Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
98.57% |
138 / 140 |
|
80.00% |
8 / 10 |
CRAP | |
0.00% |
0 / 1 |
| MarcXml | |
98.57% |
138 / 140 |
|
80.00% |
8 / 10 |
40 | |
0.00% |
0 / 1 |
| canParse | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| canParseCollection | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| canParseCollectionFile | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
3 | |||
| collectionFromString | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
2 | |||
| fromString | |
100.00% |
18 / 18 |
|
100.00% |
1 / 1 |
6 | |||
| toString | |
97.92% |
47 / 48 |
|
0.00% |
0 / 1 |
7 | |||
| loadXML | |
96.00% |
24 / 25 |
|
0.00% |
0 / 1 |
6 | |||
| openCollectionFile | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
| rewind | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
2 | |||
| getNextRecord | |
100.00% |
27 / 27 |
|
100.00% |
1 / 1 |
10 | |||
| 1 | <?php |
| 2 | |
| 3 | /** |
| 4 | * MARCXML format support class. |
| 5 | * |
| 6 | * PHP version 7 |
| 7 | * |
| 8 | * Copyright (C) The National Library of Finland 2020-2022. |
| 9 | * |
| 10 | * This program is free software; you can redistribute it and/or modify |
| 11 | * it under the terms of the GNU General Public License version 2, |
| 12 | * as published by the Free Software Foundation. |
| 13 | * |
| 14 | * This program is distributed in the hope that it will be useful, |
| 15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 17 | * GNU General Public License for more details. |
| 18 | * |
| 19 | * You should have received a copy of the GNU General Public License |
| 20 | * along with this program; if not, write to the Free Software |
| 21 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 22 | * |
| 23 | * @category VuFind |
| 24 | * @package MARC |
| 25 | * @author Ere Maijala <ere.maijala@helsinki.fi> |
| 26 | * @license http://opensource.org/licenses/gpl-2.0.php GNU General Public License |
| 27 | * @link https://vufind.org/wiki/development:plugins:record_drivers Wiki |
| 28 | */ |
| 29 | |
| 30 | namespace VuFind\Marc\Serialization; |
| 31 | |
| 32 | use function array_slice; |
| 33 | use function count; |
| 34 | use function is_string; |
| 35 | use function strlen; |
| 36 | |
| 37 | /** |
| 38 | * MARCXML format support class. |
| 39 | * |
| 40 | * @category VuFind |
| 41 | * @package MARC |
| 42 | * @author Ere Maijala <ere.maijala@helsinki.fi> |
| 43 | * @license http://opensource.org/licenses/gpl-2.0.php GNU General Public License |
| 44 | * @link https://vufind.org/wiki/development:plugins:record_drivers Wiki |
| 45 | */ |
| 46 | class MarcXml extends AbstractSerializationFile implements SerializationInterface |
| 47 | { |
| 48 | /** |
| 49 | * Current file |
| 50 | * |
| 51 | * @var string |
| 52 | */ |
| 53 | protected $fileName = ''; |
| 54 | |
| 55 | /** |
| 56 | * XML Reader for current file |
| 57 | * |
| 58 | * @var \XMLReader |
| 59 | */ |
| 60 | protected $xml = null; |
| 61 | |
| 62 | /** |
| 63 | * Current XML element path |
| 64 | * |
| 65 | * @var array |
| 66 | */ |
| 67 | protected $currentXmlPath = []; |
| 68 | |
| 69 | /** |
| 70 | * Check if this class can parse the given MARC string |
| 71 | * |
| 72 | * @param string $marc MARC |
| 73 | * |
| 74 | * @return bool |
| 75 | */ |
| 76 | public static function canParse(string $marc): bool |
| 77 | { |
| 78 | // A pretty naïve check, but it's enough to tell the different formats apart |
| 79 | return strncmp(trim($marc), '<', 1) === 0; |
| 80 | } |
| 81 | |
| 82 | /** |
| 83 | * Check if the serialization class can parse the given MARC collection string |
| 84 | * |
| 85 | * @param string $marc MARC |
| 86 | * |
| 87 | * @return bool |
| 88 | */ |
| 89 | public static function canParseCollection(string $marc): bool |
| 90 | { |
| 91 | // A pretty naïve check, but it's enough to tell the different formats apart |
| 92 | return strncmp(trim($marc), '<', 1) === 0; |
| 93 | } |
| 94 | |
| 95 | /** |
| 96 | * Check if the serialization class can parse the given MARC collection file |
| 97 | * |
| 98 | * @param string $file File name |
| 99 | * |
| 100 | * @return bool |
| 101 | */ |
| 102 | public static function canParseCollectionFile(string $file): bool |
| 103 | { |
| 104 | if (false === ($f = @fopen($file, 'r'))) { |
| 105 | throw new \Exception("Cannot open file '$file' for reading"); |
| 106 | } |
| 107 | do { |
| 108 | $s = ltrim(fgets($f, 10)); |
| 109 | } while (!$s && !feof($f)); |
| 110 | fclose($f); |
| 111 | |
| 112 | return self::canParseCollection($s); |
| 113 | } |
| 114 | |
| 115 | /** |
| 116 | * Parse MARC collection from a string into an array |
| 117 | * |
| 118 | * @param string $collection MARC record collection in the format supported by |
| 119 | * the serialization class |
| 120 | * |
| 121 | * @throws \Exception |
| 122 | * @return array |
| 123 | */ |
| 124 | public static function collectionFromString(string $collection): array |
| 125 | { |
| 126 | $xml = static::loadXML(trim($collection)); |
| 127 | $results = []; |
| 128 | foreach ($xml->record as $record) { |
| 129 | $results[] = $record->asXML(); |
| 130 | } |
| 131 | return $results; |
| 132 | } |
| 133 | |
| 134 | /** |
| 135 | * Parse MARCXML string |
| 136 | * |
| 137 | * @param string $marc MARCXML |
| 138 | * |
| 139 | * @throws \Exception |
| 140 | * @return array |
| 141 | */ |
| 142 | public static function fromString(string $marc): array |
| 143 | { |
| 144 | $xml = static::loadXML(trim($marc)); |
| 145 | |
| 146 | // Move to the record element if we were given a collection |
| 147 | if ($xml->record) { |
| 148 | $xml = $xml->record; |
| 149 | } |
| 150 | |
| 151 | $leader = isset($xml->leader) ? (string)$xml->leader[0] : ''; |
| 152 | $fields = []; |
| 153 | |
| 154 | foreach ($xml->controlfield as $field) { |
| 155 | $fields[] = [(string)$field['tag'] => (string)$field]; |
| 156 | } |
| 157 | |
| 158 | foreach ($xml->datafield as $field) { |
| 159 | $newField = [ |
| 160 | 'ind1' => str_pad((string)$field['ind1'], 1), |
| 161 | 'ind2' => str_pad((string)$field['ind2'], 1), |
| 162 | 'subfields' => [], |
| 163 | ]; |
| 164 | foreach ($field->subfield as $subfield) { |
| 165 | $newField['subfields'][] |
| 166 | = [(string)$subfield['code'] => (string)$subfield]; |
| 167 | } |
| 168 | $fields[] = [(string)$field['tag'] => $newField]; |
| 169 | } |
| 170 | |
| 171 | return compact('leader', 'fields'); |
| 172 | } |
| 173 | |
| 174 | /** |
| 175 | * Convert record to a MARCXML string |
| 176 | * |
| 177 | * @param array $record Record data |
| 178 | * |
| 179 | * @throws \Exception |
| 180 | * @return string |
| 181 | */ |
| 182 | public static function toString(array $record): string |
| 183 | { |
| 184 | $xml = new \XMLWriter(); |
| 185 | $xml->openMemory(); |
| 186 | $xml->setIndent(true); |
| 187 | $xml->startDocument('1.0', 'UTF-8'); |
| 188 | $xml->startElementNs(null, 'collection', 'http://www.loc.gov/MARC21/slim'); |
| 189 | $xml->writeAttribute( |
| 190 | 'xmlns:xsi', |
| 191 | 'http://www.w3.org/2001/XMLSchema-instance' |
| 192 | ); |
| 193 | $xml->writeAttribute( |
| 194 | 'xsi:schemaLocation', |
| 195 | 'http://www.loc.gov/MARC21/slim' |
| 196 | . ' http://www.loc.gov/standards/marcxml/schema/MARC21slim.xsd' |
| 197 | ); |
| 198 | $xml->startElement('record'); |
| 199 | if ($record['leader']) { |
| 200 | $xml->writeElement('leader', $record['leader']); |
| 201 | } |
| 202 | |
| 203 | foreach ($record['fields'] as $fieldData) { |
| 204 | $tag = (string)key($fieldData); |
| 205 | $field = current($fieldData); |
| 206 | if (is_string($field)) { |
| 207 | $xml->startElement('controlfield'); |
| 208 | $xml->writeAttribute('tag', $tag); |
| 209 | $xml->text($field); |
| 210 | $xml->endElement(); |
| 211 | } else { |
| 212 | $xml->startElement('datafield'); |
| 213 | $xml->writeAttribute('tag', $tag); |
| 214 | $xml->writeAttribute('ind1', $field['ind1']); |
| 215 | $xml->writeAttribute('ind2', $field['ind2']); |
| 216 | foreach ($field['subfields'] ?? [] as $subfield) { |
| 217 | $subfieldData = current($subfield); |
| 218 | $subfieldCode = (string)key($subfield); |
| 219 | if ($subfieldData == '') { |
| 220 | continue; |
| 221 | } |
| 222 | $xml->startElement('subfield'); |
| 223 | $xml->writeAttribute('code', $subfieldCode); |
| 224 | $xml->text($subfieldData); |
| 225 | $xml->endElement(); |
| 226 | } |
| 227 | $xml->endElement(); |
| 228 | } |
| 229 | } |
| 230 | $xml->endElement(); |
| 231 | $xml->endDocument(); |
| 232 | |
| 233 | // Strip illegal characters from XML: |
| 234 | $xmlString = preg_replace( |
| 235 | '/[^\x{0009}\x{000a}\x{000d}\x{0020}-\x{D7FF}\x{E000}-\x{FFFD}]+/u', |
| 236 | '', |
| 237 | $xml->outputMemory(true) |
| 238 | ); |
| 239 | if ($xmlString === null) { |
| 240 | throw new \Exception('Error processing XML'); |
| 241 | } |
| 242 | return $xmlString; |
| 243 | } |
| 244 | |
| 245 | /** |
| 246 | * Load XML into SimpleXMLElement |
| 247 | * |
| 248 | * @param string $xml XML |
| 249 | * |
| 250 | * @throws \Exception |
| 251 | * @return \SimpleXMLElement |
| 252 | */ |
| 253 | protected static function loadXML(string $xml): \SimpleXMLElement |
| 254 | { |
| 255 | // Make sure we have an XML prolog with proper encoding: |
| 256 | $xmlHead = '<?xml version'; |
| 257 | if (strcasecmp(substr($xml, 0, strlen($xmlHead)), $xmlHead) === 0) { |
| 258 | $decl = substr($xml, 0, strpos($xml, '?>')); |
| 259 | if (strstr($decl, 'encoding') === false) { |
| 260 | $xml = $decl . ' encoding="utf-8"' . substr($xml, strlen($decl)); |
| 261 | } |
| 262 | } else { |
| 263 | $xml = '<?xml version="1.0" encoding="utf-8"?>' . "\n\n$xml"; |
| 264 | } |
| 265 | $saveUseErrors = libxml_use_internal_errors(true); |
| 266 | try { |
| 267 | libxml_clear_errors(); |
| 268 | $doc = simplexml_load_string( |
| 269 | $xml, |
| 270 | \SimpleXMLElement::class, |
| 271 | LIBXML_COMPACT |
| 272 | ); |
| 273 | if (false === $doc) { |
| 274 | $errors = libxml_get_errors(); |
| 275 | $messageParts = []; |
| 276 | foreach ($errors as $error) { |
| 277 | $messageParts[] = '[' . $error->line . ':' . $error->column |
| 278 | . '] Error ' . $error->code . ': ' . $error->message; |
| 279 | } |
| 280 | throw new \Exception(implode("\n", $messageParts)); |
| 281 | } |
| 282 | libxml_use_internal_errors($saveUseErrors); |
| 283 | return $doc; |
| 284 | } catch (\Exception $e) { |
| 285 | libxml_use_internal_errors($saveUseErrors); |
| 286 | throw $e; |
| 287 | } |
| 288 | } |
| 289 | |
| 290 | /** |
| 291 | * Open a collection file |
| 292 | * |
| 293 | * @param string $file File name |
| 294 | * |
| 295 | * @return void |
| 296 | * |
| 297 | * @throws \Exception |
| 298 | */ |
| 299 | public function openCollectionFile(string $file): void |
| 300 | { |
| 301 | $this->fileName = $file; |
| 302 | $this->xml = new \XMLReader(); |
| 303 | $result = $this->xml->open($file); |
| 304 | if (false === $result) { |
| 305 | throw new \Exception("Cannot open file '$file' for reading"); |
| 306 | } |
| 307 | $this->currentXmlPath = []; |
| 308 | } |
| 309 | |
| 310 | /** |
| 311 | * Rewind the collection file |
| 312 | * |
| 313 | * @return void |
| 314 | * |
| 315 | * @throws \Exception |
| 316 | */ |
| 317 | public function rewind(): void |
| 318 | { |
| 319 | if ('' === $this->fileName) { |
| 320 | throw new \Exception('Collection file not open'); |
| 321 | } |
| 322 | $this->openCollectionFile($this->fileName); |
| 323 | } |
| 324 | |
| 325 | /** |
| 326 | * Get next record from the file or an empty string on EOF |
| 327 | * |
| 328 | * @return string |
| 329 | * |
| 330 | * @throws \Exception |
| 331 | */ |
| 332 | public function getNextRecord(): string |
| 333 | { |
| 334 | if (null === $this->xml) { |
| 335 | throw new \Exception('Collection file not open'); |
| 336 | } |
| 337 | while ($this->xml->read()) { |
| 338 | if ($this->xml->nodeType !== \XMLReader::ELEMENT) { |
| 339 | continue; |
| 340 | } |
| 341 | $this->currentXmlPath |
| 342 | = array_slice($this->currentXmlPath, 0, $this->xml->depth); |
| 343 | $this->currentXmlPath[] = $this->xml->name; |
| 344 | $ns = $this->xml->namespaceURI; |
| 345 | |
| 346 | $currentPathString = '/' . implode('/', $this->currentXmlPath); |
| 347 | $pathOk = '/collection/record' === $currentPathString; |
| 348 | $namespaceOk = !$ns || 'http://www.loc.gov/MARC21/slim' === $ns; |
| 349 | if (!$pathOk || !$namespaceOk) { |
| 350 | if (count($this->currentXmlPath) === 2) { |
| 351 | if (!$pathOk) { |
| 352 | $this->message( |
| 353 | "Unknown element \"$currentPathString\"", |
| 354 | E_NOTICE |
| 355 | ); |
| 356 | } elseif (!$namespaceOk) { |
| 357 | $this->message( |
| 358 | "Unknown namespace \"$ns\" for element \"" |
| 359 | . $currentPathString . '"', |
| 360 | E_NOTICE |
| 361 | ); |
| 362 | } |
| 363 | } |
| 364 | continue; |
| 365 | } |
| 366 | return $this->xml->readOuterXML(); |
| 367 | } |
| 368 | return ''; |
| 369 | } |
| 370 | } |