stack usage example (html parser)

parser.java

FileWorker.java:

package com.alex.stack.htmlparser;

import java.io.BufferedReader;
import java.io.FileReader;
import java.io.File;
import java.io.FileWriter;
import java.io.IOException;
import java.util.Vector;

class FileWorker extends File {

    public FileWorker(String name) {
        super(name);
    }

    public void erase() {
        try {
            try (FileWriter file = new FileWriter(this, false)) {
                file.write("");
            }
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
    }

    public boolean write(String string) {
        try {
            try (FileWriter file = new FileWriter(this, true)) {
                file.write(string);
            }
        } catch (IOException e) {
            return false;
        }
        return true;
    }

    public boolean writeln(String string) {
        return write(string + "\n");
    }

    public Vector<String> readLines() {
        Vector<String> result = new Vector<>();
        try {
            try (BufferedReader file = new BufferedReader(new FileReader(this))) {
                String line;
                while ((line = file.readLine()) != null) {
                    result.add(line + "\n");
                }
            }
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
        return result;
    }

    public StringBuilder readAll() {
        StringBuilder result = new StringBuilder();
        try {
            try (BufferedReader file = new BufferedReader(new FileReader(this))) {
                String line;
                while ((line = file.readLine()) != null) {
                    result.append(line).append("\n");
                }
            }
        } catch (IOException e) {
            throw new RuntimeException(e);
        }
        return result;
    }

    @Override
    public String toString() {
        return readAll().toString();
    }

}

====================================================================================================================

UniversalStack.java:

class UniversalStack<Type> {

    boolean repeatEnabled = false;
    private Vector<Type> data;

    public UniversalStack() {
        data = new Vector<>(0);
    }

    public UniversalStack(boolean repeatEnabled) {
        this.repeatEnabled = repeatEnabled;
        data = new Vector<>(0);
    }

    public void clear() {
        data.clear();
    }

    public boolean isEmpty() {
        return data.isEmpty();
    }

    public int getCount() {
        return data.size();
    }

    public void push(Type data) {
        if (repeatEnabled && contains(data)) {
            return;
        }
        this.data.add(data);
    }

    public Type pop() {
        if (!isEmpty()) {
            Type temp = data.lastElement();
            data.removeElementAt(data.size() - 1);
            return temp;
        }
        return null;
    }

    public Type peek() {
        if (!isEmpty()) {
            return data.lastElement();
        }
        return null;
    }

    public Type peekEqual() {
        if (!isEmpty()) {
            return data.lastElement();
        }
        return null;
    }

    public boolean contains(Type data) {
        for (int i = 0; i < this.data.size(); i++) {
            if (this.data.elementAt(i).equals(data)) {
                return true;
            }
        }
        return false;
    }

    @Override
    public String toString() {
        StringBuilder result = new StringBuilder();
        for (Type data1 : data) {
            result.append(data1.toString()).append("\n");
        }
        return result.toString();
    }
}

====================================================================================================================

HTMLTags.java:

class HTMLTags {

    private UniversalStack<HTMLTag> tagsStack = new UniversalStack<HTMLTag>(true);

    private Vector<HTMLTag> dualTags = new Vector<>();
    private Vector<HTMLTag> singleTags = new Vector<>();

    public HTMLTags() {

        FileWorker dualTagsFile = new FileWorker("C:\\1\\dualtags.txt"); // !!! MAYBE YOU NEED TO CHANGE PATH TO FILES !!!
        FileWorker singleTagsFile = new FileWorker("C:\\1\\singletags.txt");
        Vector<String> tagsList;

        tagsList = dualTagsFile.readLines();
        for (String tagsList1 : tagsList) {
            dualTags.add(new HTMLTag(tagsList1));
        }

        tagsList = singleTagsFile.readLines();
        for (String tagsList1 : tagsList) {
            singleTags.add(new HTMLTag(tagsList1, true));
        }
    }

    public boolean dualTagsContains(String tag) {
        for (HTMLTag dualTag : dualTags) {
            if (dualTag.getName().equals(new HTMLTag(tag).getName())) {
                return true;
            }
        }
        return false;
    }

    public boolean singleTagsContains(String tag) {
        for (HTMLTag singleTag : singleTags) {
            if (singleTag.getName().equals(new HTMLTag(tag).getName())) {
                return true;
            }
        }
        return false;
    }

    public String parseTags(String htmlCode) {
        String tag;
        tagsStack.clear();
        htmlCode = htmlCode.toLowerCase();
        try {
            while (htmlCode.contains("<") && htmlCode.contains(">")) {
                try {
                    tag = htmlCode.substring(htmlCode.indexOf("<"), htmlCode.indexOf(">") + 1);
                } catch (IndexOutOfBoundsException e) {
                    throw new Exception("Неправильно расставлены скобки '<' и '>'");
                }
                htmlCode = htmlCode.replaceFirst(tag, "[xxx]");
                if (dualTagsContains(tag)) {
                    HTMLTag tempTag = new HTMLTag(tag);
                    if (tempTag.getOpeningTag().equals((tag.contains(" ")) ? tag.substring(0, tag.indexOf(" ")) + ">" : tag)) {
                        tagsStack.push(tempTag);
                    } else if (tempTag.getClosingTag().equals(tag)) {
                        if (tagsStack.peek().getName().equals(tempTag.getName())) {
                            tagsStack.pop();
                        } else {
                            throw new Exception("Вначале необходимо открыть тег, а лишь потом закрывать: " + tempTag);
                        }
                    } else {
                        throw new Exception("Недопустимый тег: " + tag);
                    }
                } else if (singleTagsContains(tag)) {
                    if (tag.charAt(1) == '/') {
                        throw new Exception("Закрывающий тег не нужен. Это тег-одиночка: " + new HTMLTag(tag, true));
                    } else {
                    }
                } else {
                    throw new Exception("Такого тега не существует: " + tag);
                }
            }
            if (htmlCode.contains("<") || htmlCode.contains(">")) {
                throw new Exception("Остались лишние скобки '<' или '>'");
            }
            if (tagsStack.getCount() > 0) {
                throw new Exception("Не все теги были закрыты: \n" + tagsStack);
            }
            return "HTML код правильный";
        } catch (Exception e) {
            return e.getMessage();
        }
    }

    @Override
    public String toString() {
        return " Двойные теги: \n" + dualTags.toString()
                + " \n Одиночные теги: \n" + singleTags.toString();
    }

    public static class HTMLTag {

        private String tag;
        private boolean single = false;

        public HTMLTag(String tag) {
            setTag(tag);
        }

        public HTMLTag(String tag, boolean single) {
            setTag(tag, single);
        }

        public boolean setTag(String tag) {
            return setTag(tag, false);
        }

        public boolean setTag(String tag, boolean single) {
            String tempTag;
            if (tag.contains("<") && tag.contains(">")) {
                tempTag = tag.substring(tag.indexOf("<") + 1, tag.indexOf(">"));
                if (tempTag.contains(" ")) {
                    tempTag = tempTag.substring(0, tempTag.indexOf(" "));
                }
                if (tempTag.charAt(0) == '/') {
                    tempTag = tempTag.replace("/", "");
                }
                this.tag = tempTag;
                this.single = single;
                return true;
            } else {
                return false;
            }
        }

        public String getName() {
            return tag;
        }

        public String getOpeningTag() {
            return "<" + tag + ">";
        }

        public String getClosingTag() {
            return "</" + tag + ">";
        }

        @Override
        public String toString() {
            if (single) {
                return getOpeningTag();
            } else {
                return getOpeningTag() + "..." + getClosingTag();
            }
        }

    }

}

====================================================================================================================

Program.java:

class Program { // change class name!

    public static void main(String[] args) {

        String htmlFileName = "C:\\1\\some page.html"; // !!! PATH TO HTML-FILE !!!
        FileWorker htmlFile = new FileWorker(htmlFileName);
        System.out.println(htmlFile);

        HTMLTags htmlTags = new HTMLTags();

        StringBuilder htmlCode = htmlFile.readAll();
        System.out.println(htmlTags.parseTags(htmlCode.toString()));
    }
}

====================================================================================================================

dualtags.txt:

<html>...</html>
<head>...</head>
<title>...</title>
<body>...</body>
<strong>...</strong>
<p>...</p>
<h1>...</h1>
<h2>...</h2>
<h3>...</h3>
<h4>...</h4>
<h5>...</h5>
<h6>...</h6>
<blockquote>...</blockquote>
<q>...</q>
<address>...<address>
<pre>...</pre>
<b>...</b>
<i>...</i>
<tt>...</tt>
<font>...</font>
<ol>...</ol>
<ul>...</ul>
<li>...</li>
<dl>...</dl>
<dt>...</dt>
<dd>...</dd>
<table>...</table>
<td>...</td>
<tr>...</tr>
<th>...</th>
<thead>...</thead>
<caption>...</caption>
<a>...</a>

========================================================================================================================

singletags.txt:

<base>
<meta>
<br>
<hr>
<img>

========================================================================================================================

some page.html:

<html>
<head>
<meta http-equiv="Content-Type" content="text/html" />
<title>Пример страницы с таблицей №1.</title>
</head>
<body>
<table width="100%" border="1" cellspacing="1" cellpadding="1" bgcolor="">
 <tr>
   <th height="91" colspan=3 align="center" bgcolor="#66FF00"><strong>ПРИМЕР НАПИСАНИЯ ТАБЛИЦ</strong></th>
 </tr>
 <tr>
  <th height="47" colspan=3 align="center" bgcolor="#BCEE68">
      <table width="50%" border="2" cellspacing="1" bgcolor="#4876FF">
       <tr>
         <td align="center"><a href="#" style="text-decoration: none"><font color="#FFFFFF">страница 1</font></a></td>
         <td align="center"><a href="#" style="text-decoration: none"><font color="#FFFFFF">страница 2</font></a></td>
         <td align="center"><a href="#" style="text-decoration: none"><font color="#FFFFFF">страница 3</font></a></td>
         <td align="center"><a href="#" style="text-decoration: none"><font color="#FFFFFF">страница 4</font></a></td>
       </tr>
     </table>
 </th>
 </tr>
 <tr>
 <th width="17%" height="498" align="center" bgcolor="#BFEFFF">левый блок</th>
 <th width="65%" align="center" background="images/bek_564.jpg">
    <table width="90%" border="2" cellspacing="1" bordercolor="#0000CC">
       <caption>Пример таблицы для тренировок.</caption>
      <tr>
       <td colspan="3" align="center">ячейка 1</td>
      </tr>
      <tr>
       <td align="center">ячейка 2</td>
       <td align="center">ячейка 3</td>
       <td align="center">ячейка 4</td>
      </tr>
      <tr>
       <td align="center" colspan="2" rowspan="2">ячейка 5</td>
       <td align="center">ячейка 6</td>
      </tr>
      <tr>
       <td align="center">ячейка 7</td>
      </tr>
      <tr>
       <td align="center" colspan="2">ячейка8</td>
       <td align="center">ячейка 9</td>
      </tr>
      <tr>
       <td align="center" colspan="3">ячейка 10</td>
      </tr>
      <tr>
       <td align="center">ячейка 11</td>
       <td align="center" rowspan="3" colspan="3">ячейка 12</td>
     </tr>
     <tr>
      <td align="center">ячейка 13</td>
     </tr>
     <tr>
      <td align="center">ячейка 14</td>
     </tr>
     <tr>
      <td align="center" colspan="3">ячейка 15</td>
     </tr>
   </table>
  <br>
   <p align="center"><a href="http://vk.com/sunmeat">about author</a></p>
  <br>
  </th>
  <th width="18%" align="center" bgcolor="#98F5FF">правый блок</th>
 </tr>
 <tr>
  <th height="49" colspan=3 align="center" bgcolor="#1E90FF">НИЖНИЙ БЛОК</th>
 </tr>
 </table>
</body>
</html>