asko-get.php

<?php

$url = "http://www.asko-nabytek.cz/katalog-letak-asko";
$template = array
(
    "title" => "ASKO nábytek",
    "descr" => null,
    "datetime_end" => null,
    "datetime_begin" => null,
    "file_orig" => null,
    "source_url" => $url
);

$html = file_get_contents($url);

preg_match_all("/katalog.asko.cz\/([^\"]+)/",$html,$matches_codes);
preg_match_all("/Platnost:<\/strong> ([^<]+)/",$html,$matches_validity);
preg_match_all("/Leták je platný pro(.+)/",$html,$matches_validity_type);

$buffer = array();
for($i=1; $i<=count($matches_codes[1]); $i++)
{
    // kazdy sudy ulozime do bufferu
    if($i%2==0)
    {  
        $buffer[] = $matches_codes[1][$i-1];
    }
}

$find_dates = function($str)
{
    $date1 = mb_substr($str, 0, 13);
    $date2 = mb_substr($str, 19, 13);
    preg_match("/(\d{1,2}).[^\d]+(\d{1,2}).[^\d]+(\d{4})/", $date1, $m1); 
    preg_match("/(\d{1,2}).[^\d]+(\d{1,2}).[^\d]+(\d{4})/", $date2, $m2); 
    return array(strtotime("$m1[3]/$m1[2]/$m1[1]"), strtotime("$m2[3]/$m2[2]/$m2[1]"));
};

$out = array();
for($i=0; $i<count($buffer); $i++)
{
    $item = $template;
    $dates = $find_dates($matches_validity[1][$i]);

    $item['datetime_begin'] = date("Y-m-d H:i:s",$dates[0]);
    $item['datetime_end'] = date("Y-m-d H:i:s",strtotime("-1 second", $dates[1]));
    $item['file_orig'] = "http://katalog.asko.cz/".$buffer[$i]."/includes/".str_replace("-","_",$buffer[$i]).".pdf";
    $item['descr'] = "Platnost letáku od ".date("j.m.Y", $dates[0])." do ".date("j.m.Y", $dates[1])." pro ".(preg_match("/E-shop/",$matches_validity_type[1][$i]) ? 'E-Shop' : 'všechny pobočky');
    $out[] = $item;
}

var_dump($out);