小工具随手记：爬取蓝多多的小仓库CSDN发布的原创文章信息

写在前面：

本文已获得原博主授权转载，转载链接：Python爬虫分析CSDN个人博客数据_Caso_卡索的博客-CSDN博客

代码来源Caso_卡索博主的github：https://github.com/xiaoma101017/ParseCSDNBlog

转载目的仅供存档学习及日常使用，请勿做商业用途。

Step one：检查网页源代码：

爬取代码运行过程：

爬取结果：

附：完整代码（有部分修改）

import requests
from bs4 import BeautifulSoup
import pandas as pd
import os
import re
Host = "lddwarehouse.blog.csdn.net" # 请求头host参数
User_Agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/83.0.4103.61 Safari/537.36 "
Source = 'html.txt'   # 临时保存博客列表html源码
EachSource = 'each.txt' # 临时保存每篇博客html源码
OUTPUT = "蓝多多博客汇总.csv"  # 输出博客信息到 csv 文件
results = [] # 存储全部数据
#相关函数定义
def parseEachBlog(link):
    referer = "Referer: " + link
    headers = {"Referer": referer, "User-Agent": User_Agent}
    r = requests.post(link, headers=headers)
    html = r.text
    with open(EachSource, 'w', encoding='UTF-8') as f:
        f.write(html)
    soup = BeautifulSoup(open(EachSource, 'r', encoding='UTF-8'), features="html.parser")
    readcontent = soup.select('.bar-content  .read-count')
    collection = soup.select('.bar-content .get-collection')
    readcounts = re.sub(r'\D', "", str(readcontent[0]))
    collections = re.sub(r'\D', "", str(collection[0]))
    print(soup.select('.title-article')[0])
    blogname = soup.select('.title-article')[0].text
    time = soup.select('.bar-content .time')[0].text
    programa = soup.select('.blog-tags-box .tag-link')[0].text
    print(programa)
    type = soup.select('.text-center .font')[0].text #这里代码有误，需要用到下面的图像识别
    print(type)
    eachBlog = [blogname, link, readcounts, collections, time, programa, type]
    return eachBlog
def getBlogList(pages):
    listhome = "https://" + Host + "/article/list/"
    pagenums = [] # 转换后的pages页数
    for i in range(1, int(pages)+1):
        pagenums.append(str(i))

    for number in pagenums:
        url = listhome + number + "?t=1"
        headers = {"Referer": url, "Host": Host, "User-Agent": User_Agent}
        response = requests.post(url, headers=headers)
        html = response.text
        with open(Source, 'a', encoding='UTF-8') as f:
            f.write(html)
    # 获取全部博客的链接
    soup = BeautifulSoup(open(Source, 'r', encoding='UTF-8'), features="html.parser")
    hrefs = []
    re_patterm = "^https://lddwarehouse.blog.csdn.net/article/details/\d+$"
    for a in soup.find_all('a', href=True):
        if a.get_text(strip=True):
            href = a['href']
            if re.match(re_patterm, href):
                if hrefs.count(href) == 0:
                    hrefs.append(href)
    return hrefs
def parseData():
    results.sort(key=lambda result:int(result[2]), reverse=True) # 按浏览量排序
    dataframe = pd.DataFrame(data=results)
    dataframe.columns = ['文章标题', '文章链接', '浏览量', '收藏量', '发布时间','第一所属专栏','文章类型']
    dataframe.to_csv(OUTPUT, index=False, sep=',',encoding='gbk')
def delTempFile():
    if os.path.exists(Source):
        os.remove(Source)
    if os.path.exists(EachSource):
        os.remove(EachSource)

if __name__ == '__main__':
    pages = input("请输入博客列表页数: ")
    linklist = getBlogList(pages)
    print("开始获取数据...")
    for i in linklist:
        print("当前获取: %s"%(i))
        results.append(parseEachBlog(i))
    parseData()
    delTempFile()

附：文章类型的图像识别

川川大佬：python菜鸟_川川菜鸟_CSDN博客 帮蓝多多解答了关于文章类型是图片如何处理的问题，非常感谢！

Step one ：先使用soup.select获取该篇文章类型的图片描述：

Step two ：使用save函数将这个图保存到本地（在下面的识别完成以后再删除）

 Step three：读取图片并提取文字：

import ddddocr

ocr = ddddocr.DdddOcr()
with open('c.png', 'rb') as f:
    img_bytes = f.read()
res = ocr.classification(img_bytes)

print(res)

即可获得该文章类型的text：

交流：

（代码还没有修改，有兴趣可以自己先试试嗷）