/*
 * [CollectBornsInBookMacros.java]
 *
 * Summary: Find Author/Birth/Death in Book Macros.
 *
 * 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;

/**
 * Find Author/Birth/Death in Book Macros.
 * <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 CollectBornsInBookMacros
    {
    /**
     * pattern of quote already handled, used to extract author in {}
     */
    private static final Pattern AUTHOR_PATTERN_IN_BOOK_MACRO_IN_BRACES = Pattern.compile( "author=\\{(.+?)\\}",
            Pattern.DOTALL );

    /**
     * pattern of quote already handled, used to extract author in " "
     */
    private static final Pattern AUTHOR_PATTERN_IN_BOOK_MACRO_IN_QUOTES = Pattern.compile( "author=\"(.+?)\"",
            Pattern.DOTALL );

    /**
     * pattern birth="yyyy-mm-dd"
     */
    private static final Pattern BIRTH_PATTERN_IN_BOOK_MACRO = Pattern.compile( "birth=\"(.+?)\"" );

    /**
     * look for <!-- macro Book/DVD ... -->
     */
    private static final Pattern BOOK_MACRO_PATTERN = Pattern.compile(
            "<!\\-\\-\\s+macro\\s(?:Book|DVD)\\s+(.+?)\\s+\\-\\->",
            Pattern.DOTALL );

    /**
     * death=yyyy-mm-dd
     */
    private static final Pattern DEATH_PATTERN_IN_BOOK_MACRO = Pattern.compile( "death=\"(.+?)\"" );

    /**
     * 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 BookMacros on page. Add details of authors to HashMap
     *
     * @param page               the whole page
     * @param fileBeingProcessed file we are processing
     *                           .
     */
    //  CollectBornsInBookMacros
    static void collectBornsInBookMacroForPage( final String page, final File fileBeingProcessed )
        {
        Matcher m = BOOK_MACRO_PATTERN.matcher( page );
        while ( m.find() )
            {
            collectBornsInBookMacrosForPage( m.group( 1 ), fileBeingProcessed );
            }
        }
    // <span class="said">Stephen Harper</span> <!-- macro Born 1959-04-30 -->

    /**
     * find all blockquotes on page. Give new CSS classes. Add details of authors to HashMap
     *
     * @param chunk              the whole blockquote with head and teal tags trimmed off.
     * @param fileBeingProcessed file we are processing
     *                           .
     */
    private static void collectBornsInBookMacrosForPage( final String chunk, final File fileBeingProcessed )
        {
        // if it is not there, no problem. We will detect it and complain later in repair if we can't fix it.
        if ( chunk.contains( "birth=" ) )
            {
            final Matcher a1 = AUTHOR_PATTERN_IN_BOOK_MACRO_IN_QUOTES.matcher( chunk );
            final Matcher a2 = AUTHOR_PATTERN_IN_BOOK_MACRO_IN_BRACES.matcher( chunk );
            final Matcher b = BIRTH_PATTERN_IN_BOOK_MACRO.matcher( chunk );
            final Matcher d = DEATH_PATTERN_IN_BOOK_MACRO.matcher( chunk );
            boolean founda1 = a1.find();
            boolean founda2 = a2.find();
            if ( ( founda1 || founda2 ) && b.find() )
                {
                final Matcher a = founda1 ? a1 : a2;
                final String allAuthors = a.group( 1 );
                final String[] individualAuthors = MULTI_AUTHOR_SPLIT.split( allAuthors );
                final String author = individualAuthors.length >= 1 ? individualAuthors[ 0 ].trim() : "";
                final String authorFlat = DeEntifyStrings.flattenHTML( author, ' ' );
                final AuthorBio old = AuthorBio.get( authorFlat );
                // we collect unknowns too.
                String birthdate = b.group( 1 );  // found done earlier
                String deathdate = d.find() ? d.group( 1 ) : "";
                final AuthorBio recent = new AuthorBio( author, birthdate, deathdate, fileBeingProcessed );
                if ( old == null )
                    {
                    AuthorBio.put( authorFlat, recent );
                    }
                else
                    {
                    if ( !( old.getBirthdate().equals( birthdate ) && old.getDeathdate().equals( deathdate ) ) )
                        {
                        AuthorWarning.warn( "inconsistent dates in book macro", old, recent );
                        }
                    // if matches, do nothing. We already have all their is to know about the author on fil.
                    }
                }
            }
        }
    }