gov.nih.nlm.nls.lvg.Flows.ToStripPunctuation Maven / Gradle / Ivy

Go to download
Show more of this group Show more artifacts with this name
Show all versions of lvg2010dist Show documentation
LVG tools is used by Apache cTAKES.
The newest version!
package gov.nih.nlm.nls.lvg.Flows;
import java.util.*;
import gov.nih.nlm.nls.lvg.Lib.*;
import gov.nih.nlm.nls.lvg.Util.*;
/*****************************************************************************
* This class strips punctuations from a specified term.  Punctuations include:
* 
* DASH_PUNCTUATION (20): -
* 
START_PUNCTUATION (21): ( { [
* 
END_PUNCTUATION (22): ) } ]
* 
CONNECTOR_PUNCTUATION (23): _
* 
OTHER_PUNCTUATION (24): ! @ # % & * \ : ; " ' ,  . ? /
* 
MATH_SYMBOL (25): ~ + = | < >
* 
CURRENCY_SYMBOL (26): $
* 
MODIFIER_SYMBOL (27): ` ^
* 
*
* History:
* 

* 
*
* @author NLM NLS Development Team
*
* @see 
* Design Document 
*
* @version    V-2010
****************************************************************************/
public class ToStripPunctuation extends Transformation implements Cloneable
{
    // public methods
    /**
    * Performs the mutation of this flow component.
    *
    * @param   in   a LexItem as the input for this flow component
    * @param   detailsFlag   a boolean flag for processing details information
    * @param   mutateFlag   a boolean flag for processing mutate information
    *
    * @return  Vector - results from this flow component 
    */
    public static Vector Mutate(LexItem in, boolean detailsFlag, 
        boolean mutateFlag)
    {
        // Mutate the term:
        String term = StripPunctuation(in.GetSourceTerm());
        // details & mutate
        String details = null;
        String mutate = null;
        if(detailsFlag == true)
        {
            details = INFO;
        }
        if(mutateFlag == true)
        {
            mutate = Transformation.NO_MUTATE_INFO;
        }
        // update target
        Vector out = new Vector();
        LexItem temp = UpdateLexItem(in, term, Flow.STRIP_PUNCTUATION, 
            Transformation.UPDATE, Transformation.UPDATE, details, mutate);
        out.addElement(temp);
        return out;
    }
    /**
    * A unit test driver for this flow component.
    */
    public static void main(String[] args)
    {
        String testStr = GetTestStr(args, "Left's 12.34.56");
        // Mutate
        LexItem in = new LexItem(testStr);
        Vector outs = ToStripPunctuation.Mutate(in, true, true);
        PrintResults(in, outs);     // print out result
    }
    // package method
    /**
    * Strip punctuations from a specified term.
    *
    * @param   inStr   the specified term to be stripped punctuations from.
    *
    * @return  the term after being stripped punctuations
    */
    static String StripPunctuation(String inStr)
    {
        int length = inStr.length();
        char[] temp = new char[length];
        int index = 0;
        for(int i = 0; i < length; i++)
        {
            char tempChar = inStr.charAt(i);
            if(Char.IsPunctuation(tempChar) == false)
            {
                temp[index] = tempChar;
                index++;
            }
        }
        String out = new String(temp);
        return out.trim();            // must be trimmed 
    }
    // data members
    private static final String INFO = "Strip Punctuation";
}