python利用scrapy框架爬取起点

先上自己做完之后回顾细节和思路的东西,之后代码一起上。

1.Mongodb 建立一个叫QiDian的库,
然后建立了一个叫Novelclass(小说类别表)
Novelclass(可以把一级类别二级类别都存进去:玄幻--一级类别,东方玄幻--二级类别)的表

client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Novelclass

2.用parse回调方法,获得一级类别。循环取出(要注意拼接问题--"https:"+)。
并将一级类别存入Mongodb(注意一级类别的pid此时为None)。一级类别的链接不需要存入redis数据库(一级类别的链接只为了找到二级)。

3.此时获取二级类别(东方玄幻),此时回调parse方法
取得二级类别(名称+链接)同时二级类别名称与一级类别名称放入同一个Mongodb(Novelclass)里,链接则存入redis库(classid = self.insertMongo(name[0],pid),self.pushRedis(classid,url,pid))

def insertMongo(self,classname,pid):
classid = collection.insert({'classname':classname,'pid':pid})
return classid
def pushRedis(self,classid,url,pid,):
novelurl = '%s,%s,%s' %(classid,url,pid)
r.lpush('novelurl',novelurl)
此时第一步完成。一级类别与二级类别都获取到。

4.有了二级链接(东方玄幻),接下来要获取每个二级链接下的小说名字与链接。
(同样,将小说名字存入Mongodb--Novelname表,链接存入到redis--novelnameurl)
这里注意我们定义了一个字典,(为了我们能取到arr[0]--也就是二级的id(东方玄幻的id))
因为我们要确定每本小说是属于哪一类(东方玄幻或者西方玄幻。)
dict = {}
novelurl = bytes.decode(item)
arr = novelurl.split(',') # 分割字符串
qidianNovelSpider.start_urls.append(arr[1])
pid = arr[0] 二级(东方玄幻)的_id,也就是流水id,注意不要取成东方玄幻的pid(他的pid则是玄幻的id)
url = arr[1] 二级(东方玄幻)的链接
self.dict[url] = {"pid":pid,"num":0}
此时num是为了控制我们爬几页。
同样一个parse回调方法,因为我们要去下一页,所以我们要确定取到的链接是相同
classInfo = self.dict[response.url]--response.url固定就是这么写
pid = classInfo['pid']--确定了那么pid=arr[0]
num = classInfo['num']

if num>3:--------此处num的用处就来了,我只去每个二级链接的前4页(因为你是取完了前四页才循环回来,所以不是3页)
return None
同样注意链接的拼接问题。(不然会报keyerror错误)
取到就分别把名字和链接存入Mongodb和redis
classid = collection.insert({'novelname': name, 'pid': pid})此时pid就是(东方玄幻的id也就是唯一id,而不是玄幻的id)
print(name)
self.pushRedis(classid, c, pid)-----(classid就是流水id--_id,c是拼接后的链接,pid则是东方玄幻的id)
现在分类第一页能取,现在开始写下一页的,
hxs = HtmlXPathSelector(response)
hx = hxs.select('//li[@class="lbf-pagination-item"]/a[@class="lbf-pagination-next "]')
urls = hx.select("@href").extract()
d = "https:" + urls[0]
classInfo['num'] +=1-------(每取一页,num就+1)
self.dict[d] = classInfo
print(d)
request = Request(d, callback=self.parse)-------(调用上一个回调方法,也就是取那页名字和链接那个地方。)
yield request
这段代码就是为了取到下一页的链接然后去调用上一个方法,把下一页的名字和链接都拿下来。
之后就是入库操作,上面有。
Mongodb---Novelname,Redis----novelnameurl

5.接下来要做的是更新书籍信息(也就是把novelname表更新)
现在Mongodb----novelname表只有书籍名字,而没有具体信息。(所以要把作者,签约,连载或完结,免费或Vip都更新进去)
注意:我是更新而不是新建表。所以说我取得链接同样还是novelnameurl里的,只不过把得到的信息不用另起一个表,而是插入之前有的novelname
client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Novelname(这个表会同上一个py文件的相同)
同样的,我还是需要(东方玄幻的_id)
pid = arr[0]
url = arr[1]
self.dict[url] = {"pid":pid}
不过这次是为了更新Mongodb数据库准备的,
nameInfo = self.dict[response.url]
pid1 = nameInfo['pid']
pid = ObjectId(pid1)-------(此处是为了等我更新时,"_id"这个键对应相同的ObjectId)
之后就是取信息。
hx = hxs.select('//div[@class="book-info "]/h1/span/a[@class="writer"]')
hx1 =hxs.select('//p[@class="tag"]/span[@class="blue"]')
hx2 =hxs.select('//p[@class="tag"]/a[@class="red"]')
for secItem in hx:
writer = secItem.select("text()").extract()
print(writer)
for secItem1 in hx1:
state = secItem1.select("text()").extract()
print(state)
for secItem2 in hx2:
classes = secItem2.select("text()").extract()
print(classes)
可以这么取并且打印出来也是顺序,而不是说先把每个小说的writer都打印出来,在打印state。
更新Mongodb----novelname表
db.Novelname.update({"_id": pid}, {"$set": {"writer": writer, "state": state, "classes": classes}})
此时上面的pid就起了作用,更新完成。

6.接下来就是爬取每本小说的章节名和链接了。(同上面的取小说名字,只不过更简单,没有下一页)
这里还是要注意id问题,章节名称和链接都要对应上每本小说自己的_id(而不是二级的pid)。
插入Mongodb-----Chaptername,redis----chapterurl

7.最后一步,根据章节链接取小说内容,因为我们取了每本小说的所有的链接,所以也不用考虑下一章的问题。
同时,我们取完内容,也要将内容更新到章节表里,这里我们需要注意的是,我们取得小说的内容是p标签的(是字符串形式),
所以说我们插入Mongodb就遇到了问题,字符串放进去,一条就给你一个id,这不是想要的,要的是一章小说的就一个_id就好。
用到了字符串拼接。
ii=""-------先给个空的
hx = hxs.select('//div[@class="read-content j_readContent"]/p')

for secItem in hx:
contents = secItem.select("text()").extract()
content1 = contents[0]-----取到一个p下的内容
# print(content1)
ii=content1+ii-------取到的就加进去
print(ii)------我们想要的结果
最后更新进Chaptername,
db.Chaptername.update({"_id": pid}, {"$set": {"content": ii}})


Mongodb(Novelclass(一级,二级--玄幻,东方玄幻);Novelname(小说名字,----之后更新进作者,连载等等);Chaptername(章节名字,----更新进章节对应的内容))
Redis(novelurl(单纯二级链接---东方玄幻,用不到一级链接);novelnameurl(小说名字链接);chapterurl(章节链接))

 

第一个py文件

# -*- coding: utf-8 -*-
import re
from urllib.request import urlopen
from scrapy.http import Request
# from urllib.request import Request
from bs4 import BeautifulSoup
from lxml import etree
import pymongo
import scrapy
from scrapy.selector import HtmlXPathSelector
client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Novelclass          #表名classification


import redis        #导入redis数据库
r = redis.Redis(host='127.0.0.1', port=6379, db=0)

class qidianClassSpider(scrapy.Spider):
    name = "qidianClass2"
    allowed_domains = ["qidian.com"]   #允许访问的域
    start_urls = [
        "https://www.qidian.com/all",
    ]

    #每爬完一个网页会回调parse方法
    def parse(self, response):
        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//div[@class="work-filter type-filter"]/ul[@type="category"]/li[@class=""]/a')
        for secItem in hx:
            url = secItem.select("@href").extract()
            c = "https:"+url[0]
            name = secItem.select("text()").extract()
            classid = self.insertMongo(name[0],None)
            print(c)
            # a = db.Novelclass.find()
            # for item in a:
            #     print(item.get('_id'))
            # b = item.get('_id')
            # novelurl = '%s,%s' % (item.get('_id'), c)
            # r.lpush('novelurl', novelurl)
            request = Request(c,callback=lambda response,pid=str(classid):self.parse_subclass(response,pid))
            yield request
    def parse_subclass(self, response,pid):
        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//div[@class="sub-type"]/dl[@class=""]/dd[@class=""]/a')
        for secItem in hx:
            urls = secItem.select("@href").extract()
            url = "https:" + urls[0]
            name = secItem.select("text()").extract()
            classid = self.insertMongo(name[0],pid)
            self.pushRedis(classid,url,pid)

    def insertMongo(self,classname,pid):
        classid = collection.insert({'classname':classname,'pid':pid})
        return classid
    def pushRedis(self,classid,url,pid,):
        novelurl = '%s,%s,%s' %(classid,url,pid)
        r.lpush('novelurl',novelurl)

  第二个py文件

# -*- coding: utf-8 -*-
import re
from urllib.request import urlopen
from scrapy.http import Request
import pymongo
import scrapy
from time import sleep
from scrapy.selector import HtmlXPathSelector

client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Novelname

import redis  # 导入redis数据库

r = redis.Redis(host='127.0.0.1', port=6379, db=0)

ii = 0


class qidianNovelSpider(scrapy.Spider):
    name = "qidianClass3"
    allowed_domains = ["qidian.com"]
    dict = {}
    start_urls = []

    def __init__(self):  # 定义一个方法

        a = r.lrange('novelurl', 0, -1)
        # ii = 0
        for item in a:
            novelurl = bytes.decode(item)
            arr = novelurl.split(',')  # 分割字符串
            qidianNovelSpider.start_urls.append(arr[1])
            pid = arr[0]
            url = arr[1]
            self.dict[url] = {"pid":pid,"num":0}
            # ii +=1
            # if ii>3:
            #     break
                # qidianNovelSpider.start_urls = start_urls
            # print(start_urls)

    def parse(self, response):

        classInfo = self.dict[response.url]
        pid = classInfo['pid']
        num = classInfo['num']
        # print(self.dict)
        if num>3:
            return None
        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//div[@class="book-mid-info"]/h4/a')
        for secItem in hx:
            url = secItem.select("@href").extract()
            c = "https:" + url[0]
            name = secItem.select("text()").extract()
            classid = collection.insert({'novelname': name, 'pid': pid})
            print(name)
            self.pushRedis(classid, c, pid)

        print('-----------递归--------------')

        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//li[@class="lbf-pagination-item"]/a[@class="lbf-pagination-next "]')
        urls = hx.select("@href").extract()
        d = "https:" + urls[0]
        classInfo['num'] +=1
        self.dict[d] = classInfo
        print(d)
        request = Request(d, callback=self.parse)
        yield request
        print('--------end--------------')

    def pushRedis(self, classid, c, pid):
        novelnameurl = '%s,%s,%s' % (classid, c, pid)
        r.lpush('novelnameurl', novelnameurl)

  第三个py文件

# -*- coding: utf-8 -*-
import re
from urllib.request import urlopen
from scrapy.http import Request
import pymongo
import scrapy
from time import sleep
from scrapy.selector import HtmlXPathSelector
from bson.objectid import ObjectId

client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Novelname

import redis  # 导入redis数据库

r = redis.Redis(host='127.0.0.1', port=6379, db=0)

# ii = 0


class qidianNovelSpider1(scrapy.Spider):
    name = "qidianClass4"
    allowed_domains = ["qidian.com"]
    dict = {}
    start_urls = []

    def __init__(self):  # 定义一个方法

        a = r.lrange('novelnameurl', 0, -1)
        # ii = 0
        for item in a:
            novelnameurl = bytes.decode(item)
            arr = novelnameurl.split(',')  # 分割字符串
            qidianNovelSpider1.start_urls.append(arr[1])
            pid = arr[0]
            url = arr[1]
            self.dict[url] = {"pid":pid}

    def parse(self, response):
        nameInfo = self.dict[response.url]
        pid1 = nameInfo['pid']
        pid = ObjectId(pid1)
        print(pid)
        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//div[@class="book-info "]/h1/span/a[@class="writer"]')
        hx1 =hxs.select('//p[@class="tag"]/span[@class="blue"]')
        hx2 =hxs.select('//p[@class="tag"]/a[@class="red"]')
        for secItem in hx:
            writer = secItem.select("text()").extract()
            print(writer)
        for secItem1 in hx1:
            state = secItem1.select("text()").extract()
            print(state)
        for secItem2 in hx2:
            classes = secItem2.select("text()").extract()
            print(classes)
            # for item in a:
            #     b = item.get('_id')
            #     print(b)

            db.Novelname.update({"_id": pid}, {"$set": {"writer": writer, "state": state, "classes": classes}})
            print('------------------------------------------')


            # classid = collection.insert({'novelname': name, 'pid': Pid})
            # print(name)
            # self.pushRedis(classid, c, Pid)

  第四个py文件

# -*- coding: utf-8 -*-
import re
from urllib.request import urlopen
from scrapy.http import Request
import pymongo
import scrapy
from time import sleep
from scrapy.selector import HtmlXPathSelector
from bson.objectid import ObjectId

client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Chaptername

import redis  # 导入redis数据库

r = redis.Redis(host='127.0.0.1', port=6379, db=0)


class qidianNovelSpider1(scrapy.Spider):
    name = "qidianClass5"
    allowed_domains = ["qidian.com"]
    dict = {}
    start_urls = []

    def __init__(self):  # 定义一个方法

        a = r.lrange('novelnameurl', 0, -1)
        # ii = 0
        for item in a:
            novelnameurl = bytes.decode(item)
            arr = novelnameurl.split(',')  # 分割字符串
            qidianNovelSpider1.start_urls.append(arr[1])
            pid = arr[0]
            url = arr[1]
            self.dict[url] = {"pid":pid}
            print(url)

    def parse(self, response):
        nameInfo = self.dict[response.url]
        pid = nameInfo['pid']
        hxs = HtmlXPathSelector(response)
        hx = hxs.select('//div[@class="volume-wrap"]/div[@class="volume"]/ul[@class="cf"]/li/a[@target="_blank"]')
        for secItem in hx:
            urls = secItem.select("@href").extract()
            url = "https:"+urls[0]
            chapter = secItem.select("text()").extract()

            print(chapter)
            print(url)
            classid = collection.insert({'chaptername': chapter, 'pid': pid})
            self.pushRedis(classid,url, pid)

    def pushRedis(self, classid, url, pid):
        chapterurl = '%s,%s,%s' % (classid, url, pid)
        r.lpush('chapterurl', chapterurl)

  第五个py文件

# -*- coding: utf-8 -*-
import re
from urllib.request import urlopen
from scrapy.http import Request
import pymongo
import scrapy
from time import sleep
from scrapy.selector import HtmlXPathSelector
from bson.objectid import ObjectId

client = pymongo.MongoClient(host="127.0.0.1")
db = client.QiDian
collection = db.Chaptername

import redis  # 导入redis数据库

r = redis.Redis(host='127.0.0.1', port=6379, db=0)


class qidianNovelSpider1(scrapy.Spider):
    name = "qidianClass6"
    allowed_domains = ["qidian.com"]
    dict = {}
    start_urls = []

    def __init__(self):  # 定义一个方法

        a = r.lrange('chapterurl', 0, -1)
        # ii = 0
        for item in a:
            chapterurl = bytes.decode(item)
            arr = chapterurl.split(',')  # 分割字符串
            qidianNovelSpider1.start_urls.append(arr[1])
            pid = arr[0]
            url = arr[1]
            self.dict[url] = {"pid":pid}
            # print(url)


    def parse(self, response):
        nameInfo = self.dict[response.url]
        pid1 = nameInfo['pid']
        pid = ObjectId(pid1)
        hxs = HtmlXPathSelector(response)
        ii=""
        hx = hxs.select('//div[@class="read-content j_readContent"]/p')
        for secItem in hx:
            contents = secItem.select("text()").extract()
            content1 = contents[0]
            # print(content1)
            ii=content1+ii

            # content = bytes(content1,'GBK')
        # classid = collection.insert({'content': ii, 'pid': pid1})
        db.Chaptername.update({"_id": pid}, {"$set": {"content": ii}})
            # print(content)
            # f = open('1.txt','wb')
            # f.write(content)
            # f.close()

  好了 ,大功告成

posted @ 2017-12-02 17:30  都是他舅  阅读(975)  评论(0编辑  收藏  举报