D-Net

package eu.dnetlib.functionality.cql.parse;

import java.text.ParseException;

import java.text.SimpleDateFormat;

import java.util.List;

import java.util.Map;

import java.util.Map.Entry;

import com.google.common.base.Function;

import com.google.common.base.Joiner;

import com.google.common.base.Splitter;

import com.google.common.collect.BiMap;

import com.google.common.collect.HashBiMap;

import com.google.common.collect.Iterables;

import com.google.common.collect.Maps;

import org.apache.commons.lang.StringUtils;

import org.apache.lucene.queryparser.classic.QueryParserBase;

public class TermNode extends Node {

	public static final String dnetDefaultField = "__all";

	private String field;

	private Relation rel;

	private String value;

	private Map<String, List<String>> options = Maps.newHashMap();

	private BiMap<String, String> aliases = HashBiMap.create();

	private Map<String, String> weights = Maps.newHashMap();

	private SimpleDateFormat simpleDateFormat = new SimpleDateFormat("yyyy-MM-dd");

	public TermNode(final String field, final Relation rel, final String value) {

		this.field = field;

		this.rel = rel;

		this.value = value;

		simpleDateFormat.setLenient(false);

	}

	public TermNode(final String field, final Relation rel, final String value, final Map<String, List<String>> options, final BiMap<String, String> aliases,

			final Map<String, String> weights) {

		this(field, rel, value);

		this.options = options;

		this.aliases = aliases;

		this.weights = weights;

	}

	@Override

	public String toString() {

		return TermNode.class.getSimpleName() + "(" + field + " " + rel + " " + value + ")";

	}

	@Override

	public String toLucene() {

		//StringTokenizer termTokenizer = new StringTokenizer(value, " ");

		//StringTokenizer weightTokenizer = new StringTokenizer(value, " ");

		final Iterable<String> termTokenizer = Splitter.on(" ").omitEmptyStrings().split(value);

		final Iterable<String> weightTokenizer = Splitter.on(" ").omitEmptyStrings().split(value);

		switch (rel) {

		case EXACT:

			final String lucene = getFieldName() + ":" + "\"" + value + "\"";

			return StringUtils.isNotBlank(weight()) ? lucene + weight() + " " + expand(value) : lucene;

		case EQUAL:

		case ALL:

			return "(" + handleTokens(termTokenizer, "AND") + " " + expandTokens(weightTokenizer) + ")";

		case ANY:

			return "(" + handleTokens(termTokenizer, "OR") + " " + expandTokens(weightTokenizer) + ")";

		case NOT:

			return "NOT " + field + ":" + "\"" + value + "\"";

		case LT:

			if (isDate(value)) {

				value = checkDate(value);

			}

			return field + ":" + "{* TO " + value + "}" + weight();

		case GT:

			if (isDate(value)) {

				value = checkDate(value);

			}

			return field + ":" + "{" + value + " TO *}" + weight();

		case LTE:

			if (isDate(value)) {

				value = checkDate(value);

			}

			return field + ":" + "[* TO " + value + "]" + weight();

		case GTE:

			if (isDate(value)) {

				value = checkDate(value);

			}

			return field + ":" + "[" + value + " TO *]" + weight();

		case WITHIN:

			String lowerValue = value.split(" ")[0];

			String upperValue = value.split(" ")[1];

			if (isDate(lowerValue)) {

				lowerValue = checkDate(lowerValue);

			}

			if (isDate(upperValue)) {

				upperValue = checkDate(upperValue);

			}

			return field + ":[" + lowerValue + " TO " + upperValue + "]" + weight();

		default:

			throw new RuntimeException("unable to serialize: " + toString());

		}

	}

	private String getFieldName() {

		return aliases.get(field) != null ? aliases.get(field) : field;

	}

	private String weight() {

		return (weights != null) && (weights.get(field) != null) ? "^" + weights.get(field) : "";

	}

	private String expandTokens(final Iterable<String> tokens) {

		return Joiner.on("").skipNulls().join(Iterables.transform(tokens, new Function<String, String>() {

			@Override

			public String apply(final String s) {

				if (field.equals(dnetDefaultField.toLowerCase()) || field.equals(dnetDefaultField.toLowerCase())) {

					return expand(s);

				}

				return null;

			}

		})).trim();

	}

	private String expand(final String token) {

		String ret = "";

		if (!weights.keySet().contains(field)) {

			for (Entry<String, String> e : weights.entrySet()) {

				ret += e.getKey() + ":\"" + checkEscaping(token) + "\"^" + e.getValue() + " ";

			}

		}

		return ret.trim();

	}

	private String handleTokens(final Iterable<String> tokens, final String op) {

		final String separator = " " + op + " ";

		return Joiner.on(separator).join(Iterables.transform(tokens, new Function<String, String>() {

			@Override

			public String apply(final String s) {

				return field + ":" + checkEscaping(s) + weight();

			}

		})).trim();

	}

	private String checkEscaping(String token) {

		boolean isWildcard = token.contains("*") || token.contains("?");

		boolean isWildcardEnabled = ((options.get("wildcard") != null) && options.get("wildcard").contains("true")) || token.equals("*");

		if (!(isWildcard & isWildcardEnabled)) {

			token = QueryParserBase.escape(token);

		}

		return token;

	}

	private boolean isDate(final String aPossibleDate) {

		try {

			simpleDateFormat.parse(aPossibleDate);

		} catch (ParseException pe) {

			return false;

		}

		return true;

	}

	private String checkDate(final String date) {

		if (!date.endsWith("Z")) return date + "T00:00:00Z";

		return date;

	}

	public String getField() {

		return field;

	}

	public Relation getRel() {

		return rel;

	}

	public String getValue() {

		return value;

	}

}