//Dstl (c) Crown Copyright 2017
package uk.gov.dstl.baleen.annotators.regex;

import java.util.Arrays;
import java.util.Collections;
import java.util.TimeZone;
import java.util.regex.Matcher;
import java.util.stream.Collectors;

import org.apache.commons.lang.StringUtils;
import org.apache.uima.fit.descriptor.ConfigurationParameter;
import org.apache.uima.jcas.JCas;

import com.google.common.collect.ImmutableSet;

import uk.gov.dstl.baleen.annotators.regex.helpers.AbstractRegexAnnotator;
import uk.gov.dstl.baleen.core.pipelines.orderers.AnalysisEngineAction;
import uk.gov.dstl.baleen.types.semantic.Temporal;

/**
 * Annotate times within a document using regular expressions
 * 
 * <p>The document content is run through a regular expression matcher looking for things that match the following time regular expression,
 * where UTC is being used to represent all time zone acronyms defined in Java:</p>
 * <pre>\\b(((0?[0-9])|([0-9]{2}))[:][0-9]{2}\\h*((UTC)([ ]?[+-][ ]?((0?[0-9])|(1[0-2])))?)?\\h*(pm|am)?)\\b|\\b(((1[0-2])|([1-9]))(pm|am))\\b|\\b(midnight)\\b|\\b(midday)\\b|\\b((12\\h)?noon)\\b|\\b([0-2][0-9][0-5][0-9][ ]?(hr(s)?)?[ ]?((UTC)([ ]?[+-][ ]?((0?[0-9])|(1[0-2])))?)?)\\b</pre>
 * <p>This will only capture times that match the regular expression, and will miss times expressed in a different format.</p>
 * 
 * By default, only times that contain alphabetical characters or colons will be accepted to minimise false positives.
 */
public class Time extends AbstractRegexAnnotator<Temporal> {

	/**
	 * Do we require that there are alphabetical characters in the time?
	 * This helps avoid picking out things like 2015 as a time when it should be a year,
	 * as it forces the time to be written like 2015hrs or 8:15pm.
	 * 
	 * For the purposes of the TimeRegex annotator, colons are treated as alphabetical characters,
	 * such that times such as 20:15 are captured. Other punctuation isn't, as 20.15 is more like to
	 * be an amount than a time. 
	 * 
	 * @baleen.config true
	 */
	public static final String PARAM_REQUIRE_ALPHA = "requireAlpha";
	@ConfigurationParameter(name = PARAM_REQUIRE_ALPHA, defaultValue = "true")
	private Boolean requireAlpha;
	
	private static final String TIME_ZONES = StringUtils.join(
		Arrays.asList(TimeZone.getAvailableIDs())
			.stream().filter(s -> StringUtils.isAllUpperCase(s) && s.length() <= 3)
			.collect(Collectors.toList()),
		"|");
	
	private static final String TIME_REGEX = "\\b(([0-1]?[0-9]|2[0-4])[:\\.][0-5][0-9]\\h*(("+TIME_ZONES+")([ ]?[+-][ ]?((0?[0-9])|(1[0-2])))?)?\\h*(pm|am)?)\\b|\\b(((1[0-2])|([1-9]))(pm|am))\\b|\\b(midnight)\\b|\\b(midday)\\b|\\b((12\\h)?noon)\\b|\\b([0-1][0-9]|2[0-4])[0-5][0-9][ ]?(((hr(s)?)?[ ]?(("+TIME_ZONES+")([ ]?[+-][ ]?((0?[0-9])|(1[0-2])))?)?)|hours|h)\\b";
	
	/** New instance.
	 * 
	 */
	public Time() {
		super(TIME_REGEX, false, 1.0);
	}
	
	@Override
	protected Temporal create(JCas jCas, Matcher matcher) {
		if(requireAlpha){
			String time = matcher.group();
			if(!time.matches(".*[a-zA-Z:].*")){
				return null;
			}
		}
		
		Temporal dtg = new Temporal(jCas);
		
		dtg.setPrecision("UNQUALIFIED");
		dtg.setScope("SINGLE");
		dtg.setTemporalType("TIME");

		return dtg;
	}
	
	@Override
	public AnalysisEngineAction getAction() {
		return new AnalysisEngineAction(Collections.emptySet(), ImmutableSet.of(Temporal.class));
	}
}