Validate spreadsheet file and it's row data as UTF-8 (csv, xls, etc)

csv_parser.php

<?php

const MIMES = array(
    'text/csv',
    'text/plain', 
    'application/csv',
    'application/vnd.ms-excel', 
    'application/vnd.ms-excel.sheet.binary.macroEnabled.12',
    'application/vnd.ms-excel.sheet.macroEnabled.12',
    'application/vnd.openxmlformats-officedocument.spreadsheetml.sheet'
);

/**
 * [Running from command line. Comment out if you want to run somewhere else]
 * 
 * Check to see if the user has specified any arguments to for the script to consume.
 * If no arguments were passed to the script, return the error.
 * Check to see if the file path specified in argument 1 is valid.
 * If the path is not valid, return the error.
 */
if ($argc < 2 )
{
    die("Error: Missing arguments. Please specify the full file path and name you wish to scan.\r\n");
}
else 
{
    $arguments = $argv;
    $file = $arguments[1];

    check_file_is_csv($file);
}
//

/**
 * Validate the encoding of the file specified against the UTF-8 format.
 * If not valid UTF-8 format, return the exact or closes file format identified.
 */
if (mb_check_encoding(file_get_contents($file), 'UTF-8')) 
{
    echo "File format encoding is valid UTF-8. \r\n";
}
else
{
    $format = exec('file -I ' . $file, $output);
    $charset = preg_match('/charset=(.*)/', $format, $matches);
    echo "File format encoding is not valid UTF-8, it is " .$matches[1]. "!\r\n"; 
}

/**
 * Loop through the specified file 
 */
$row = 0;
if (($handle = fopen($file,"r")) !== FALSE) {
  while (($data = fgetcsv($handle, 1000, ",")) !== FALSE) {

    if ($row !== 0) 
    {
        for ($c=0; $c < count($data); $c++) {

            $unsupported_char = has_unsupported_char($data[$c]);
            if ($unsupported_char)
            {
                /**
                 * Take into account that we remove the first row due to it being a header row.
                 * Take into account that the $col is array and starts from 0.
                 */
                $actual_row = $row + 1;
                $actual_col = $c + 1;
                echo "Unsupported character '$unsupported_char' identified found on row #$actual_row in column $c next to $data[$c].\n";
            }

            $enc = mb_detect_encoding($data[$c], 'UTF-8');
            if($enc !== "UTF-8")
            {
                echo "None UTF-8 character identified found on #$row in column $c!\n";   
            }
        }
    } 
    $row++;
  }

  echo "\r\n";
  echo "Total of $row rows checked.";
  fclose($handle);
}

/**
 * Functions
 */

function has_unsupported_char($string){
    if($string == "") return FALSE;
    $string_array = str_split($string);
    
    for($i=0; $i < count($string_array); $i++)
    {
        $byte = ord($string_array[$i]);

        if($byte > 127) return $string_array[$i];
    }
}

function check_file_is_csv($file)
{
    $finfo = finfo_open(FILEINFO_MIME_TYPE);
    $mime = @finfo_file($finfo, $file);

    if(!$mime)
    {
        die("Unable to locate file specified on the system.");
    }

    if(!in_array($mime, MIMES))
    {
        die("Sorry, this mime type not allowed");
    }

    finfo_close($finfo);

    return;
}
?>