D-Net

package eu.dnetlib.pace.clustering;

import java.util.Collection;

import java.util.LinkedHashSet;

import java.util.Map;

import java.util.StringTokenizer;

/**

 * The Class Ngrams.

 */

public class Ngrams extends AbstractClusteringFunction {

	/**

	 * Instantiates a new ngrams.

	 * 

	 * @param params

	 *            the params

	 */

	public Ngrams(final Map<String, Integer> params) {

		super(params);

	}

	/*

	 * (non-Javadoc)

	 * 

	 * @see eu.dnetlib.pace.clustering.AbstractClusteringFunction#doApply(java.lang.String)

	 */

	@Override

	protected Collection<String> doApply(final String s) {

		return getNgrams(s, param("ngramLen"), param("max"), param("maxPerToken"), param("minNgramLen"));

	}

	/**

	 * Gets the ngrams.

	 * 

	 * @param s

	 *            the s

	 * @param ngramLen

	 *            the ngram len

	 * @param max

	 *            the max

	 * @param maxPerToken

	 *            the max per token

	 * @param minNgramLen

	 *            the min ngram len

	 * @return the ngrams

	 */

	protected Collection<String> getNgrams(final String s, final int ngramLen, final int max, final int maxPerToken, final int minNgramLen) {

		final Collection<String> ngrams = new LinkedHashSet<String>();

		final StringTokenizer st = new StringTokenizer(s);

		while (st.hasMoreTokens()) {

			final String token = st.nextToken();

			if (!token.isEmpty()) {

				for (int i = 0; (i < maxPerToken) && ((ngramLen + i) <= token.length()); i++) {

					String ngram = (token + "    ").substring(i, ngramLen + i).trim();

					if (ngrams.size() >= max) return ngrams;

					if (ngram.length() >= minNgramLen) {

						ngrams.add(ngram);

					}

				}

			}

		}

		// System.out.println(ngrams + " n: " + ngrams.size());

		return ngrams;

	}

}