source: gli/branches/rtl-gli/src/org/greenstone/gatherer/cdm/CommandTokenizer.java@ 18368

/**
 *#########################################################################
 *
 * A component of the Gatherer application, part of the Greenstone digital
 * library suite from the New Zealand Digital Library Project at the
 * University of Waikato, New Zealand.
 *
 * <BR><BR>
 *
 * Author: John Thompson, Greenstone Digital Library, University of Waikato
 *
 * <BR><BR>
 *
 * Copyright (C) 1999 New Zealand Digital Library Project
 *
 * <BR><BR>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * <BR><BR>
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * <BR><BR>
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 *########################################################################
 */
package org.greenstone.gatherer.cdm;

/**************************************************************************************
 * Title:       Gatherer
 * Description:  The Gatherer: a tool for gathering and enriching a digital collection.
 * Company:   The University of Waikato
 * Written:   07/05/02
 * Revised:   03/10/02 - Commented
 **************************************************************************************/
import java.io.BufferedReader;
import java.util.StringTokenizer;
import org.greenstone.gatherer.DebugStream;
import org.greenstone.gatherer.util.StaticStrings;

/** This class provides an extension to the standard StringTokenizer in that it recognizes quotes (or some form of bracketting) enclose a single token so in something like: <br>format Search '&lt;table&gt;&lt;img src=... &lt;/table&gt;'<br> the formatting string is parsed as a single token. Unfortunately this makes countTokens() unreliable for exact measurement of tokens remaining, and only useful for determining if there are tokens left to be processed (includes any that have already been read into command buffer).
 * @author John Thompson, Greenstone Digital Library, University of Waikato
 * @version 2.3
 */
public class CommandTokenizer {
    
    static final public int BRACKET_ENCLOSED        = 0;
    static final public int DOUBLE_QUOTE_ENCLOSED   = 1;
    static final public int NORMAL          = 2;
    static final public int QUOTE_ENCLOSED      = 3;
    
    private BufferedReader in_stream;
    private int count = -1;
    private StringTokenizer internal_tokenizer;
    private boolean strip_quotes = true;
    /** Basic Constructor. Used to parse tokens from a string keeping tokens surrounded by speechmarks or square brackets intact. Thus something like:<br>
     * collectionmeta collectionextra [l = en] "Hello World"<br>
     * is tokenized thus<br>
     * {'collectionmeta', 'collectionextra', 'l = en', 'Hello World'}
     * @param command the command String you wish to tokenize
     */
    public CommandTokenizer(String command) {
    this.internal_tokenizer = new StringTokenizer(command);
    this.in_stream = null;
    }
    
    /** Advanced Constructor. As above but with one major difference. Since it is provided an input stream (presumably where the command string originated from), it is able to parse a quote enclosed command token that stretches over several lines. Each newline is preserved in the resulting token. There is an extra bitchslap here as comething like a collection extra might have html code in them that contain escaped speechmarks, so extra care must be taken not to break at them. Thus something like:<br>
     * collectionmeta collectionextra [l = en] "<br>
     *     an example of the crazy as description we sometimes get which includes of all things something like <a href=\"this.html\"<br>
     *     >this</a> which you could easily see might be a problem if I parse this niavely."<br>
     * is tokenized thus<br>
     * {'collectionmeta', 'collectionextra', 'l = en', '\nan example of the crazy as description we sometimes get which includes of all things something like <a href=\"this.html\"\n>this</a> which you could easily see might be a problem if I parse this niavely.'}
     * @param command the command String you wish to tokenize
     * @param in_stream a BufferedReader from which the tokenizer can draw further lines as necessary
     */
    public CommandTokenizer(String command, BufferedReader in_stream) {
    ///atherer.println("***** CommandTokenizer *****\nparse:\t" + command + "\n****************************");
    this.internal_tokenizer = new StringTokenizer(command);
    this.in_stream = in_stream;
    }
    
    public CommandTokenizer(String command, BufferedReader in_stream, boolean strip_quotes) {
    this.internal_tokenizer = new StringTokenizer(command);
    this.in_stream = in_stream;
    this.strip_quotes = strip_quotes;
    }
    
    /** Returns the minumum number of remaining tokens before the tokenizer runs out of string. There may be more tokens than this count, but never less. The discrepancy is due to internal functionality and the fact we can't read ahead in the string or associated stream without risking the need for unpredictable push-back
     * @return the minumum number of tokens available as an int
     */
    public int countTokens() {
    if(count == 0 && internal_tokenizer.countTokens() > 1) {
        return 1;
    }
    if(count == -1) {
        count = internal_tokenizer.countTokens();
    }
    return count;
    }
    
    /** Determine if there are still tokens available.
     * @return true if there are more tokens, false otherwise
     */
    public boolean hasMoreTokens() {
    return internal_tokenizer.hasMoreTokens();
    }
    
    /** Method to retrieve the next token from the command, taking care to group tokens enclosed in speech marks.
     * @return a String containing the next token from the command
     */
    public String nextToken() {
    String result = null;
    if(internal_tokenizer.hasMoreTokens()) {
        StringBuffer buffer = new StringBuffer(internal_tokenizer.nextToken());
        switch(buffer.charAt(0)) {
        case StaticStrings.DOUBLEQUOTE_CHAR:
        ///ystem.err.println("Building token wrapped by double quotes.");
        result = buildToken(buffer, StaticStrings.DOUBLEQUOTE_CHAR, this.strip_quotes);
        break;
        case StaticStrings.SINGLEQUOTE_CHAR:
        ///ystem.err.println("Building token wrapped by single quotes.");
        result = buildToken(buffer, StaticStrings.SINGLEQUOTE_CHAR, this.strip_quotes);
        break;
        case StaticStrings.OPENBRACKET_CHAR:
        ///ystem.err.println("Building token wrapped by brackets.");
        result = buildToken(buffer, StaticStrings.CLOSEBRACKET_CHAR, false);
        break;
        default:
        ///ystem.err.println("Returning plain string.");
        result = buffer.toString();
        }
        buffer = null;
    }
    // Because of our tricky counting system we never want to have negative tokens remaining. In fact, unless the internal string buffer is empty, we will return a count of 1 anyway
    if(count > 0) {
        count = count - 1;
    }
    ///ystem.err.println("----- CommandTokenizer -----\ntoken:\t" + result + "\n----------------------------");
    return result;
    }
    
    /** Parse in the next token. paying heed to enclosing characters demands, escaped characters, newlines and empty buffers and consequential unexpected end of tokens
     * @param buffer the StringBuffer in which the partial token is stored (at the first bit that caused this method to be called)
     * @param end_char the sentinel char we are watching for as it encloses a token
     * @param strip_characters a boolean denoting whether the enclosing characters should be stripped off
     * @return the token, either in its entirity less the enclosing characters if required or, if an unexpected end occured, whatever we parsed without its starting enclosing character, again only if required. In fact if we weren't asked to strip characters then we add the enclosing character back in
     */
    private String buildToken(StringBuffer buffer, char end_char, boolean strip_characters) {
    while(buffer.length() == 1 || buffer.charAt(buffer.length() - 1) != end_char || (buffer.length() > 3 && buffer.charAt(buffer.length() - 2) == StaticStrings.BACKSLASH_CHAR)) {
        try {
        // The first version is for the basic tokenizer which has no idea of an input stream, so runs out tokens at the same time as the internal tokenizer does
        if(internal_tokenizer.hasMoreTokens()) {
            buffer.append(StaticStrings.SPACE_CHAR);
            buffer.append(internal_tokenizer.nextToken());
        }
        // While the second version can draw more lines from the stream until eof occurs
        else if(in_stream != null) {
            String line_str = null;
            while(!internal_tokenizer.hasMoreTokens() && (line_str = in_stream.readLine()) != null) {
            ///atherer.println("+++++ CommandTokenizer +++++\nappend:\t" + line_str + "\n+++++++++++++++++++++++++++++");
            // Its at this stage the our token count becomes completely putu
            internal_tokenizer = new StringTokenizer(line_str);
            buffer.append(StaticStrings.NEW_LINE_CHAR); // A new line in the final token
            }
            line_str = null;
            if(internal_tokenizer.hasMoreTokens()) {
            // Don't add a space if we just added a newline
            if(buffer.charAt(buffer.length() - 1) != StaticStrings.NEW_LINE_CHAR) {
                buffer.append(StaticStrings.SPACE_CHAR);
            }
            buffer.append(internal_tokenizer.nextToken());
            }
            // We've prematurely run out of content, so throw the dummy, or at least return whatever we managed to parse sans its opening character
            else {
            if(strip_characters) {
                return buffer.substring(1);
            }
            else {
                buffer.append(end_char);
                return buffer.toString();
            }
            }
        }
        // We've prematurely run out of content, so throw the dummy, or at least return whatever we managed to parse sans its opening character
        else {
            if(strip_characters) {
            return buffer.substring(1);
            }
            else {
            buffer.append(end_char);
            return buffer.toString();
            }
        }
        }
        // Exception thrown when we attempted reading from the input stream, so throw the dummy, or at least return whatever we managed to parse sans its opening character
        catch(Exception exception) {
        DebugStream.printStackTrace(exception);
        if(strip_characters) {
            return buffer.substring(1);
        }
        else {
            buffer.append(end_char);
            return buffer.toString();
        }
        }
    }
    // Return the string sans enclosing characters
    if(buffer.length() >= 2 && strip_characters) {
        return buffer.substring(1, buffer.length() - 1);
    }
    else {
        return buffer.toString();
    }
    }
}