[爬虫] demo 以及使用etree示例

Posted 2023-02-13 长虹剑

记载一下使用的爬虫代码
主要是活用 etree

文章目录

• etree 多跳转解析
• • 讲解
• 普通网站爬取
• • 讲解
• 使用 chrome 分析网站
• 小的实例汇总

etree 多跳转解析

from _chj.comm.pic import *
import requests
from lxml import etree
import pandas as pd
import urllib.parse
from requests_html import HTMLSession

session = HTMLSession()

class params:
    base_url="https://www.fangpi.net"
    sbase_url="https://www.fangpi.net/s/"
    fsing="tmp/sing.txt"
    furl_out="tmp/sing_url.sh"
    dsong = "tmp/songs"

def main():
    f1_get_files()
    f2_mk_wget()

def f2_mk_wget():
    chj_file.mkdir( params.dsong )
    exec_cmd( f" bash params.furl_out " )

def f1_get_files():
    fp = open(params.furl_out, "w")
    for i, line in enumerate( readlines( params.fsing ) ):
        arr = line.split()
        if len(arr) == 1:
            nm, uname = line, None
        else:
            nm, uname = line.split()
            uname = uname.strip()
            outurl, nm = get_link( nm.strip(), uname )
        if outurl is not None:
            #fp.write( f"i outurl nm\\n")
            ss = f"wget -c 'outurl' -O params.dsong/i-nm\\n"
            fp.write( ss )

def get_link( query_nm, query_uname ):
    url = params.sbase_url +  urllib.parse.quote( query_nm )

    r = requests.get(url)
    tree = etree.HTML( r.text )
	r.close()
    items = tree.xpath(".//table[@class='table']/tbody/tr")

    for tr in items:
        #tds = tr.xpath(".//td")
        nm = tr[0].xpath("./a/text()")[0].strip()
        unm = tr[1].text
        urlsing = tr[2].xpath("./a/@href")[0]
        #print( nm, unm, urlsing )
        if nm == query_nm:
            if query_uname is None or unm == query_uname:
                url_sing = params.base_url + urlsing
                href, download = get_sing_url( url_sing )
                return href, download
    return None, None
    #df = pd.read_html(url, encoding='utf-8',header=0)
    # 这个会失去链接

def get_sing_url( url ):
    r = session.get(url) # 必须要渲染否则不行
    r.html.render()
    tree = etree.HTML( r.html.html )
    r.close()
    #print( r.html.html )
    a = tree.xpath(".//div[@class='input-group-append']/a[@id='btn-download-mp3']")[0]
    href = a.xpath("./@href")[0]
    download = a.xpath("./@download")[0]
    #print( href, download )
    #r = requests.get( href )
    return href, download

if __name__ == "__main__":
    main()
    

讲解

先获得list中合适的条目，然后解析获得最终音频的位置

普通网站爬取

先放上整体代码，再讲解

# -*- coding:utf-8 -* 
import urllib.request
from lxml import etree

root_url="http://www.itangyuan.com/"    
url=f"root_url/book/catalogue/14432108.html"
def main(): 
	
	# 第一步
    html = urllib.request.urlopen(url).read()
    tree = etree.HTML(html)
    
    links = tree.xpath(".//div[@class='catalog']/ul/li/a/@href")[2:]
    
    # 第二部
    for i, link in enumerate( links ):
        link = f"root_urllink"

        html = urllib.request.urlopen(link).read()
        tree = etree.HTML(html).xpath(".//div[@class='section-main-con']")
        
        if len(tree) == 0: continue
        if len(tree) != 1: p("WARNING", link, len(len(tree)))
        tree = tree[0]
        
        title = tree.xpath(".//h1/text()")[0]
        content = "\\n\\n".join( tree.xpath(".//p/text()")[:-1] )
        
        with open(f"res/i+1:03d.title.txt", "w", encoding='utf-8') as fp:
            fp.write(content)
        
# 下面这个函数忽略
def main2(): 
    with open("虹猫蓝兔七侠传小说.md", "w", encoding="utf-8") as fpout:
        for fnm in glob.glob("res/*"):
            with open(fnm, encoding="utf-8") as fp:
                nm = fnm.split('\\\\')[1].split('.txt')[0]
                fpout.write(f"# nm\\n")
                fpout.write(fp.read()+"\\n")
 

if __name__ == '__main__': 
    main() 

讲解

1. 分析网页 http://www.itangyuan.com/book/catalogue/14432108.html ， 确定每个章节的网址
2. 加载每个网址，分析其中一个页面，获得相应的内容

由于这个网页内容比较简单，直接 xpath 索引就全部获得了。

注意事项：
1）每次 xpath 完是个数组，可能应为是通过 class 索引缘故，未详细探索
2）用 utf-8 保存

使用 chrome 分析网站

想爬取一些数据
比如
https://tv.cctv.com/2019/12/31/VIDEOX9ykqMX1J0rlAhEmjeo191231.shtml

chrome 中分析发现一个链接

使用 ffmpeg 下载之后发现有问题 ( ffmpeg -i xx.m3u8 -c copy demo.mp4 ), 视频是模糊的

进一步分析网络包
先 clear 一下

刷新一下页面，分析包，然后发现下面那个请求，获得一个本质上是 json 的 url

可以在这个 url 中 分析得到

不过chrome 默认是没有装 jsonview插件的可以装一下

jsonView

小的实例汇总

lyric

divs = tree.xpath(".//div[@label-module='para*']")
for e in divs: print( e.text)