/*
 * [ProofreadInserts.java]
 *
 * Summary: Proofread Insert macros. Find dups on page. Find orphan Books/electronics/DVDs. Ensure ascending price.
 *
 * Copyright: (c) 2012-2017 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 2012-03-17 initial version
 */
package com.mindprod.stores;

import com.mindprod.common18.Misc;
import com.mindprod.filter.ExtensionListFilter;
import com.mindprod.htmlmacros.macro.Global;
import com.mindprod.htmlmacros.support.ConfigurationForMindprod;
import com.mindprod.hunkio.HunkIO;

import java.io.File;
import java.io.IOException;
import java.text.DecimalFormat;
import java.util.ArrayList;
import java.util.HashSet;
import java.util.List;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

import static java.lang.System.*;

/**
 * Proofread Insert macros. Find dups on page. Find orphan Books/electronics/DVDs. Ensure ascending price.
 * <p/>
 * All Insert and Cite macros must be expanded first.
 *
 * @author Roedy Green, Canadian Mind Products
 * @version 1.0 2012-03-17 initial version
 * @since 2012-03-17
 */
public class ProofreadInserts
    {
    // TidyBookMacros: reflow Book macros.  Update BookHead with recent title and published. Add missing published to BookHead.
    // RefreshInserts: refresh data on Insert macros from book page. Add missing published to book Insert.
    // InsertBookMarkers: put <!-- +++ --> between books to make it easier to sort.  One shot.
    // SortBooks: sort book Inserts on page so most recently published are first.
    // ProofreadInserts: check that sort worked.

    /**
     * format for non-display of price
     */
    private static final DecimalFormat DOLLAR_FORMAT = new DecimalFormat( "'$'###,##0.00" );

    /**
     * name of file e.g. book/9789506710118.html referenced by the insert
     */
    private static final HashSet<String> insertsFound = new HashSet<>( 1000 );

    /**
     * regex to look for hand coded <cite></cite>
     */
    private static final Pattern CITE_PATTERN = Pattern.compile( "<cite><a class=\"(?:book|electronics|dvd)\"\\s+"
                                                                 + "href=\"[\\./]*(?:book|electronics|dvd)/([0-9A-Z]+)\\.html\">" );

    private static List<String> dirsToProcess;

    /**
     * dir where Book macros are kept.
     */
    private static File bookDir;

    /**
     * dir where DVD macros are kept.
     */
    private static File dvdDir;

    /**
     * dir where electronic macros are kept in files of form B0057A5RA6.html
     */
    private static File electronicDir;

    /**
     * webrootDir directory of website
     */
    private static File webrootDir;

    /**
     * scan file contents for <!-- macro Cite
     *
     * @param big  contents of file
     * @param file file, used only for error messages. does not touch file
     */
    private static void scanForCiteMacros( String big, File file )
        {
        int start = 0;
        while ( ( start = big.indexOf( "<!-- macro Cite ", start ) ) >= 0 )
            {
            start += "<!-- macro Cite ".length();
            // now pointing at 9789506710118
            final int end = big.indexOf( ' ', start );
            if ( end >= 0 )
                {
                // do books, electronics and dvds
                // extract String we would for corresponding Insert macro
                final String ref = big.substring( start, end );
                // multiple cites on same page ok, also Insert and cite.
                // Record cites so we won't think this book is an orphan.
                insertsFound.add( ref );
                start = end;
                }
            else
                {
                err.println( "malformed Cite macro on page " + file );
                break;
                }
            }
        }

    /**
     * scan file contents for <cite><a class="book" href="../book/9780670020539.html">
     * <p/>
     * Hand-coded cites, or Cite macros expanded in Include.
     *
     * @param big contents of file
     */
    private static void scanForCites( String big )
        {
        final Matcher m = CITE_PATTERN.matcher( big );
        while ( m.find() )
            {
            insertsFound.add( m.group( 1 ) );
            } // end while
        } // /method

    /**
     * scan file contents for <!-- macro Insert
     *
     * @param big  contents of file
     * @param file file, used only for error messages. does not touch file
     */
    private static void scanForInsertMacros( String big, File file )
        {
        /* inserts found on current page.  Used to find dups */
        final HashSet<String> insertsFoundOnThisPage = new HashSet<>( 100 );
        int start = 0;
        while ( ( start = big.indexOf( "<!-- macro Insert ", start ) ) >= 0 )
            {
            start += "<!-- macro Insert ".length();
            // now pointing at 9789506710118
            final int end = big.indexOf( ' ', start );
            if ( end >= 0 )
                {
                // only electronic Inserts
                final String ref = big.substring( start, end );
                if ( insertsFoundOnThisPage.contains( ref ) )
                    {
                    err.println( ref + " appears more than once in " + file );
                    }
                insertsFoundOnThisPage.add( ref );
                insertsFound.add( ref );
                start = end;
                }
            else
                {
                err.println( "malformed Insert macro in " + file );
                break;
                }
            }
        }

    /**
     * scan file contents for <!-- macro Cite
     *
     * @param big  contents of file
     * @param file file, used only for error messages. does not touch file
     */
    private static void scanForPrices( String big, File file )
        {
        // we are scanning in expanded macros, not the macros themselves.
        // We are scanning ordinary pages, not the electronic pages for individual products.
        double previousPriceOnPage = 0;
        int start = 0;
        while ( ( start = big.indexOf( "<span class=\"hidden\">$", start ) ) >= 0 )
            {
            start += "<span class=\"hidden\">$".length();
            // now pointing at 9.00</span>
            final int end = big.indexOf( '<', start );
            if ( end >= 0 )
                {
                final double price = Double.parseDouble( big.substring( start, end ).replaceAll( ",", "" ) );
                start = end;
                if ( price == 0 )
                    {
                    err.println( "0 price in " + file );
                    }
                else if ( price < previousPriceOnPage - 1 )  // allow slop in sort of $1
                    {
                    err.println( "prices out of order between "
                                 + DOLLAR_FORMAT.format( previousPriceOnPage )
                                 + " and "
                                 + DOLLAR_FORMAT.format( price )
                                 + " in "
                                 + file );
                    }
                else
                    {
                    previousPriceOnPage = price;
                    }
                }
            else
                {
                err.println( "malformed price on page " + file );
                break;
                }
            }
        }

    /**
     * @param args not used
     *
     * @throws java.io.IOException if trouble reading or writing files containing Book macros.
     */
    public static void main( String[] args ) throws IOException
        {
        Global.installConfiguration( new ConfigurationForMindprod() );
        // combine dirsWithMacros and dirsWithIncludes into dirsToProcess;
        final List<String> a = Global.configuration.getDirsContainingHTMLMacros();
        final List<String> b = Global.configuration.getDirsContainingIncludes();
        dirsToProcess = new ArrayList<>( a.size() + b.size() );
        dirsToProcess.addAll( a );
        dirsToProcess.addAll( b );
        webrootDir = new File( Global.configuration.getLocalWebrootWithSlashes() );
        electronicDir = new File( webrootDir, "electronic" );
        bookDir = new File( webrootDir, "book" );
        dvdDir = new File( webrootDir, "dvd" );
        for ( String dirname : dirsToProcess )
            {
            final File dir = new File( webrootDir, dirname );
            // out.println("Processing " + dir );
            final String[] filenames = dir.list( new ExtensionListFilter( "html" ) );
            for ( String filename : filenames )
                {
                final File file = new File( dir, filename );
                final String big = HunkIO.readEntireFile( file, HunkIO.UTF8 );
                scanForInsertMacros( big, file );
                scanForCiteMacros( big, file );
                scanForCites( big );
                scanForPrices( big, file );
                } // end scan  all files in dir
            } // end scan all dirs.
        // look for unreferenced electronic ASINs
        final String[] electronicFilenames = electronicDir.list( new ElectronicMacroFileFilter() );
        err.println( "\nOrphaned files have no <cites>, Cites or Inserts, though they appear on the summary page.\n" );
        for ( String electronicFilename : electronicFilenames )
            {
            if ( !insertsFound.contains( electronicFilename.substring( 0, 10 ) ) )
                {
                err.println( "Orphaned electronic file " + electronicFilename );
                }
            }
        // look for unreferenced books
        final String[] bookFilenames = bookDir.list( new BookMacroFileFilter() );
        for ( String bookFilename : bookFilenames )
            {
            if ( !insertsFound.contains( bookFilename.substring( 0, 13 ) ) )
                {
                err.println( "Orphaned book file " + bookFilename );
                }
            }
        // look for unreferenced DVDs
        final String[] dvdFilenames = dvdDir.list( new DVDMacroFileFilter() );
        for ( String dvdFilename : dvdFilenames )
            {
            if ( !insertsFound.contains( dvdFilename.substring( 0, 12 ) ) )
                {
                err.println( "Orphaned dvd file " + dvdFilename );
                }
            }
        Misc.trackLastThread();
        System.exit( 0 );
        } // end main
    }