package eu.dnetlib.data.collector.plugins.archive.zip;

import java.io.File;
import java.net.MalformedURLException;
import java.net.URL;
import java.util.Iterator;

import com.google.common.base.Function;
import com.google.common.collect.Iterators;

import eu.dnetlib.data.collector.plugins.oai.engine.XmlCleaner;
import eu.dnetlib.data.collector.rmi.CollectorServiceException;
import eu.dnetlib.data.collector.rmi.InterfaceDescriptor;

/**
 *
 * @author Andrea
 *
 */
public class ZipIterable implements Iterable<String> {

	/** The path to .zip archive. */
	private File zipFile;

	public ZipIterable(final InterfaceDescriptor interfaceDescriptor) throws CollectorServiceException {
		try {
			final String zipPath = interfaceDescriptor.getBaseUrl();
			URL zipUrl = new URL(zipPath);
			this.zipFile = new File(zipUrl.getPath());
			if (!zipFile.exists()) { throw new CollectorServiceException(String.format("The base ULR %s, does not exist", zipFile.getPath())); }
		} catch (MalformedURLException e) {
			throw new CollectorServiceException("Zip collector failed! ", e);
		}
	}

	@Override
	public Iterator<String> iterator() {
		final ZipIterator zipIterator = new ZipIterator(zipFile.getAbsolutePath());
		return Iterators.transform(zipIterator, new Function<String, String>() {

			@Override
			public String apply(final String inputRecord) {
				return XmlCleaner.cleanAllEntities(inputRecord.startsWith("\uFEFF") ? inputRecord.substring(1) : inputRecord);
			}
		});
	}

}