conn.close()
    try:
            startIndex = text.index('<div class="mod newslist clear">')
            endIndex = text.index('<div class="mod curPosition clear">',startIndex)
            text = text[startIndex:endIndex]
            patt = re.compile('href="([^"]+?).htm"><img', re.DOTALL | re.IGNORECASE)
            jsurls = [x+'.hdBigPic.js' for x in patt.findall(text)]
            pageurllist = []
            for jsurl in jsurls:
                    if jsurl in visitedURL:
                            continue
                    jsID = createPageID(jsurl)
                    pageDict[jsID] = {'url':jsurl,'links':None}
                    jslinks = []
                    try:
                            conn = urllib.request.urlopen(jsurl)
                    except BaseException as e:
                            print('failed')
                            continue
                    try:
                        text = conn.readall().decode('GBK').encode('utf-8').decode('utf-8')
                        text = text[:text.index('/*  |xGv00|')]
                        obj = ast.literal_eval(text)
                        picnum = int(obj['Children'][0]['Children'][0]['Children'][0]['Content'])
                        picsobj = obj['Children'][0]['Children'][1]['Children']
                        for x in picsobj:
                            picurl = x['Children'][2]['Children'][0]['Content']
                            jslinks.append(picurl)
                        if jslinks:
                            pageDict[jsID]['links'] = jslinks
                            print(jsurl,'{} pics'.format(len(jslinks)))
                        try:
                            title = obj['Children'][0]['Children'][8]['Children'][0]['Content']
                        except:
                            title = 'unknown'
                        pageurllist.append(jsurl)
                        for picurl in jslinks:
                            load_pic(picurl,jsID)
                    except BaseException as e:
                            print(jsurl,'failed')
                            raise e
            pageDict[pID]['links'] = pageurllist
            
                    
    except ValueError as e:
            print('error',e)
            #can't find proper place
            pass
    runMachine()
	
urls = ['http://games.qq.com/l/photo/gmcos/yxcos.htm']

load_page(urls[0])

来源：oschina

链接：https://my.oschina.net/u/1167335/blog/153501