scrub_divinum

#! /usr/bin/env php
<?php

array_shift( $argv );

if( $argc < 2 )
{
    $argv = glob( __DIR__ . "/M*.html" );
    //echo "defaulting to everything " . implode( ', ', $argv );
}

$index = "<html><head><title>Liturgy for ";
$links = array();
foreach( $argv as $file )
{
    $f = f( $file );
    if( !$f ) continue;
    $links[] = $f;
    run( $f );
}
$index .= implode( ', ', $links ) . '</title></head><body>';
foreach( $links as $d )
{
    $index .= "<div><a href=\"Horas-{$d}.html\">Horas-{$d}.html</a></div>";
}
$index .= '</body></html>';

file_put_contents( 'index.html', $index );

function f( $arg )
{
    if( preg_match("#([0-9]{1,2}-[0-9]{1,2})#", $arg, $matches ) )
        return $matches[ 0 ];
    return false;
}


function run( $targ  )
{
    $d = __DIR__ . DIRECTORY_SEPARATOR;
    $dir = glob( $d . $targ . DIRECTORY_SEPARATOR . '*' );

    @mkdir( "{$d}{$targ}_clean" );

    foreach( $dir as $fl )
    {
        process( $fl, dirname( $fl ) . '_clean' .
            DIRECTORY_SEPARATOR . basename( $fl ), $targ );
    }

    process( "${d}Master-$targ.html", "${d}Horas-$targ.html", $targ );
}

function process( $fl, $out, $targ )
{
    $cont = file_get_contents( $fl );
    $dom = new DomDocument();
    @$dom->loadHTML( $cont );
    remove( $dom, "font", "td", "tr", "tbody", "table" );
    clean_references( $dom, $targ );
    $cont = @$dom->saveXML();
    file_put_contents( $out, $cont );
}

function remove( $dom )
{
    $args = array_slice( func_get_args(), 1 );
    foreach( $args as $tag )
    {
        while( $dom->getElementsByTagName( $tag )->length )
        {
            foreach( $dom->getElementsByTagName( $tag ) as $node )
            {
                $parent = $node->parentNode;

                $nodes = array();
                foreach( $node->childNodes as $child )
                {
                    $nodes[] = $child;
                }
                graduate( $nodes, $node, $parent );
                $parent->removeChild( $node );
            }
        }
    }
}

function graduate( $list, $parent, $grandparent )
{
    foreach( $list as $node )
        $grandparent->insertBefore( $node, $parent );
}

function clean_references( $data, $targ )
{
    foreach( $data->getElementsByTagName( "a" ) as $a )
    {
        $href = $a->getAttribute( "href" );
        $href = str_replace( $targ, "{$targ}_clean", $href );
        $a->setAttribute( "href", $href );
    }
}