Java解析html

https://www.jianshu.com/p/fd5caaaa950d

1、从URL获取HTML来解析

Document doc = Jsoup.connect("http://www.baidu.com/").get();

2、从磁盘加载html，再解析成txt

import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.select.Elements;

import java.io.*;

public class TEST {
    public static void main(String[] args) {
        try {
            String from = "C:\\Users\\user\\Desktop\\新建文件夹\\novel\\";
            //目标txt的地址
            String to = "C:\\Users\\user\\Desktop\\新建文件夹\\novel\\xx.txt";
            BufferedWriter bw = new BufferedWriter(new OutputStreamWriter(new FileOutputStream(to)));
            for (int i = 0; i < 62; i++) {
                //把html文件转换成String类型
                String str1 = from + (i + 1) + ".html";
                BufferedReader bufferedReader = new BufferedReader(new FileReader(str1));
                StringBuilder stringBuilder = new StringBuilder();
                String str;
                while ((str = bufferedReader.readLine()) != null) {
                    stringBuilder.append(str);//把磁盘中的html拼接成一个字符串
                }

                //给小说添加每章题目
                bw.newLine();
                bw.newLine();
                bw.newLine();
                bw.newLine();
                bw.write("第"+(i+1)+"章");
                bw.newLine();//回车

                String html = new String(stringBuilder);
                //Jsoup解析html
                Document document = Jsoup.parse(html);
                //我的小说html里固定有16个<dd data-id="13">标签，属性名data-id=0~14是有效标签
                for (int j = 0; j < 15; j++) {
                    //通过指定的标签属性名=data-id和属性值=i获得标签，<dd data-id="i>
                    Elements elements = document.getElementsByAttributeValue("data-id", String.valueOf(j));
                    for (Element element : elements) {
                        //通过标签名字<p></p>来获取获得标签，一个<p>小说换行一次
                        Elements p = element.getElementsByTag("p");
                        for (Element element1 : p) {
                            bw.write("        ");
                            bw.write(element1.text());//element1.text()获得标签里的文本
                            bw.newLine();
                            bw.flush();
                            System.out.println(element1.text());
                        }
                    }
                }

                bufferedReader.close();
            }

            bw.close();
            /*
            //通过标签的属性名来获取内容 <dd data-id="13">
            Elements elements = document.getElementsByAttribute("data-id");
            int size = elements.size();//正文里面共有几个data-id
            for (int i = 0; i < size - 1; i++) {
                //通过指定的属性名=data-id和属性值=13获得标签，<dd data-id="13">
                Elements target = document.getElementsByAttributeValue("data-id", String.valueOf(i));

                bw.write(target.text());
                bw.newLine();
                bw.flush();
                System.out.println(target.text());
            }*/

        } catch (FileNotFoundException e) {
            e.printStackTrace();
        } catch (IOException e) {
            e.printStackTrace();
        }
    }
}