Python寫的簡易採集器

#!/usr/bin/python
#-*-coding:utf-8-*-

# 簡易採集爬蟲
#    1.採集Yahoo!Answers，parseData函數修改一下，可以採集任何網站
#    2.需要sqlite3或者pysqlite支援
#    3.可以在DreamHost.com空間上面運行
#    4.可以修改User-Agent冒充搜尋引擎蜘蛛
#    5.可以設定暫停時間，控制採集速度
#    6.採集Yahoo會被封IP數小時，所以這個採集用處不大
# Author: Lukin<mylukin@gmail.com>
# Date  : 2008-09-25

# 匯入採集需要用到的模組
import re, sys, time
import httplib, os.path as osp
from urlparse import urlparse
# 使用sqite資料庫，為了相容DreamHost.com的空間，只能這麼寫了
try :
    import sqlite3 as sqlite
except ImportError:
    from pysqlite2 import dbapi2 as sqlite

# 採集速度控制，單位秒
sleep = 0
# 資料庫路徑
dbname = './database.db'
# 設定提交的header頭
headers = {"Accept": "*/*","Referer": "http://answers.yahoo.com/","User-Agent": "Mozilla/5.0+(compatible;+Googlebot/2.1;++http://www.google.com/bot.html)"}
# 串連伺服器
dl = httplib.HTTPConnection('answers.yahoo.com')
# 串連資料庫
conn = sqlite.connect(osp.abspath(dbname))

# 建立資料庫
def createDatabase():
    global conn,dbname;
    if osp.isfile(osp.abspath(dbname)) : return
    c = conn.cursor()
    # 建立url列表存放表
    c.execute('''CREATE TABLE IF NOT EXISTS [collect]([cid] INTEGER PRIMARY KEY,[curl] TEXT,[state] INTEGER DEFAULT '0',UNIQUE([curl]));''')
    c.execute('''CREATE INDEX IF NOT EXISTS [collect_idx_state] ON [collect]([state]);''')
    # 建立分類表
    c.execute('''CREATE TABLE IF NOT EXISTS [sorts]([sortid] INTEGER PRIMARY KEY,[sortname] TEXT,[sortpath] TEXT,[sortfoot] INTEGER DEFAULT '0',[sortnum] INTEGER DEFAULT '0',UNIQUE([sortpath]));''')
    c.execute('''CREATE INDEX IF NOT EXISTS [sorts_idx_sortname] ON [sorts]([sortname]);''')
    c.execute('''CREATE INDEX IF NOT EXISTS [sorts_idx_sortfoot] ON [sorts]([sortfoot]);''')
    # 建立文章表
    c.execute('''CREATE TABLE IF NOT EXISTS [article]([aid] INTEGER PRIMARY KEY,[sortid] INTEGER DEFAULT '0',[hits] INTEGER DEFAULT '0',[title] TEXT,[path] TEXT,[question] TEXT,[banswer] TEXT,[oanswer] TEXT,UNIQUE([path]));''')
    c.execute('''CREATE INDEX IF NOT EXISTS [article_idx_sortid] ON [article]([sortid]);''')
    # 事物提交
    conn.commit()
    c.close()

# 執行採集
def collect(url="http://answers.yahoo.com/"):
    global dl,error,headers; R = 0
    print "GET:",url
    urls = urlparse(url); path = urls[2];
    if urls[4]!='' : path += '?' + urls[4]
    dl.request(method="GET", url=path, headers=headers); rs = dl.getresponse()
    if rs.status==200 :
        R = parseData(rs.read(),url);
    else :
        print "3 seconds, try again ..."; time.sleep(3)
        dl.request(method="GET", url=path, headers=headers); rs = dl.getresponse()
        if rs.status==200 :
            R = parseData(rs.read(),url);
        else :
            print "3 seconds, try again ..."; time.sleep(3)
            dl.request(method="GET", url=path, headers=headers); rs = dl.getresponse()
            if rs.status==200 :
                R = parseData(rs.read(),url);
            else :
                print "Continue to collect ..."
                R = 3
    # 更新記錄
    updateOneUrl(url,R)
    # 返回結果
    return R

# 處理採集到的資料
def parseData(html,url):
    global dl,conn; R = 2;
    c = conn.cursor()
    # 格式化html代碼
    format = formatURL(clearBlank(html),url)
    # 取出所有的串連
    urls = re.findall(r'''(<a[^>]*?href="([^"]+)"[^>]*?>)|(<a[^>]*?href='([^']+)'[^>]*?>)''',format,re.I)
    if urls != None :
        i = 0
        # 迴圈所有的串連
        for regs in urls :
            # 得到一個單一的url
            sUrl = en2chr(regs[1].strip())
            # 判斷url是否符合規則，符合，則插入資料庫
            if re.search('http(.*?)/(dir|question)/index(.*?)',sUrl,re.I) != None :
                if re.search('http(.*?)/dir/index(.*?)',sUrl,re.I) != None:
                    if sUrl.find('link=list') == -1 and sUrl.find('link=over') == -1 :
                        sUrl+= '&link=over'
                    else:
                        sUrl = sUrl.replace('link=list','link=over')
                if sUrl[-11:]=='link=mailto' : continue
                try :
                    c.execute('INSERT INTO [collect]([curl])VALUES(?);',(sUrl,))
                    i = i + 1
                except sqlite.IntegrityError :
                    pass
        if i>0 : print "Message: %d get a new URL." % (i,)
    # 截取資料
    if re.search('http(.*)/question/index(.*)',url,re.I) != None :
        sortfoot = 0
        # 自動建立分類和分類別關係
        guide  = sect(format,'<ol id="yan-breadcrumbs">','</ol>','(<li>(.*?)Home(.*?)</li>)')
        aGuide = re.findall('<a[^>]*href="[^"]*"[^>]*>(.*?)</a>',guide,re.I)
        if aGuide != None :
            sortname = ""
            for sortname in aGuide :
                sortname = sortname.strip()
                sortpath = en2path(sortname)
                # 查詢分類是否存在
                c.execute('SELECT [sortid],[sortname] FROM [sorts] WHERE [sortpath]=? LIMIT 0,1;',(sortpath,))
                row = c.fetchone();
                # 分類不存在，添加分類
                if row==None :
                    c.execute('INSERT INTO [sorts]([sortname],[sortpath],[sortfoot])VALUES(?,?,?);',(sortname,sortpath,sortfoot))
                    sortfoot = c.lastrowid
                else:
                    sortfoot = row[0]
            # 標題
            title = sect(format,'<h1 class="subject">','</h1>')
            # 最佳答案
            BestAnswer = sect(format,'(<h2><span>Best Answer</span>(.*?)</h2>(.*?)<div class="content">)','(</div>)')
            # 最佳答案不存在，則不採集
            if BestAnswer != None :
                # 文章路徑
                path = en2path(sortname + '-' + title.strip())
                # 問題
                adddata = sect(format,'<div class="additional-details">','</div>')
                content = sect(format,'(<h1 class="subject">(.*?)<div class="content">)','(</div>)')
                if adddata != None : content += '<br/>' + adddata
                # 其他回答
                OtherAnswer = ''
                for regs in re.findall('<div class="qa-container">(.+?)<div class="utils-container">',format):
                    if regs.find('<h2>') == -1 and regs.find('</h2>') == -1 :
                        a1 = sect(regs,'<div class="content">','</div>')
                        a2 = sect(regs,'<div class="reference">','</div>')
                        OtherAnswer+= '<div class="oAnswer">' + a1
                        if a2 != None : OtherAnswer+= '<div class="reference">' + a2 + '</div>'
                        OtherAnswer+= '</div>'
                # 判斷採集成功
                if title != None and content != None :
                    # 將資料寫入到資料
                    try :
                        c.execute('INSERT INTO [article]([sortid],[title],[path],[question],[banswer],[oanswer])VALUES(?,?,?,?,?,?);',(sortfoot,title,path,content,BestAnswer,OtherAnswer))
                        print "Message：%s.html" % (path,)
                        R = 1
                    except sqlite.IntegrityError :
                        pass
    # 提交寫入資料庫
    conn.commit(); c.close()
    return R
# 取得一條URL
def getOneUrl():
    global conn; c = conn.cursor()
    c.execute('SELECT [curl] FROM [collect] WHERE [state] IN(0,3) LIMIT 0,1;')
    row = c.fetchone(); c.close()
    if row==None : return ""
    return row[0].encode('utf-8')

# 更新一條記錄的狀態
def updateOneUrl(url,state):
    global conn; c = conn.cursor()
    c.execute('UPDATE [collect] SET [state]=? WHERE [curl]=?;',(state,url))
    conn.commit(); c.close()

# 清除html代碼裡的多餘空格
def clearBlank(html):
    if len(html) == 0 : return ''
    html = re.sub('\r|\n|\t','',html)
    while html.find("  ")!=-1 or html.find('&nbsp;')!=-1 :
        html = html.replace('&nbsp;',' ').replace('  ',' ')
    return html

# 格式化url
def formatURL(html,url):
    urls = re.findall('''(<a[^>]*?href="([^"]+)"[^>]*?>)|(<a[^>]*?href='([^']+)'[^>]*?>)''',html,re.I)
    if urls == None : return html
    for regs in urls :
        html = html.replace(regs[0],matchURL(regs[0],url))
    return html

# 格式化單個url
def matchURL(tag,url):
    urls = re.findall('''(.*)(src|href)=(.+?)( |/>|>).*|(.*)url\(([^\)]+)\)''',tag,re.I)
    if urls == None :
        return tag
    else :
        if urls[0][5] == '' :
            urlQuote = urls[0][2]
        else:
            urlQuote = urls[0][5]

    if len(urlQuote) > 0 :
        cUrl = re.sub('''['"]''','',urlQuote)
    else :
        return tag

    urls = urlparse(url); scheme = urls[0];
    if scheme!='' : scheme+='://'
    host = urls[1]; host = scheme + host
    if len(host)==0 : return tag
    path = osp.dirname(urls[2]);
    if path=='/' : path = '';
    if cUrl.find("#")!=-1 : cUrl = cUrl[:cUrl.find("#")]
    # 判斷類型
    if re.search('''^(http|https|ftp):(//|\\\\)(([\w/\\\+\-~`@:%])+\.)+([\w/\\\.\=\?\+\-~`@':!%#]|(&amp;)|&)+''',cUrl,re.I) != None :
        # http開頭的url類型要跳過
        return tag
    elif cUrl[:1] == '/' :
        # 絕對路徑
        cUrl = host + cUrl
    elif cUrl[:3]=='../' :
        # 相對路徑
        while cUrl[:3]=='../' :
            cUrl = cUrl[3:]
            if len(path) > 0 :
                path = osp.dirname(path)
    elif cUrl[:2]=='./' :
        cUrl = host + path + cUrl[1:]
    elif cUrl.lower()[:7]=='mailto:' or cUrl.lower()[:11]=='javascript:' :
        return tag
    else :
        cUrl = host + path + '/' + cUrl
    R = tag.replace(urlQuote,'"' + cUrl + '"')
    return R

# html代碼截取函數
def sect(html,start,end,cls=''):
    if len(html)==0 : return ;
    # Regex截取
    if start[:1]==chr(40) and start[-1:]==chr(41) and end[:1]==chr(40) and end[-1:]==chr(41) :
        reHTML = re.search(start + '(.*?)' + end,html,re.I)
        if reHTML == None : return
        reHTML = reHTML.group()
        intStart = re.search(start,reHTML,re.I).end()
        intEnd = re.search(end,reHTML,re.I).start()
        R = reHTML[intStart:intEnd]
    # 字串截取
    else :
        # 取得開始字串的位置
        intStart = html.lower().find(start.lower())
        # 如果搜尋不到開始字串，則直接返回空
        if intStart == -1 : return
        # 取得結束字串的位置
        intEnd = html[intStart+len(start):].lower().find(end.lower())
        # 如果搜尋不到結束字串，也返回為空白
        if intEnd == -1 : return
        # 開始和結束字串都有了，可以開始截取了
        R = html[intStart+len(start):intStart+intEnd+len(start)]
    # 清理內容
    if cls != '' :
        R = clear(R,cls)
    # 返回截取的字元
    return R

# 正則清除
def clear(html,regexs):
    if regexs == '' : return html
    for regex in regexs.split(chr(10)):
        regex = regex.strip()
        if regex != '' :
            if regex[:1]==chr(40) and regex[-1:]==chr(41):
                html = re.sub(regex,'',html,re.I|re.S)
            else :
                html = html.replace(regex,'')
    return html

# 格式化為路徑
def en2path(enStr):
    return re.sub('[\W]+','-',en2chr(enStr),re.I|re.U).strip('-')

# 替換實體為正常字元
def en2chr(enStr):
    return enStr.replace('&amp;','&')

# ------------------------------------- 開始執行程式 -------------------------------------------

# 首先建立資料庫
createDatabase()

# 開始採集
loops = 0
while True:
    if loops>0 :
        url = getOneUrl()
        if url == "" :
            loops = 0
        else :
            loops = collect(url)
    else :
        loops = collect()
    # 暫停
    time.sleep(sleep)
    if loops==0 : break
# 關閉HTTP串連
dl.close()
# 退出程式
sys.exit()

 

本文是引用 http://www.cnblogs.com/kuyuecs/archive/2008/10/15/1311346.html