/*
 * [Chaos.java]
 *
 * Summary: Measures how chaotic the line lengths of a text or html file are.
 *
 * Copyright: (c) 2013-2017 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 2013-02-16 initial version
 */
package com.mindprod.chaos;

import com.mindprod.commandline.CommandLine;
import com.mindprod.common18.EIO;
import com.mindprod.common18.ST;
import com.mindprod.filter.AllButSVNDirectoriesFilter;
import com.mindprod.filter.ExtensionListFilter;
import com.mindprod.hunkio.HunkIO;

import java.io.File;
import java.io.FileNotFoundException;
import java.io.IOException;
import java.util.regex.Pattern;

import static java.lang.System.*;

/**
 * Measures how chaotic the line lengths of a text or html file are.
 * <p/>
 * An indication the file needs to be tidied.
 *
 * @author Roedy Green, Canadian Mind Products
 * @version 1.0 2013-02-16 initial version
 * @since 2013-02-16
 */
public class Chaos
    {
    /**
     * year first released
     */
    private static final int FIRST_COPYRIGHT_YEAR = 2013;

    /**
     * undisplayed copyright notice.
     *
     * @noinspection UnusedDeclaration
     */
    private static final String EMBEDDED_COPYRIGHT =
            "Copyright: (c) 2013-2017 Roedy Green, Canadian Mind Products, http://mindprod.com";

    /**
     * date this version released.
     *
     * @noinspection UnusedDeclaration
     */
    private static final String RELEASE_DATE = "2013-02-16";

    /**
     * how to use the program
     */
    private static final String USAGE = "\n\nChaos requires a list of *.html or *.txt files to process. -q means all " +
                                        "files in dirs.";

    /**
     * embedded version string.
     *
     * @noinspection UnusedDeclaration
     */
    private static final String VERSION_STRING = "1.0";

    /**
     * break file into lines at \n or \r\n
     */
    private static final Pattern LINE_SPLITTER = Pattern.compile( "\\r\\n|\\n", Pattern.MULTILINE );

    /**
     * break file into words at space
     */
    private static final Pattern WORD_SPLITTER = Pattern.compile( "\\s+" );

    /**
     * calculate average line length andf aveerage deviation (not least square)
     *
     * @param file file to compute
     *
     * @throws IOException
     */
    private static void displayChaos( File file ) throws IOException
        {
        final String contents = HunkIO.readEntireFile( file );
        final String[] lines = LINE_SPLITTER.split( contents );
        final int divisor = lines.length != 0 ? lines.length : 1;
        // calc average line length
        int total = 0;
        for ( String line : lines )
            {
            total += line.length();
            }
        final int average = total / divisor;
        //calc average deviation from average
        int totalDev = 0;
        for ( String line : lines )
            {
            totalDev += Math.abs( line.length() - average );
            }
        final int averageDev = totalDev / divisor;
        // find longest line
        int max = 0;
        for ( String line : lines )
            {
            if ( line.length() > max )
                {
                max = line.length();
                }
            }
        // find word count
        int wordCount = 0;
        for ( String line : lines )
            {
            final String[] wordsInLine = WORD_SPLITTER.split( line );
            wordCount += wordsInLine.length;
            // don't count poosible  dummy blank word at head of line
            if ( wordsInLine.length > 0 && wordsInLine[ 0 ].length() == 0 )
                {
                wordCount--;
                }
            }
        // output average line length and average deviation from average
        out.println( ST.rightJustified( average, 4, false )
                     + ST.rightJustified( averageDev, 4, false )
                     + ST.rightJustified( max, 6, false )
                     + ST.rightJustified( lines.length, 7, false )
                     + ST.rightJustified( wordCount, 7, false )
                     + ST.rightJustified( contents.length(), 8, false )
                     + "  "
                     + EIO.getCanOrAbsPath( file ) );
        }

    /**
     * takes list of files to process, displays chaos numbers.
     *
     * @param args command line arguments names of files to process or -s
     */
    public static void main( final String args[] )
        {
        CommandLine commandLine = new CommandLine( args,
                new AllButSVNDirectoriesFilter(),
                new ExtensionListFilter( "html", "htm", "txt", "bat", "btm" ) );
        if ( commandLine.size() == 0 )
            {
            throw new IllegalArgumentException( "No files found to process\n" + USAGE );
            }
        out.println( " -line length-  -------count--------" );
        out.println( " avg dev   max  lines  words   chars" );
        for ( File file : commandLine )
            {
            try
                {
                displayChaos( file );
                }
            catch ( FileNotFoundException e )
                {
                out.println( "Error: "
                             + EIO.getCanOrAbsPath( file )
                             + " not found." );
                }
            catch ( Exception e )
                {
                out.println( e.getMessage()
                             + " in file "
                             + EIO.getCanOrAbsPath( file ) );
                }
            } // end for
        } // end main
    }