RISReader.php

Go to the documentation of this file.

<?php
namespace LibRIS;
 
class RISReader {
 
        const RIS_EOL = "\r\n";
        const LINE_REGEX = '/^(([A-Z1-9]{2})\s+-(.*))|(.*)$/';
        protected $data = NULL;
 
 
        public function __construct($options = array()) {
        }
 
 
        public function parseFile($filename, $context = NULL) {
                if (! is_file($filename)) {
                        throw new ParseException(sprintf('File %s not found.', htmlentities($filename)));
                }
                $flags = FILE_SKIP_EMPTY_LINES | FILE_TEXT;
                $contents = file($filename, $flags, $context);
                $this->parseArray($contents);
        }
 
 
        public function parseString($string) {
                $contents = explode(RISReader::RIS_EOL, $string);
                $this->parseArray($contents);
        }
 
 
        protected function parseArray($lines) {
                $recordset = array();
                // Do any cleaning and normalizing.
                $this->cleanData($lines);
                $record = array();
                $lastTag = NULL;
                foreach ($lines as $line) {
                        $line = trim($line);
                        $matches = array();
                        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;
                                $record = array();
                        } elseif (! empty($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 (! empty($record)) {
                        $recordset[] = $record;
                }
                $this->data = $recordset;
        }
 
 
        public function getRecords() {
                return $this->data;
        }
 
 
        public function printRecords() {
                $format = "%s:\n\t%s\n";
                foreach ($this->data as $record) {
                        foreach ($record as $key => $values) {
                                foreach ($values as $value) {
                                        printf($format, RISTags::describeTag($key), $value);
                                }
                        }
                        print PHP_EOL;
                }
        }
 
 
        protected function cleanData(&$lines) {
                if (empty($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 = $lines[0];
                if (substr($first, 0, 3) == pack('CCC', 0xef, 0xbb, 0xbf)) {
                        $lines[0] = substr($first, 3);
                }
        }
}