PHPOffice /
PhpSpreadsheet
| Total Complexity | 103 |
| Total Lines | 638 |
| Duplicated Lines | 0 % |
| Coverage | 69.37% |
| Changes | 0 | ||
Complex classes like Html often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
While breaking up the class, it is a good idea to analyze how other classes use Html, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 17 | class Html extends BaseReader |
||
| 18 | { |
||
| 19 | /** |
||
| 20 | * Sample size to read to determine if it's HTML or not. |
||
| 21 | */ |
||
| 22 | const TEST_SAMPLE_SIZE = 2048; |
||
| 23 | |||
| 24 | /** |
||
| 25 | * Input encoding. |
||
| 26 | * |
||
| 27 | * @var string |
||
| 28 | */ |
||
| 29 | protected $inputEncoding = 'ANSI'; |
||
| 30 | |||
| 31 | /** |
||
| 32 | * Sheet index to read. |
||
| 33 | * |
||
| 34 | * @var int |
||
| 35 | */ |
||
| 36 | protected $sheetIndex = 0; |
||
| 37 | |||
| 38 | /** |
||
| 39 | * Formats. |
||
| 40 | * |
||
| 41 | * @var array |
||
| 42 | */ |
||
| 43 | protected $formats = [ |
||
| 44 | 'h1' => [ |
||
| 45 | 'font' => [ |
||
| 46 | 'bold' => true, |
||
| 47 | 'size' => 24, |
||
| 48 | ], |
||
| 49 | ], // Bold, 24pt |
||
| 50 | 'h2' => [ |
||
| 51 | 'font' => [ |
||
| 52 | 'bold' => true, |
||
| 53 | 'size' => 18, |
||
| 54 | ], |
||
| 55 | ], // Bold, 18pt |
||
| 56 | 'h3' => [ |
||
| 57 | 'font' => [ |
||
| 58 | 'bold' => true, |
||
| 59 | 'size' => 13.5, |
||
| 60 | ], |
||
| 61 | ], // Bold, 13.5pt |
||
| 62 | 'h4' => [ |
||
| 63 | 'font' => [ |
||
| 64 | 'bold' => true, |
||
| 65 | 'size' => 12, |
||
| 66 | ], |
||
| 67 | ], // Bold, 12pt |
||
| 68 | 'h5' => [ |
||
| 69 | 'font' => [ |
||
| 70 | 'bold' => true, |
||
| 71 | 'size' => 10, |
||
| 72 | ], |
||
| 73 | ], // Bold, 10pt |
||
| 74 | 'h6' => [ |
||
| 75 | 'font' => [ |
||
| 76 | 'bold' => true, |
||
| 77 | 'size' => 7.5, |
||
| 78 | ], |
||
| 79 | ], // Bold, 7.5pt |
||
| 80 | 'a' => [ |
||
| 81 | 'font' => [ |
||
| 82 | 'underline' => true, |
||
| 83 | 'color' => [ |
||
| 84 | 'argb' => Color::COLOR_BLUE, |
||
| 85 | ], |
||
| 86 | ], |
||
| 87 | ], // Blue underlined |
||
| 88 | 'hr' => [ |
||
| 89 | 'borders' => [ |
||
| 90 | 'bottom' => [ |
||
| 91 | 'borderStyle' => Border::BORDER_THIN, |
||
| 92 | 'color' => [ |
||
| 93 | Color::COLOR_BLACK, |
||
| 94 | ], |
||
| 95 | ], |
||
| 96 | ], |
||
| 97 | ], // Bottom border |
||
| 98 | ]; |
||
| 99 | |||
| 100 | protected $rowspan = []; |
||
| 101 | |||
| 102 | /** |
||
| 103 | * Create a new HTML Reader instance. |
||
| 104 | */ |
||
| 105 | 18 | public function __construct() |
|
| 108 | 18 | } |
|
| 109 | |||
| 110 | /** |
||
| 111 | * Validate that the current file is an HTML file. |
||
| 112 | * |
||
| 113 | * @param string $pFilename |
||
| 114 | * |
||
| 115 | * @return bool |
||
| 116 | */ |
||
| 117 | 16 | public function canRead($pFilename) |
|
| 118 | { |
||
| 119 | // Check if file exists |
||
| 120 | try { |
||
| 121 | 16 | $this->openFile($pFilename); |
|
| 122 | } catch (Exception $e) { |
||
| 123 | return false; |
||
| 124 | } |
||
| 125 | |||
| 126 | 16 | $beginning = $this->readBeginning(); |
|
| 127 | 16 | $startWithTag = self::startsWithTag($beginning); |
|
| 128 | 16 | $containsTags = self::containsTags($beginning); |
|
| 129 | 16 | $endsWithTag = self::endsWithTag($this->readEnding()); |
|
| 130 | |||
| 131 | 16 | fclose($this->fileHandle); |
|
|
1 ignored issue
–
show
|
|||
| 132 | |||
| 133 | 16 | return $startWithTag && $containsTags && $endsWithTag; |
|
| 134 | } |
||
| 135 | |||
| 136 | 16 | private function readBeginning() |
|
| 137 | { |
||
| 138 | 16 | fseek($this->fileHandle, 0); |
|
| 139 | |||
| 140 | 16 | return fread($this->fileHandle, self::TEST_SAMPLE_SIZE); |
|
| 141 | } |
||
| 142 | |||
| 143 | 16 | private function readEnding() |
|
| 144 | { |
||
| 145 | 16 | $meta = stream_get_meta_data($this->fileHandle); |
|
| 146 | 16 | $filename = $meta['uri']; |
|
| 147 | |||
| 148 | 16 | $size = filesize($filename); |
|
| 149 | 16 | if ($size === 0) { |
|
| 150 | 1 | return ''; |
|
| 151 | } |
||
| 152 | |||
| 153 | 15 | $blockSize = self::TEST_SAMPLE_SIZE; |
|
| 154 | 15 | if ($size < $blockSize) { |
|
| 155 | 2 | $blockSize = $size; |
|
| 156 | } |
||
| 157 | |||
| 158 | 15 | fseek($this->fileHandle, $size - $blockSize); |
|
| 159 | |||
| 160 | 15 | return fread($this->fileHandle, $blockSize); |
|
| 161 | } |
||
| 162 | |||
| 163 | 16 | private static function startsWithTag($data) |
|
| 164 | { |
||
| 165 | 16 | return '<' === substr(trim($data), 0, 1); |
|
| 166 | } |
||
| 167 | |||
| 168 | 16 | private static function endsWithTag($data) |
|
| 169 | { |
||
| 170 | 16 | return '>' === substr(trim($data), -1, 1); |
|
| 171 | } |
||
| 172 | |||
| 173 | 16 | private static function containsTags($data) |
|
| 176 | } |
||
| 177 | |||
| 178 | /** |
||
| 179 | * Loads Spreadsheet from file. |
||
| 180 | * |
||
| 181 | * @param string $pFilename |
||
| 182 | * |
||
| 183 | * @throws Exception |
||
| 184 | * |
||
| 185 | * @return Spreadsheet |
||
| 186 | */ |
||
| 187 | 9 | public function load($pFilename) |
|
| 188 | { |
||
| 189 | // Create new Spreadsheet |
||
| 190 | 9 | $spreadsheet = new Spreadsheet(); |
|
| 191 | |||
| 192 | // Load into this instance |
||
| 193 | 9 | return $this->loadIntoExisting($pFilename, $spreadsheet); |
|
| 194 | } |
||
| 195 | |||
| 196 | /** |
||
| 197 | * Set input encoding. |
||
| 198 | * |
||
| 199 | * @param string $pValue Input encoding, eg: 'ANSI' |
||
| 200 | * |
||
| 201 | * @return Html |
||
| 202 | */ |
||
| 203 | public function setInputEncoding($pValue) |
||
| 204 | { |
||
| 205 | $this->inputEncoding = $pValue; |
||
| 206 | |||
| 207 | return $this; |
||
| 208 | } |
||
| 209 | |||
| 210 | /** |
||
| 211 | * Get input encoding. |
||
| 212 | * |
||
| 213 | * @return string |
||
| 214 | */ |
||
| 215 | public function getInputEncoding() |
||
| 216 | { |
||
| 217 | return $this->inputEncoding; |
||
| 218 | } |
||
| 219 | |||
| 220 | // Data Array used for testing only, should write to Spreadsheet object on completion of tests |
||
| 221 | protected $dataArray = []; |
||
| 222 | |||
| 223 | protected $tableLevel = 0; |
||
| 224 | |||
| 225 | protected $nestedColumn = ['A']; |
||
| 226 | |||
| 227 | 9 | protected function setTableStartColumn($column) |
|
| 228 | { |
||
| 229 | 9 | if ($this->tableLevel == 0) { |
|
| 230 | 9 | $column = 'A'; |
|
| 231 | } |
||
| 232 | 9 | ++$this->tableLevel; |
|
| 233 | 9 | $this->nestedColumn[$this->tableLevel] = $column; |
|
| 234 | |||
| 235 | 9 | return $this->nestedColumn[$this->tableLevel]; |
|
| 236 | } |
||
| 237 | |||
| 238 | 9 | protected function getTableStartColumn() |
|
| 241 | } |
||
| 242 | |||
| 243 | 9 | protected function releaseTableStartColumn() |
|
| 244 | { |
||
| 245 | 9 | --$this->tableLevel; |
|
| 246 | |||
| 247 | 9 | return array_pop($this->nestedColumn); |
|
| 248 | } |
||
| 249 | |||
| 250 | 9 | protected function flushCell(Worksheet $sheet, $column, $row, &$cellContent) |
|
| 251 | { |
||
| 252 | 9 | if (is_string($cellContent)) { |
|
| 253 | // Simple String content |
||
| 254 | 9 | if (trim($cellContent) > '') { |
|
| 255 | // Only actually write it if there's content in the string |
||
| 256 | // Write to worksheet to be done here... |
||
| 257 | // ... we return the cell so we can mess about with styles more easily |
||
| 258 | 9 | $sheet->setCellValue($column . $row, $cellContent); |
|
| 259 | 9 | $this->dataArray[$row][$column] = $cellContent; |
|
| 260 | } |
||
| 261 | } else { |
||
| 262 | // We have a Rich Text run |
||
| 263 | // TODO |
||
| 264 | $this->dataArray[$row][$column] = 'RICH TEXT: ' . $cellContent; |
||
| 265 | } |
||
| 266 | 9 | $cellContent = (string) ''; |
|
| 267 | 9 | } |
|
| 268 | |||
| 269 | /** |
||
| 270 | * @param DOMNode $element |
||
| 271 | * @param Worksheet $sheet |
||
| 272 | * @param int $row |
||
| 273 | * @param string $column |
||
| 274 | * @param string $cellContent |
||
| 275 | */ |
||
| 276 | 9 | protected function processDomElement(DOMNode $element, Worksheet $sheet, &$row, &$column, &$cellContent) |
|
| 515 | } |
||
| 516 | } |
||
| 517 | } |
||
| 518 | 9 | } |
|
| 519 | |||
| 520 | /** |
||
| 521 | * Loads PhpSpreadsheet from file into PhpSpreadsheet instance. |
||
| 522 | * |
||
| 523 | * @param string $pFilename |
||
| 524 | * @param Spreadsheet $spreadsheet |
||
| 525 | * |
||
| 526 | * @throws Exception |
||
| 527 | * |
||
| 528 | * @return Spreadsheet |
||
| 529 | */ |
||
| 530 | 9 | public function loadIntoExisting($pFilename, Spreadsheet $spreadsheet) |
|
| 531 | { |
||
| 532 | // Validate |
||
| 533 | 9 | if (!$this->canRead($pFilename)) { |
|
| 534 | throw new Exception($pFilename . ' is an Invalid HTML file.'); |
||
| 535 | } |
||
| 536 | |||
| 537 | // Create new sheet |
||
| 538 | 9 | while ($spreadsheet->getSheetCount() <= $this->sheetIndex) { |
|
| 539 | $spreadsheet->createSheet(); |
||
| 540 | } |
||
| 541 | 9 | $spreadsheet->setActiveSheetIndex($this->sheetIndex); |
|
| 542 | |||
| 543 | // Create a new DOM object |
||
| 544 | 9 | $dom = new DOMDocument(); |
|
| 545 | // Reload the HTML file into the DOM object |
||
| 546 | 9 | $loaded = $dom->loadHTML(mb_convert_encoding($this->securityScanFile($pFilename), 'HTML-ENTITIES', 'UTF-8')); |
|
| 547 | 9 | if ($loaded === false) { |
|
| 548 | throw new Exception('Failed to load ' . $pFilename . ' as a DOM Document'); |
||
| 549 | } |
||
| 550 | |||
| 551 | // Discard white space |
||
| 552 | 9 | $dom->preserveWhiteSpace = false; |
|
| 553 | |||
| 554 | 9 | $row = 0; |
|
| 555 | 9 | $column = 'A'; |
|
| 556 | 9 | $content = ''; |
|
| 557 | 9 | $this->rowspan = []; |
|
| 558 | 9 | $this->processDomElement($dom, $spreadsheet->getActiveSheet(), $row, $column, $content); |
|
| 559 | |||
| 560 | // Return |
||
| 561 | 9 | return $spreadsheet; |
|
| 562 | } |
||
| 563 | |||
| 564 | /** |
||
| 565 | * Get sheet index. |
||
| 566 | * |
||
| 567 | * @return int |
||
| 568 | */ |
||
| 569 | public function getSheetIndex() |
||
| 572 | } |
||
| 573 | |||
| 574 | /** |
||
| 575 | * Set sheet index. |
||
| 576 | * |
||
| 577 | * @param int $pValue Sheet index |
||
| 578 | * |
||
| 579 | * @return HTML |
||
| 580 | */ |
||
| 581 | public function setSheetIndex($pValue) |
||
| 586 | } |
||
| 587 | |||
| 588 | /** |
||
| 589 | * Scan theXML for use of <!ENTITY to prevent XXE/XEE attacks. |
||
| 590 | * |
||
| 591 | * @param string $xml |
||
| 592 | * |
||
| 593 | * @return string |
||
| 594 | */ |
||
| 595 | 9 | public function securityScan($xml) |
|
| 603 | } |
||
| 604 | |||
| 605 | /** |
||
| 606 | * Apply inline css inline style. |
||
| 607 | * |
||
| 608 | * NOTES : |
||
| 609 | * Currently only intended for td & th element, |
||
| 610 | * and only takes 'background-color' and 'color'; property with HEX color |
||
| 611 | * |
||
| 612 | * TODO : |
||
| 613 | * - Implement to other propertie, such as border |
||
| 614 | * |
||
| 615 | * @param Worksheet $sheet |
||
| 616 | * @param int $row |
||
| 617 | * @param string $column |
||
| 618 | * @param array $attributeArray |
||
| 619 | */ |
||
| 620 | 9 | private function applyInlineStyle(&$sheet, $row, $column, $attributeArray) |
|
| 659 |
Loading history...