package de.mpiwg.anteater.results.filter;

import java.util.ArrayList;
import java.util.List;

import de.mpiwg.anteater.results.ApplicantResult;
import de.mpiwg.anteater.results.LocationResult;
import de.mpiwg.anteater.results.SpeciesScientificResult;
import de.mpiwg.anteater.text.Paragraph;
import de.mpiwg.anteater.text.TextInformation;
import de.mpiwg.anteater.text.TextPart;
import de.mpiwg.anteater.text.TextType;

public class SpeciesInBetweenRegexApplicantFilter implements IResultFilter {

	@Override
	public void filterElements(TextInformation info,
			List<ApplicantResult> applicantResults,
			List<SpeciesScientificResult> speciesResults,
			List<LocationResult> locationResults) {
		
		List<TextPart> summaries = info.getSummaries();
		List<TextPart> suppleInfs = info.getSupplInfos();
		for (ApplicantResult applicant : applicantResults) {
			
			// if applicant was found with regex there can't be a species in it
			if (applicant.getPrediction() == 2.0) {
				TextPart part;
				
				if (applicant.getResult().getType() == TextType.TYPE_SUMMARY)
					part = summaries.get(applicant.getResult().getTextIdx());
				else 
					part = suppleInfs.get(applicant.getResult().getTextIdx());
				
				Paragraph paragraph = part.getParagraphOfIndex(applicant.getFinding().getStart());
				
				List<SpeciesScientificResult> toBeRemoved = new ArrayList<SpeciesScientificResult>();
				for (SpeciesScientificResult species : speciesResults) {
					// if species is from same text type and text index
					if (species.getResult().getType() == applicant.getResult().getType() && species.getResult().getTextIdx() == applicant.getResult().getTextIdx()) {
						// if species is not in same paragraph as applicant all is good
						if (part.getParagraphOfIndex(species.getFinding().getStart()) !=  paragraph)
							continue;
					}
					toBeRemoved.add(species);
				}
				
				speciesResults.removeAll(toBeRemoved);
			}
		}
	}

}