妹子图爬虫

-- coding:utf-8 --

import requests
import os
import re
import time
import threading
from lxml import etree
from bs4 import BeautifulSoup
from multiprocessing import Pool, cpu_count

HEADERS = {
'X-Requested-With': 'XMLHttpRequest',
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; WOW64) AppleWebKit/537.36 '
'(KHTML, like Gecko) Chrome/56.0.2924.87 Safari/537.36',
'Referer': 'http://www.mzitu.com'
}

page_urls = ['http://www.mzitu.com/page/{cnt}'.format(cnt=cnt) for cnt in range(1, int(bs) + 1)]

DIR_PATH = "H:/mzitu"

def get_url():
    index_url = "https://www.mzitu.com"
    bs = etree.HTML(requests.get(url=index_url, headers=HEADERS).text).xpath(
    '/html/body/div[2]/div[1]/div[2]/nav/div/a[4]/text()')[0]
    print(bs)
    page_urls = ['http://www.mzitu.com/page/{cnt}'.format(cnt=cnt) for cnt in range(1, int(bs) + 1)]
    img_urls = []
    for i in page_urls:
        print('正在获取' + i + '链接图片地址')
        try:
            bs = BeautifulSoup(requests.get(url=i, headers=HEADERS, timeout=10).text, 'lxml').find('ul', id='pins')
            res = re.findall(r'href="(.*?)" target="_blank"><img', str(bs))
            img_url = [url.replace('"', "") for url in res]
            img_urls.extend(img_url)
        except Exception as e:
            print(e)
    return set(img_urls)

lock = threading.Lock()

def urls_crawler(url):

r = requests.get(url, headers=HEADERS, timeout=10).text

img_name = etree.HTML(r).xpath('//div[@class="main-image"]/p/a/img/@alt')[0]
print(img_name)
# with lock:
if mark_dir(img_name):
    max_count = etree.HTML(r).xpath('//div[@class="pagenavi"]/a[5]/span/text()')[0]
    page_url = [url + '/{cnt}'.format(cnt=cnt) for cnt in range(1, int(max_count) + 1)]
    img_urls = []
    for i, j in enumerate(page_url):
        time.sleep(0.3)
        r = requests.get(j, headers=HEADERS, timeout=10).text
        img_url = etree.HTML(r).xpath('//div[@class="main-image"]/p/a/img/@src')[0]
        img_urls.append(img_url)
    for cnt, url in enumerate(img_urls):
        save_pic(cnt, url)


def save_pic(cnt, url):

try:
    img = requests.get(url, headers=HEADERS, timeout=10).content
    img_name = '{}.jpg'.format(cnt)
    with open(img_name, 'ab') as f:
        f.write(img)
except Exception as e:
    print(e)


def mark_dir(flot_name):

PATH = os.path.join(DIR_PATH, flot_name)
if not os.path.exists(PATH):  
    os.makedirs(PATH)
    os.chdir(PATH)
    return True
print("Folder has existed! {}".format(flot_name))
return False


def delete_empty_dir(save_dir):

if os.path.exists(save_dir):  
    if os.path.isdir(save_dir):  
        for i in os.listdir(save_dir):  
            path = os.path.join(save_dir, i)  
            if os.path.isdir(path):  
                delete_empty_dir(path)  
    if not os.listdir(save_dir):  
        try:
            os.rmdir(save_dir)
        except:
            pass


if name == 'main':
    stratTime = time.time()
    urls = get_url()
    pool = Pool(processes=cpu_count())
    try:
    delete_empty_dir(DIR_PATH)
    pool.map(urls_crawler, urls)
    except Exception:
    time.sleep(30)
    delete_empty_dir(DIR_PATH)
    try:
    pool.map(urls_crawler, urls)
    except:
    pass
    stopTime = time.time()
    print(int(stopTime) - int(stratTime))

 

☆版权☆

* 网站名称:obaby@mars
* 网址:https://h4ck.org.cn/
* 个性:https://oba.by/
* 本文标题: 《妹子图爬虫》
* 本文链接:https://h4ck.org.cn/2021/05/8093
* 短链接:https://oba.by/?p=8093
* 转载文章请标明文章来源,原文标题以及原文链接。请遵从 《署名-非商业性使用-相同方式共享 2.5 中国大陆 (CC BY-NC-SA 2.5 CN) 》许可协议。


猜你喜欢:

发表回复

您的电子邮箱地址不会被公开。 必填项已用*标注