class.ilRisReader.php

Go to the documentation of this file.

<?php
 
class ilRisReader
{
    public const RIS_EOL = "\r\n";
    public const LINE_REGEX = '/^(([A-Z1-9]{2})\s+-(.*))|(.*)$/';
 
    protected array $data = [];
 
    public function __construct($options = [])
    {
    }
 
    public function parseString(string $string): void
    {
        $contents = explode(self::RIS_EOL, $string);
        $this->parseArray($contents);
    }
 
    protected function parseArray(array $lines): void
    {
        $recordset = [];
 
        // Do any cleaning and normalizing.
        $this->cleanData($lines);
 
        $record = [];
        $lastTag = null;
        foreach ($lines as $line) {
            $line = trim((string) $line);
            $matches = [];
 
            preg_match(self::LINE_REGEX, $line, $matches);
            if (!empty($matches[3])) {
                $lastTag = $matches[2];
                $record[$matches[2]][] = trim($matches[3]);
            } // End record and prep a new one.
            elseif (!empty($matches[2]) && $matches[2] === 'ER') {
                $lastTag = null;
                $recordset[] = $record;
                $record = [];
            } elseif (!empty($matches[4])) {
                // Append to the last one.
                // We skip leading info (like BOMs).
                if (!empty($lastTag)) {
                    $lastEntry = count($record[$lastTag]) - 1;
                    // We trim because some encoders add tabs or multiple spaces.
                    // Standard is silent on how this should be handled.
                    $record[$lastTag][$lastEntry] .= ' ' . trim($matches[4]);
                }
            }
        }
        if ($record !== []) {
            $recordset[] = $record;
        }
 
        $this->data = $recordset;
    }
 
    public function getRecords(): array
    {
        return $this->data;
    }
 
    protected function cleanData(array &$lines): void
    {
        if ($lines === []) {
            return;
        }
 
        // Currently, we only need to strip a BOM if it exists.
        // Thanks to Derik Badman (http://madinkbeard.com/) for finding the
        // bug and suggesting this fix:
        // http://blog.philipp-michels.de/?p=32
        $first = $lines[0];
        if (substr((string) $first, 0, 3) === pack('CCC', 0xef, 0xbb, 0xbf)) {
            $lines[0] = substr((string) $first, 3);
        }
    }
 
}