Python爬取网站文章并自动发布到wordpress博客网站

一、抓取数据

# -*- coding:utf-8 -*-
import json
import sys
import base64
from typing import Optional

import requests
from pydantic import BaseModel
from lxml import etree
from loguru import logger

logger.remove()

logger.add(
    sys.stdout,
    colorize=True,
    diagnose=False,
    format='<green>{time:MM-DD HH:mm:ss}</green>|{level}|<yellow>{name}:{line}</yellow>|<level>{message}</level>')


class Article(BaseModel):
    title: Optional[str]
    category: Optional[str]
    tag: Optional[str]
    author: Optional[str]
    content: Optional[str]
    date: Optional[str]


class LikeCs:

    def __init__(self):
        self.session = requests.session()

        self.articles = {}
        self.host = 'https://www.likecs.com'

    def get_tree(self, url):
        resp = self.session.get(url)
        return etree.HTML(resp.text)

    def get_category(self, page_url):
        logger.info('正在获取页面分类：{}'.format(page_url))
        tree = self.get_tree(page_url)
        lst = []
        for li in tree.xpath('//ul[@class="cardcont navsubs"]/li'):
            category_url = self.host + li.xpath('./a/@href')[0]
            category = li.xpath('./a/text()')[0].strip()
            item = {
                'name': category,
                'url': category_url
            }
            self.articles[category] = []
            logger.info(item)
            lst.append(item)
        logger.info('页面：{} 分类获取完毕！'.format(page_url))
        return lst

    def get_list(self, category):
        tree = self.get_tree(category['url'])
        # pages = tree.xpath()
        pages = 25
        for page in range(1, pages + 1):

            logger.info('正在爬取分类：{}  当前第{}页，总共{}页！'.format(category['name'], page, pages))
            if page == 1:
                tree2 = tree
            else:
                tree2 = self.get_tree(category['url'] + '?page={}'.format(page))

            for item in tree2.xpath('//article[@class="post tag-uifont tag-webview"]'):
                article = Article()
                article.category = category['name'].strip()
                try:
                    article.title = item.xpath('./header/h2/a/text()')[0].replace('\n', '').strip()
                except:
                    article.title = ''
                try:
                    article.author = item.xpath('./footer/text()')[0].strip()
                except:
                    article.author = ''
                try:
                    article.tag = item.xpath('./footer/a/text()')[0].strip()
                except:
                    article.tag = ''
                try:
                    article.date = item.xpath('./footer/time/text()')[0].strip()
                except:
                    article.date = ''

                article_url = self.host + item.xpath('./header/h2/a/@href')[0]
                a = self.get_article(article, article_url)
                logger.info('文章：{}  爬取成功！'.format(a.title))
                self.articles[a.category].append(a.dict())
                # self.publish_article_to_wordpress(a)

    def get_article(self, article: Article, article_url):
        tree = self.get_tree(article_url)
        article.content = tree.xpath('//section[@class="post-content"]')[0].xpath('string(.)').strip()

        return article

    def main(self):

        page_url = [
            'https://www.likecs.com/nav/1.html',
            # 'https://www.likecs.com/nav/9.html',
            # 'https://www.likecs.com/nav/31.html',
            # 'https://www.likecs.com/nav/41.html',
            # 'https://www.likecs.com/nav/52.html',
            # 'https://www.likecs.com/nav/68.html',
            # 'https://www.likecs.com/nav/72.html'
        ]
        for url in page_url:
            for category in self.get_category(url)[0:1]:
                self.get_list(category)
                logger.debug('-' * 50)
            logger.debug('-' * 50)

        with open('./article.json', 'w') as f:
            json.dump(self.articles, f, ensure_ascii=False)

二、配置wordpress

WordPress配置

if (-f $request_filename/index.html){
rewrite (.*) $1/index.html break;
}
if (-f $request_filename/index.php){
rewrite (.*) $1/index.php;
}
if (!-f $request_filename){
rewrite (.*) /index.php;
}
rewrite /wp-admin$ $scheme://$host$uri/ permanent;

WordPress插件

WP REST API
WordPress REST API Authentication

三、发布到WordPress中

#  发布文章到WordPress
def publish_article_to_wordpress(article: Article):
    session = requests.session()
    # 填写WordPress域名，管理端账号密码
    wp_host = 'xxx.xxx.xxx.xxx'
    wp_username = 'admin'
    wp_password = '123456'

    basic_auth = str(base64.b64encode('{user}:{passwd}'.format(user=wp_username, passwd=wp_password).encode()),
                     'utf-8')
    api = 'http://{wp_host}/wp-json/wp/v2/posts'.format(wp_host=wp_host)
    headers = {
        'Authorization': 'Basic {basic_auth}'.format(basic_auth=basic_auth)
    }
    data = {
        'title': article.title,
        'content': article.content,
        'categories': [3],
        'status': 'publish'
    }
    resp = session.post(api, headers=headers, data=data)
    if resp.status_code == 201:
        logger.info('文章：{}  发布成功！'.format(article.title))
    else:
        logger.error('文章：{}  发布发布失败！{}'.format(article.title, resp.json()))