python+mongodb數(shù)據(jù)抓取詳細(xì)介紹

Python數(shù)據(jù)抓取分析
編程模塊：requests,lxml，pymongo，time，BeautifulSoup

首先獲取所有產(chǎn)品的分類網(wǎng)址：    
def step():
  try:
    headers = {
      。。。。。
      }
    r = requests.get(url,headers,timeout=30)
    html = r.content
    soup = BeautifulSoup(html,"lxml")
    url = soup.find_all(正則表達(dá)式)
    for i in url:
      url2 = i.find_all('a')
      for j in url2:
         step1url =url + j['href']
         print step1url
         step2(step1url)
  except Exception,e:
    print e

我們?cè)诋a(chǎn)品分類的同時(shí)需要確定我們所訪問的地址是產(chǎn)品還是又一個(gè)分類的產(chǎn)品地址（所以需要判斷我們?cè)L問的地址是否含有if判斷標(biāo)志）：    
def step2(step1url):
  try:
    headers = {
      。。。。
      }
    r = requests.get(step1url,headers,timeout=30)
    html = r.content
    soup = BeautifulSoup(html,"lxml")
    a = soup.find('div',id='divTbl')
    if a:
      url = soup.find_all('td',class_='S-ITabs')
      for i in url:
        classifyurl = i.find_all('a')
        for j in classifyurl:
           step2url = url + j['href']
           #print step2url
           step3(step2url)
    else:
      postdata(step1url)

當(dāng)我們if判斷后為真則將第二頁(yè)的分類網(wǎng)址獲取到（第一個(gè)步驟），否則執(zhí)行postdata函數(shù)，將網(wǎng)頁(yè)產(chǎn)品地址抓取！    
def producturl(url):
  try:
    p1url = doc.xpath(正則表達(dá)式)
    for i in xrange(1,len(p1url) + 1):
      p2url = doc.xpath(正則表達(dá)式)
      if len(p2url) > 0:
        producturl = url + p2url[0].get('href')
        count = db[table].find({'url':producturl}).count()
        if count <= 0:
            sn = getNewsn()
            db[table].insert({"sn":sn,"url":producturl})
            print str(sn) + 'inserted successfully'
        else:
            'url exist'
 
  except Exception,e:
    print e

其中為我們所獲取到的產(chǎn)品地址并存入mongodb中，sn作為地址的新id。

下面我們需要在mongodb中通過(guò)新id索引來(lái)獲取我們的網(wǎng)址并進(jìn)行訪問，對(duì)產(chǎn)品進(jìn)行數(shù)據(jù)分析并抓取，將數(shù)據(jù)更新進(jìn)數(shù)據(jù)庫(kù)內(nèi)！

其中用到最多的BeautifulSoup這個(gè)模塊，但是對(duì)于存在于js的價(jià)值數(shù)據(jù)使用BeautifulSoup就用起來(lái)很吃力，所以對(duì)于js中的數(shù)據(jù)我推薦使用xpath，但是解析網(wǎng)頁(yè)就需要用到HTML.document_fromstring(url)方法來(lái)解析網(wǎng)頁(yè)。

對(duì)于xpath抓取價(jià)值數(shù)據(jù)的同時(shí)一定要細(xì)心！如果想了解xpath就在下面留言，我會(huì)盡快回答！    
def parser(sn,url):
  try:
    headers = {
      。。。。。。
      }
    r = requests.get(url, headers=headers,timeout=30)
    html = r.content
    soup = BeautifulSoup(html,"lxml")
    dt = {}
    #partno
    a = soup.find("meta",itemprop="mpn")
    if a:
      dt['partno'] = a['content']
    #manufacturer
    b = soup.find("meta",itemprop="manufacturer")
    if b:
      dt['manufacturer'] = b['content']
    #description
    c = soup.find("span",itemprop="description")
    if c:
      dt['description'] = c.get_text().strip()
    #price
    price = soup.find("table",class_="table table-condensed occalc_pa_table")
    if price:
      cost = {}
      for i in price.find_all('tr'):
        if len(i) > 1:
          td = i.find_all('td')
          key=td[0].get_text().strip().replace(',','')
          val=td[1].get_text().replace(u'\u20ac','').strip()
          if key and val:
            cost[key] = val
      if cost:
        dt['cost'] = cost
        dt['currency'] = 'EUR'
    #quantity
    d = soup.find("input",id="ItemQuantity")
    if d:
      dt['quantity'] = d['value']
    #specs
    e = soup.find("div",class_="row parameter-container")
    if e:
      key1 = []
      val1= []
      for k in e.find_all('dt'):
        key = k.get_text().strip().strip('.')
        if key:
          key1.append(key)
      for i in e.find_all('dd'):
        val = i.get_text().strip()
        if val:
          val1.append(val)
      specs = dict(zip(key1,val1))
    if specs:
      dt['specs'] = specs
      print dt
       
    if dt:
      db[table].update({'sn':sn},{'$set':dt})
      print str(sn) + ' insert successfully'
      time.sleep(3)
    else:
      error(str(sn) + '\t' + url)
  except Exception,e:
    error(str(sn) + '\t' + url)
    print "Don't data!"

最后全部程序運(yùn)行，將價(jià)值數(shù)據(jù)分析處理并存入數(shù)據(jù)庫(kù)中！