知乎話題爬蟲完美版（python）

時間 2019-11-16
原文原文鏈接
由於要作觀點，觀點的屋子相似於知乎的話題，因此得想辦法把他給爬下來，搞了半天最終仍是妥妥的搞定了，代碼是python寫的，不懂得麻煩自學哈！懂得直接看代碼，絕對可用。html
#coding:utf-8
"""
@author:haoning
@create time:2015.8.5
"""
from __future__ import division  # 精確除法
from Queue import Queue
from __builtin__ import False
import json
import os
import re
import platform
import uuid
import urllib
import urllib2
import sys
import time
import MySQLdb as mdb
from bs4 import BeautifulSoup

reload(sys)
sys.setdefaultencoding( "utf-8" )

headers = {
   'User-Agent' : 'Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:35.0) Gecko/20100101 Firefox/35.0',
   'Content-Type':'application/x-www-form-urlencoded; charset=UTF-8',
   'X-Requested-With':'XMLHttpRequest',
   'Referer':'https://www.zhihu.com/topics',
   'Cookie':'__utma=51854390.517069884.1416212035.1416212035.1416212035.1; q_c1=c02bf44d00d240798bfabcfc95baeb56|1455778173000|1416205243000; _za=b1c8ae35-f986-46a2-b24a-cb9359dc6b2a; aliyungf_tc=AQAAAJ1m71jL1woArKqF22VFnL/wRy6C; _xsrf=9d494558f9271340ab24598d85b2a3c8; cap_id="MDNiMjcwM2U0MTRhNDVmYjgxZWVhOWI0NTA2OGU5OTg=|1455864276|2a4ce8247ebd3c0df5393bb5661713ad9eec01dd"; n_c=1; _alicdn_sec=56c6ba4d556557d27a0f8c876f563d12a285f33a'
}

DB_HOST = '127.0.0.1'
DB_USER = 'root'
DB_PASS = 'root'

queue= Queue() #接收隊列
nodeSet=set()
keywordSet=set()
stop=0
offset=-20
level=0
maxLevel=7
counter=0
base=""

conn = mdb.connect(DB_HOST, DB_USER, DB_PASS, 'zhihu', charset='utf8')
conn.autocommit(False)
curr = conn.cursor()

def get_html(url):
    try:
        req = urllib2.Request(url)
        response = urllib2.urlopen(req,None,3) #在這裏應該加入代理
        html = response.read()
        return html
    except:
        pass
    return None

def getTopics():
    url = 'https://www.zhihu.com/topics'
    print url
    try:
        req = urllib2.Request(url)
        response = urllib2.urlopen(req) #鍦ㄨ繖閲屽簲璇ュ姞鍏ヤ唬鐞�
        html = response.read().decode('utf-8')
        print html
        soup = BeautifulSoup(html)
        lis = soup.find_all('li', {'class' : 'zm-topic-cat-item'})
        
        for li in lis:
            data_id=li.get('data-id')
            name=li.text
            curr.execute('select id from classify_new where name=%s',(name))
            y= curr.fetchone()
            if not y:
                curr.execute('INSERT INTO classify_new(data_id,name)VALUES(%s,%s)',(data_id,name))
        conn.commit()
    except Exception as e:
        print "get topic error",e
        

def get_extension(name):  
    where=name.rfind('.')
    if where!=-1:
        return name[where:len(name)]
    return None


def which_platform():
    sys_str = platform.system()
    return sys_str

def GetDateString():
    when=time.strftime('%Y-%m-%d',time.localtime(time.time()))
    foldername = str(when)
    return foldername 

def makeDateFolder(par,classify):
    try:
        if os.path.isdir(par):
            newFolderName=par + '//' + GetDateString() + '//'  +str(classify)
            if which_platform()=="Linux":
                newFolderName=par + '/' + GetDateString() + "/" +str(classify)
            if not os.path.isdir( newFolderName ):
                os.makedirs( newFolderName )
            return newFolderName
        else:
            return None 
    except Exception,e:
        print "kk",e
    return None 

def download_img(url,classify):
    try:
        extention=get_extension(url)
        if(extention is None):
            return None
        req = urllib2.Request(url)
        resp = urllib2.urlopen(req,None,3)
        dataimg=resp.read()
        name=str(uuid.uuid1()).replace("-","")+"_www.guandn.com"+extention
        top="E://topic_pic"
        folder=makeDateFolder(top, classify)
        filename=None
        if folder is not None:
            filename  =folder+"//"+name
        try:
            if "e82bab09c_m" in str(url):
                return True
            if not os.path.exists(filename):
                file_object = open(filename,'w+b')
                file_object.write(dataimg)
                file_object.close()
                return '/room/default/'+GetDateString()+'/'+str(classify)+"/"+name
            else:
                print "file exist"
                return None
        except IOError,e1:
            print "e1=",e1
            pass
    except Exception as e:
        print "eee",e
        pass
    return None #若是沒有下載下來就利用原來網站的連接

def getChildren(node,name):
    global queue,nodeSet
    try:
        url="https://www.zhihu.com/topic/"+str(node)+"/hot"
        html=get_html(url)
        if html is None:
            return
        soup = BeautifulSoup(html)
        p_ch='父話題'
        node_name=soup.find('div', {'id' : 'zh-topic-title'}).find('h1').text
        topic_cla=soup.find('div', {'class' : 'child-topic'})
        if topic_cla is not None:
            try:
                p_ch=str(topic_cla.text)
                aList = soup.find_all('a', {'class' : 'zm-item-tag'}) #獲取全部子節點
                if u'子話題' in p_ch:
                    for a in aList:
                        token=a.get('data-token')
                        a=str(a).replace('\n','').replace('\t','').replace('\r','')
                        start=str(a).find('>')
                        end=str(a).rfind('</a>')
                        new_node=str(str(a)[start+1:end])
                        curr.execute('select id from rooms where name=%s',(new_node)) #先保證名字毫不相同
                        y= curr.fetchone()
                        if not y:
                            print "y=",y,"new_node=",new_node,"token=",token
                            queue.put((token,new_node,node_name))
            except Exception as e:
                print "add queue error",e
    except Exception as e:
        print "get html error",e
        
    

def getContent(n,name,p,top_id):
    try:
        global counter
        curr.execute('select id from rooms where name=%s',(name)) #先保證名字毫不相同
        y= curr.fetchone()
        print "exist?? ",y,"n=",n
        if not y:
            url="https://www.zhihu.com/topic/"+str(n)+"/hot"
            html=get_html(url)
            if html is None:
                return
            soup = BeautifulSoup(html)
            title=soup.find('div', {'id' : 'zh-topic-title'}).find('h1').text
            pic_path=soup.find('a',{'id':'zh-avartar-edit-form'}).find('img').get('src')
            description=soup.find('div',{'class':'zm-editable-content'})
            if description is not None:
                description=description.text
                
            if (u"未歸類" in title or u"根話題" in title): #容許入庫，避免死循環
                description=None
                
            tag_path=download_img(pic_path,top_id)
            print "tag_path=",tag_path
            if (tag_path is not None) or tag_path==True:
                if tag_path==True:
                    tag_path=None
                father_id=2 #默認爲雜談
                curr.execute('select id from rooms where name=%s',(p))
                results = curr.fetchall()
                for r in results:
                    father_id=r[0]
                name=title
                curr.execute('select id from rooms where name=%s',(name)) #先保證名字毫不相同
                y= curr.fetchone()
                print "store see..",y
                if not y:
                    friends_num=0
                    temp = time.time()
                    x = time.localtime(float(temp))
                    create_time = time.strftime("%Y-%m-%d %H:%M:%S",x) # get time now
                    create_time
                    creater_id=None
                    room_avatar=tag_path
                    is_pass=1
                    has_index=0
                    reason_id=None  
                    #print father_id,name,friends_num,create_time,creater_id,room_avatar,is_pass,has_index,reason_id
                    ######################有資格入庫的內容
                    counter=counter+1
                    curr.execute("INSERT INTO rooms(father_id,name,friends_num,description,create_time,creater_id,room_avatar,is_pass,has_index,reason_id)VALUES(%s,%s,%s,%s,%s,%s,%s,%s,%s,%s)",(father_id,name,friends_num,description,create_time,creater_id,room_avatar,is_pass,has_index,reason_id))
                    conn.commit() #必須時時進入數據庫，否則找不到父節點
                    if counter % 200==0:
                        print "current node",name,"num",counter
    except Exception as e:
        print "get content error",e       

def work():
    global queue
    curr.execute('select id,node,parent,name from classify where status=1')
    results = curr.fetchall()
    for r in results:
        top_id=r[0]
        node=r[1]
        parent=r[2]
        name=r[3]
        try:
            queue.put((node,name,parent)) #首先放入隊列
            while queue.qsize() >0:
                n,p=queue.get() #頂節點出隊
                getContent(n,p,top_id)
                getChildren(n,name) #出隊內容的子節點
            conn.commit()
        except Exception as e:
            print "what's wrong",e  
            
def new_work():
    global queue
    curr.execute('select id,data_id,name from classify_new_copy where status=1')
    results = curr.fetchall()
    for r in results:
        top_id=r[0]
        data_id=r[1]
        name=r[2]
        try:
            get_topis(data_id,name,top_id)
        except:
            pass


def get_topis(data_id,name,top_id):
    global queue
    url = 'https://www.zhihu.com/node/TopicsPlazzaListV2'
    isGet = True;
    offset = -20;
    data_id=str(data_id)
    while isGet:
        offset = offset + 20
        values = {'method': 'next', 'params': '{"topic_id":'+data_id+',"offset":'+str(offset)+',"hash_id":""}'}
        try:
            msg=None
            try:
                data = urllib.urlencode(values)
                request = urllib2.Request(url,data,headers)
                response = urllib2.urlopen(request,None,5)
                html=response.read().decode('utf-8')
                json_str = json.loads(html)
                ms=json_str['msg']
                if len(ms) <5:
                    break
                msg=ms[0]
            except Exception as e:
                print "eeeee",e
            #print msg
            if msg is not None:
                soup = BeautifulSoup(str(msg))
                blks = soup.find_all('div', {'class' : 'blk'})
                for blk in blks:
                    page=blk.find('a').get('href')
                    if page is not None:
                        node=page.replace("/topic/","") #將更多的種子入庫
                        parent=name
                        ne=blk.find('strong').text
                        try:
                            queue.put((node,ne,parent)) #首先放入隊列
                            while queue.qsize() >0:
                                n,name,p=queue.get() #頂節點出隊
                                size=queue.qsize()
                                if size > 0:
                                    print size
                                getContent(n,name,p,top_id)
                                getChildren(n,name) #出隊內容的子節點
                            conn.commit()
                        except Exception as e:
                            print "what's wrong",e  
        except urllib2.URLError, e:
            print "error is",e
            pass 
            
        
if __name__ == '__main__':
    i=0
    while i<400:
        new_work()
        i=i+1
說下數據庫的問題，我這裏就不傳附件了，看字段本身創建，由於這確實太簡單了，我是用的mysql，你看本身的需求本身建。node
有什麼不懂得麻煩去去轉盤網找我，由於這個也是我開發的，上面會及時更新qq羣號，這裏不留qq號啥的，以避免被系統給K了。python