Java String Normalize normalize(String text)

Description

normalize

License

Open Source License

Declaration

public static String normalize(String text) 

Method Source Code

//package com.java2s;

import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;

public class Main {
    private final static List<String> words = new ArrayList<String>();

    public static String normalize(String text) {
        StringBuilder builder = new StringBuilder();
        int index = 0;
        int quote = 0;
        text = text.trim();//w  w  w .j  a v  a  2s . c o m
        while (index < text.length()) {
            char c = text.charAt(index);
            if (Character.isWhitespace(c) || Character.isSpaceChar(c))
                c = ' ';
            switch (c) {
            case '+':
            case '-':
            case '&':
            case '|':
            case '!':
            case '(':
            case ')':
            case '{':
            case '}':
            case '[':
            case ']':
            case '^':
            case '~':
            case '*':
            case '?':
            case ':':
            case '\\':
            case '=':
            case '\'':
                builder.append('\\').append(c);
                break;
            case '\"':
                builder.append(c);
                quote++;
                break;
            default:
                /*if(Character.isLetterOrDigit(c)
                || Character.isWhitespace(c)
                || Character.isSpaceChar(c) || c < 256)*/ builder.append(c);
                break;
            }
            index++;
        }
        if (quote % 2 == 1)
            builder.append('\"');

        //    System.out.println(" normalize text: "+ builder);

        if (builder.length() < 30 && words.size() > 0) {
            //      System.out.println("compile text : "+ compile(builder.toString()));
            return compile(builder.toString());
        }
        return builder.toString();
    }

    static String compile(String text) {
        StringBuilder builder = new StringBuilder(text);
        StringBuilder lower = new StringBuilder(text.toLowerCase());
        Iterator<String> iterator = words.iterator();
        while (iterator.hasNext()) {
            String key = iterator.next();
            replace(builder, lower, key);
        }

        return builder.toString();
    }

    private static void replace(StringBuilder builder, StringBuilder lower, String key) {
        int index = lower.indexOf(key);
        //    System.out.println(" kiem key "+ key + " : "+ index);
        String value = new StringBuilder("\"").append(key).append("\"").toString();
        while (index > -1) {
            if (index > 0 && Character.isLetterOrDigit(lower.charAt(index - 1))) {
                index = lower.indexOf(key, index + value.length());
                continue;
            }

            if ((index + key.length()) < lower.length() - 1
                    && Character.isLetterOrDigit(lower.charAt(index + key.length()))) {
                index = lower.indexOf(key, index + value.length());
                continue;
            }
            builder.replace(index, index + key.length(), value);
            lower.replace(index, index + key.length(), value.toLowerCase());
            index = lower.indexOf(key, index + value.length());
        }
    }
}

Related

1. normalize(String string)
2. normalize(String string)
3. normalize(String string)
4. normalize(String text)
5. normalize(String text)
6. normalize(String text)
7. normalize(String uri)
8. normalize(String value)
9. normalize(String value, Locale locale)