Text Extraction from HTML Java

Question

I'm working on a program that downloads HTML pages and then selects some of the information and write it to another file.

I want to extract the information which is intbetween the paragraph tags, but i can only get one line of the paragraph. My code is as follows;

FileReader fileReader = new FileReader(file);
BufferedReader buffRd = new BufferedReader(fileReader);
BufferedWriter out = new BufferedWriter(new FileWriter(newFile.txt));
String s;

while ((s = br.readLine()) !=null) {
    if(s.contains("<p>")) {
        try {
            out.write(s);
        } catch (IOException e) {
        }
    }
}

i was trying to add another while loop, which would tell the program to keep writing to file until the line contains the </p> tag, by saying;

while ((s = br.readLine()) !=null) {
    if(s.contains("<p>")) {
        while(!s.contains("</p>") {
            try {
                out.write(s);
            } catch (IOException e) {
            }
        }
    }
}

But this doesn't work. Could someone please help.

Undetermined

HTML parsers do exist and there are plenty of them. – Play

Answer 1

jsoup

Another html parser I really liked using was jsoup. You could get all the <p> elements in 2 lines of code.

Document doc = Jsoup.connect("http://en.wikipedia.org/").get();
Elements ps = doc.select("p");

Then write it out to a file in one more line

out.write(ps.text());  //it will append all of the p elements together in one long string

or if you want them on separate lines you can iterate through the elements and write them out separately.

Dia

If a document doesn't use p tags (non-semantic mark up), I assume this won't work – Tip

@sinθ The Question explicitly asked for p elements. This answer is spot-on correct. – Roseleeroselia

Answer 2

jericho is one of several posible html parsers that could make this task both easy and safe.

Riffle

Answer 3

JTidy can represent an HTML document (even a malformed one) as a document model, making the process of extracting the contents of a <p> tag a rather more elegant process than manually thunking through the raw text.

Coincide

Answer 4

Try (if you don't want to use a HTML parser library):

        FileReader fileReader = new FileReader(file);
        BufferedReader buffRd = new BufferedReader(fileReader);
        BufferedWriter out = new BufferedWriter(new FileWriter(newFile.txt));
        String s;
        int writeTo = 0;
        while ((s = br.readLine()) !=null) 
        {
                if(s.contains("<p>"))
                {
                        writeTo = 1;

                        try 
                        {
                            out.write(s);
                    } 
                        catch (IOException e) 
                        {

                    }
                }
                if(s.contains("</p>"))
                {
                        writeTo = 0;

                        try 
                        {
                            out.write(s);
                    } 
                        catch (IOException e) 
                        {

                    }
                }
                else if(writeTo==1)
                {
                        try 
                        {
                            out.write(s);
                    } 
                        catch (IOException e) 
                        {

                    }
                }
}

Umbilication

What happens if the <p> and </p> are on the same line? In this case the string will be written out twice. I guess it really depends on the input. – Desmund

You could add some state to see if you have already written out the line before writing it out again. – Desmund

Answer 5

I've had success using TagSoup & XPath to parse HTML.

http://home.ccil.org/~cowan/XML/tagsoup/

Hillhouse

Answer 6

Use a ParserCallback. Its a simple class thats included with the JDK. It notifies you every time a new tag is found and then you can extract the text of the tag. Simple example:

import java.io.*;
import java.net.*;
import javax.swing.text.*;
import javax.swing.text.html.*;
import javax.swing.text.html.parser.*;

public class ParserCallbackTest extends HTMLEditorKit.ParserCallback
{
    private int tabLevel = 1;
    private int line = 1;

    public void handleComment(char[] data, int pos)
    {
        displayData(new String(data));
    }

    public void handleEndOfLineString(String eol)
    {
        System.out.println( line++ );
    }

    public void handleEndTag(HTML.Tag tag, int pos)
    {
        tabLevel--;
        displayData("/" + tag);
    }

    public void handleError(String errorMsg, int pos)
    {
        displayData(pos + ":" + errorMsg);
    }

    public void handleMutableTag(HTML.Tag tag, MutableAttributeSet a, int pos)
    {
        displayData("mutable:" + tag + ": " + pos + ": " + a);
    }

    public void handleSimpleTag(HTML.Tag tag, MutableAttributeSet a, int pos)
    {
        displayData( tag + "::" + a );
//      tabLevel++;
    }

    public void handleStartTag(HTML.Tag tag, MutableAttributeSet a, int pos)
    {
        displayData( tag + ":" + a );
        tabLevel++;
    }

    public void handleText(char[] data, int pos)
    {
        displayData( new String(data) );
    }

    private void displayData(String text)
    {
        for (int i = 0; i < tabLevel; i++)
            System.out.print("\t");

        System.out.println(text);
    }

    public static void main(String[] args)
    throws IOException
    {
        ParserCallbackTest parser = new ParserCallbackTest();

        // args[0] is the file to parse

        Reader reader = new FileReader(args[0]);
//      URLConnection conn = new URL(args[0]).openConnection();
//      Reader reader = new InputStreamReader(conn.getInputStream());

        try
        {
            new ParserDelegator().parse(reader, parser, true);
        }
        catch (IOException e)
        {
            System.out.println(e);
        }
    }
}

So all you need to do is set a boolean flag when the paragraph tag is found. Then in the handleText() method you extract the text.

Thuja

Answer 7

Try this.

 public static void main( String[] args )
{
    String url = "http://en.wikipedia.org/wiki/Big_data";

    Document document;
    try {
        document = Jsoup.connect(url).get();
        Elements paragraphs = document.select("p");

        Element firstParagraph = paragraphs.first();
        Element lastParagraph = paragraphs.last();
        Element p;
        int i=1;
        p=firstParagraph;
        System.out.println("*  " +p.text());
        while (p!=lastParagraph){
            p=paragraphs.get(i);
            System.out.println("*  " +p.text());
            i++;
        } 
} catch (IOException e) {
    // TODO Auto-generated catch block
    e.printStackTrace();
}
}

Tocsin

What is this 'Element' and 'Document'. Is this any third party parser? Show the import lines too – Symptomatology

Answer 8

You may just be using the wrong tool for the job:

perl -ne "print if m|<p>| .. m|</p>|" infile.txt >outfile.txt

Microreader