/*
 * [CollectFamousBorns.java]
 *
 * Summary: Find Author/Birth/Death in near said or famous people.
 *
 * Copyright: (c) 2011-2017 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 2011-10-18 initial version
 */
package com.mindprod.repair;

import com.mindprod.entities.DeEntifyStrings;

import java.io.File;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import static java.lang.System.*;

/**
 * Find Author/Birth/Death in near said or famous people.
 * <p/>
 * Does not change the html files.
 *
 * @author Roedy Green, Canadian Mind Products
 * @version 1.0 2011-10-18 initial version
 * @since 2011-10-18
 */
class CollectFamousBorns
    {
    private static final boolean DEBUGGING = false;

    /**
     * pattern of quote with Born with two dates. Ignore quoth generated quotations.
     */
    private static final Pattern FAMOUS_DEAD_PATTERN = Pattern.compile(
            "<span class=\"(said|famous)\">([^<]+)</span>\\s*<!\\-\\-\\s+macro\\s+Born\\s+([\\d\\-ADBC]+)\\s+" +
            "([\\d\\-ADBC]|(unknown)+)\\s+(?:checked)?\\s*\\-\\->",
            Pattern.DOTALL
    );

    /**
     * pattern of quote with Born with one date. Ignore quoth generated quotations. .
     */
    private static final Pattern FAMOUS_LIVING_PATTERN = Pattern.compile(
            "<span class=\"(said|famous)\">([^<]+)</span>\\s*<!\\-\\-\\s+macro\\s+Born\\s+([\\d\\-unknown]+)" +
            "(?:\\s+checked)?\\s*\\-\\->",
            Pattern.DOTALL
    );

    /**
     * pattern of quote already handled, used to extract author
     */
    private static final Pattern MULTI_AUTHOR_SPLIT = Pattern.compile(
            "\\s+and\\s+|\\s+with\\s+|\\s+aka\\s+|,|&amp;|\\[|\\]" );

    /**
     * find all famous or said borns on page. Add details of authors to HashMap
     *
     * @param big                the whole page
     * @param fileBeingProcessed file we are processing
     */
    static void collectBornsForPage( final String big, final File fileBeingProcessed )
        {
        final Matcher living = FAMOUS_LIVING_PATTERN.matcher( big );
        // find names of authors and birth and death dates inside and outside Blockquotes
        while ( living.find() )
            {
            if ( DEBUGGING )
                {
                for ( int i = 0; i <= living.groupCount(); i++ )
                    {
                    out.println( ">>living group: " + i + " : " + living.group( i ) );
                    }
                }
            processOneBornMacro( living.group( 2 ), living.group( 3 ), "", fileBeingProcessed );
            }
        final Matcher dead = FAMOUS_DEAD_PATTERN.matcher( big );
        while ( dead.find() )
            {
            if ( DEBUGGING )
                {
                for ( int i = 0; i <= dead.groupCount(); i++ )
                    {
                    out.println( ">>dead group: " + i + " : " + dead.group( i ) );
                    }
                }
            processOneBornMacro( dead.group( 2 ), dead.group( 3 ), dead.group( 4 ), fileBeingProcessed );
            }
        }

    /**
     * process author and birth/date dates
     *
     * @param allAuthors         list of authors (ignore all but first )
     * @param birthdate          unknown or yyyy-mm-dd
     * @param deathdate          "" or yyyy-mm-dd
     * @param fileBeingProcessed file we are processing
     */
    private static void processOneBornMacro( String allAuthors,
                                             final String birthdate,
                                             final String deathdate,
                                             final File fileBeingProcessed )
        {
        allAuthors = DeEntifyStrings.stripHTMLTags( allAuthors );
        final String[] individualAuthors = MULTI_AUTHOR_SPLIT.split( allAuthors );
        final String author = individualAuthors.length >= 1 ? individualAuthors[ 0 ].trim() : "";
        final String authorFlat = DeEntifyStrings.deEntifyHTML( author, ' ' );
        final AuthorBio old = AuthorBio.get( authorFlat );
        final AuthorBio recent = new AuthorBio( author, birthdate, deathdate, fileBeingProcessed );
        // we don't keep that AuthorBio object unless it is the first for this author.
        if ( old == null )
            {
            if ( DEBUGGING )
                {
                out.println( " adding : " + recent );
                }
            AuthorBio.put( authorFlat, recent );
            }
        else if ( old.getBirthdate().equals( birthdate ) && old.getDeathdate().equals( "" ) && deathdate.length() > 0 )
            {
            AuthorBio.put( authorFlat, recent ); // update deathdate info
            }
        else if ( !( old.getBirthdate().equals( birthdate ) && old.getDeathdate().equals( deathdate ) ) )
            {
            AuthorWarning.warn( "date inconsistent with file date", old, recent );
            }
        }
    }