/*
 * [ImportQuotesFromHTML.java]
 *
 * Summary: Imports quote collection from HTML to the database.
 *
 * Copyright: (c) 2014-2017 Roedy Green, Canadian Mind Products, http://mindprod.com
 *
 * Licence: This software may be copied and used freely for any purpose but military.
 *          http://mindprod.com/contact/nonmil.html
 *
 * Requires: JDK 1.8+
 *
 * Created with: JetBrains IntelliJ IDEA IDE http://www.jetbrains.com/idea/
 *
 * Version History:
 *  1.0 2014-08-08 initial version
 */
package com.mindprod.quotedb;

import com.mindprod.common18.BigDate;
import com.mindprod.common18.Chunkify;
import com.mindprod.common18.EIO;
import com.mindprod.hunkio.HunkIO;

import java.io.File;
import java.io.IOException;
import java.sql.Connection;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

/**
 * Imports quote collection from HTML to the database.
 *
 * @author Roedy Green, Canadian Mind Products
 * @version 1.0 2014-08-08 initial version
 * @since 2014-08-08
 */
public class ImportQuotesFromHTML
    {
    private static final Pattern AUTHOR_EXTRACTOR = Pattern.compile( "<span class=\"said\">([ -;=\\?-~])</span>" );

    private static final Pattern BIRTH_EXTRACTOR = Pattern.compile( "<!-- macro Born ([0-9\\-ADBC]+)" );

    private static final Pattern DEATH_EXTRACTOR = Pattern.compile( "<!-- macro Born [0-9\\-ADBC]+ ([0-9\\-ADBC]+)" );

    private static final Pattern CSS_EXTRACTOR = Pattern.compile( "<blockquote class=\"(\\p{Alnum}+)\">" );

    private static int uniqueAuthorID = 0;

    private static String authorName = null;

    private static BigDate birthDate = null;

    private static BigDate deathDate = null;

    private static String bornInfo = null;

    private static String commentary = null;

    private static String cssClass = null;

    private static BigDate composedDate = null;

    private static String notesAboutAuthor = null;

    private static String quotation = null;

    private static String source = null;

    private static String title = null;

    private static int uniqueQuoteID = 0;
    // /declarations

    /**
     * The connection.  Handle to the database
     */
    private static Connection conn = null;

    /**
     * import set of quotes from a *.html file
     */
    public void importQuotesFromHTML( File quotationsFile ) throws IOException
        {
        // file should be stripgenerated, no syntax errors.
        String big = HunkIO.readEntireFile( quotationsFile, EIO.UTF8 );
        Chunkify chunks = new Chunkify( big, "<blockquote", "</blockquote" );
        for ( String chunk : chunks )
            {
            Matcher m = AUTHOR_EXTRACTOR.matcher( chunk );
            if ( m.find() )
                {
                authorName = m.group( 1 );
                }
            else
                {
                authorName = null;
                }
            m = BIRTH_EXTRACTOR.matcher( chunk );
            if ( m.find() )
                {
                birthDate = new BigDate( m.group( 1 ) );
                }
            else
                {
                birthDate = null;
                }
            m = DEATH_EXTRACTOR.matcher( chunk );
            if ( m.find() )
                {
                deathDate = new BigDate( m.group( 1 ) );
                }
            else
                {
                deathDate = null;
                }
            //
            // commentary
            m = CSS_EXTRACTOR.matcher( chunk );
            if ( m.find() )
                {
                cssClass = m.group( 1 );
                }
            else
                {
                cssClass = null;
                }
            }
        }   // /method
    // /methods
    }