PHPOffice /
PhpSpreadsheet
| Total Complexity | 101 |
| Total Lines | 627 |
| Duplicated Lines | 0 % |
| Coverage | 69.39% |
| 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 |
||
| 18 | class Html extends BaseReader |
||
| 19 | { |
||
| 20 | /** |
||
| 21 | * @var XmlScanner |
||
| 22 | */ |
||
| 23 | private $securityScanner; |
||
| 24 | |||
| 25 | /** |
||
| 26 | * Sample size to read to determine if it's HTML or not. |
||
| 27 | */ |
||
| 28 | const TEST_SAMPLE_SIZE = 2048; |
||
| 29 | |||
| 30 | /** |
||
| 31 | * Input encoding. |
||
| 32 | * |
||
| 33 | * @var string |
||
| 34 | */ |
||
| 35 | protected $inputEncoding = 'ANSI'; |
||
| 36 | |||
| 37 | /** |
||
| 38 | * Sheet index to read. |
||
| 39 | * |
||
| 40 | * @var int |
||
| 41 | */ |
||
| 42 | protected $sheetIndex = 0; |
||
| 43 | |||
| 44 | /** |
||
| 45 | * Formats. |
||
| 46 | * |
||
| 47 | * @var array |
||
| 48 | */ |
||
| 49 | protected $formats = [ |
||
| 50 | 'h1' => [ |
||
| 51 | 'font' => [ |
||
| 52 | 'bold' => true, |
||
| 53 | 'size' => 24, |
||
| 54 | ], |
||
| 55 | ], // Bold, 24pt |
||
| 56 | 'h2' => [ |
||
| 57 | 'font' => [ |
||
| 58 | 'bold' => true, |
||
| 59 | 'size' => 18, |
||
| 60 | ], |
||
| 61 | ], // Bold, 18pt |
||
| 62 | 'h3' => [ |
||
| 63 | 'font' => [ |
||
| 64 | 'bold' => true, |
||
| 65 | 'size' => 13.5, |
||
| 66 | ], |
||
| 67 | ], // Bold, 13.5pt |
||
| 68 | 'h4' => [ |
||
| 69 | 'font' => [ |
||
| 70 | 'bold' => true, |
||
| 71 | 'size' => 12, |
||
| 72 | ], |
||
| 73 | ], // Bold, 12pt |
||
| 74 | 'h5' => [ |
||
| 75 | 'font' => [ |
||
| 76 | 'bold' => true, |
||
| 77 | 'size' => 10, |
||
| 78 | ], |
||
| 79 | ], // Bold, 10pt |
||
| 80 | 'h6' => [ |
||
| 81 | 'font' => [ |
||
| 82 | 'bold' => true, |
||
| 83 | 'size' => 7.5, |
||
| 84 | ], |
||
| 85 | ], // Bold, 7.5pt |
||
| 86 | 'a' => [ |
||
| 87 | 'font' => [ |
||
| 88 | 'underline' => true, |
||
| 89 | 'color' => [ |
||
| 90 | 'argb' => Color::COLOR_BLUE, |
||
| 91 | ], |
||
| 92 | ], |
||
| 93 | ], // Blue underlined |
||
| 94 | 'hr' => [ |
||
| 95 | 'borders' => [ |
||
| 96 | 'bottom' => [ |
||
| 97 | 'borderStyle' => Border::BORDER_THIN, |
||
| 98 | 'color' => [ |
||
| 99 | Color::COLOR_BLACK, |
||
| 100 | ], |
||
| 101 | ], |
||
| 102 | ], |
||
| 103 | ], // Bottom border |
||
| 104 | ]; |
||
| 105 | |||
| 106 | protected $rowspan = []; |
||
| 107 | |||
| 108 | /** |
||
| 109 | * Create a new HTML Reader instance. |
||
| 110 | */ |
||
| 111 | 18 | public function __construct() |
|
| 112 | { |
||
| 113 | 18 | $this->readFilter = new DefaultReadFilter(); |
|
| 114 | 18 | $this->securityScanner = new XmlScanner('<!ENTITY'); |
|
| 115 | 18 | } |
|
| 116 | |||
| 117 | /** |
||
| 118 | * Validate that the current file is an HTML file. |
||
| 119 | * |
||
| 120 | * @param string $pFilename |
||
| 121 | * |
||
| 122 | * @return bool |
||
| 123 | */ |
||
| 124 | 16 | public function canRead($pFilename) |
|
| 125 | { |
||
| 126 | // Check if file exists |
||
| 127 | try { |
||
| 128 | 16 | $this->openFile($pFilename); |
|
| 129 | } catch (Exception $e) { |
||
| 130 | return false; |
||
| 131 | } |
||
| 132 | |||
| 133 | 16 | $beginning = $this->readBeginning(); |
|
| 134 | 16 | $startWithTag = self::startsWithTag($beginning); |
|
| 135 | 16 | $containsTags = self::containsTags($beginning); |
|
| 136 | 16 | $endsWithTag = self::endsWithTag($this->readEnding()); |
|
| 137 | |||
| 138 | 16 | fclose($this->fileHandle); |
|
|
1 ignored issue
–
show
|
|||
| 139 | |||
| 140 | 16 | return $startWithTag && $containsTags && $endsWithTag; |
|
| 141 | } |
||
| 142 | |||
| 143 | 16 | private function readBeginning() |
|
| 148 | } |
||
| 149 | |||
| 150 | 16 | private function readEnding() |
|
| 151 | { |
||
| 152 | 16 | $meta = stream_get_meta_data($this->fileHandle); |
|
| 153 | 16 | $filename = $meta['uri']; |
|
| 154 | |||
| 155 | 16 | $size = filesize($filename); |
|
| 156 | 16 | if ($size === 0) { |
|
| 157 | 1 | return ''; |
|
| 158 | } |
||
| 159 | |||
| 160 | 15 | $blockSize = self::TEST_SAMPLE_SIZE; |
|
| 161 | 15 | if ($size < $blockSize) { |
|
| 162 | 2 | $blockSize = $size; |
|
| 163 | } |
||
| 164 | |||
| 165 | 15 | fseek($this->fileHandle, $size - $blockSize); |
|
| 166 | |||
| 167 | 15 | return fread($this->fileHandle, $blockSize); |
|
| 168 | } |
||
| 169 | |||
| 170 | 16 | private static function startsWithTag($data) |
|
| 171 | { |
||
| 172 | 16 | return '<' === substr(trim($data), 0, 1); |
|
| 173 | } |
||
| 174 | |||
| 175 | 16 | private static function endsWithTag($data) |
|
| 176 | { |
||
| 177 | 16 | return '>' === substr(trim($data), -1, 1); |
|
| 178 | } |
||
| 179 | |||
| 180 | 16 | private static function containsTags($data) |
|
| 181 | { |
||
| 182 | 16 | return strlen($data) !== strlen(strip_tags($data)); |
|
| 183 | } |
||
| 184 | |||
| 185 | /** |
||
| 186 | * Loads Spreadsheet from file. |
||
| 187 | * |
||
| 188 | * @param string $pFilename |
||
| 189 | * |
||
| 190 | * @throws Exception |
||
| 191 | * |
||
| 192 | * @return Spreadsheet |
||
| 193 | */ |
||
| 194 | 9 | public function load($pFilename) |
|
| 195 | { |
||
| 196 | // Create new Spreadsheet |
||
| 197 | 9 | $spreadsheet = new Spreadsheet(); |
|
| 198 | |||
| 199 | // Load into this instance |
||
| 200 | 9 | return $this->loadIntoExisting($pFilename, $spreadsheet); |
|
| 201 | } |
||
| 202 | |||
| 203 | /** |
||
| 204 | * Set input encoding. |
||
| 205 | * |
||
| 206 | * @param string $pValue Input encoding, eg: 'ANSI' |
||
| 207 | * |
||
| 208 | * @return Html |
||
| 209 | */ |
||
| 210 | public function setInputEncoding($pValue) |
||
| 211 | { |
||
| 212 | $this->inputEncoding = $pValue; |
||
| 213 | |||
| 214 | return $this; |
||
| 215 | } |
||
| 216 | |||
| 217 | /** |
||
| 218 | * Get input encoding. |
||
| 219 | * |
||
| 220 | * @return string |
||
| 221 | */ |
||
| 222 | public function getInputEncoding() |
||
| 223 | { |
||
| 224 | return $this->inputEncoding; |
||
| 225 | } |
||
| 226 | |||
| 227 | // Data Array used for testing only, should write to Spreadsheet object on completion of tests |
||
| 228 | protected $dataArray = []; |
||
| 229 | |||
| 230 | protected $tableLevel = 0; |
||
| 231 | |||
| 232 | protected $nestedColumn = ['A']; |
||
| 233 | |||
| 234 | 9 | protected function setTableStartColumn($column) |
|
| 235 | { |
||
| 236 | 9 | if ($this->tableLevel == 0) { |
|
| 237 | 9 | $column = 'A'; |
|
| 238 | } |
||
| 239 | 9 | ++$this->tableLevel; |
|
| 240 | 9 | $this->nestedColumn[$this->tableLevel] = $column; |
|
| 241 | |||
| 242 | 9 | return $this->nestedColumn[$this->tableLevel]; |
|
| 243 | } |
||
| 244 | |||
| 245 | 9 | protected function getTableStartColumn() |
|
| 248 | } |
||
| 249 | |||
| 250 | 9 | protected function releaseTableStartColumn() |
|
| 255 | } |
||
| 256 | |||
| 257 | 9 | protected function flushCell(Worksheet $sheet, $column, $row, &$cellContent) |
|
| 258 | { |
||
| 259 | 9 | if (is_string($cellContent)) { |
|
| 260 | // Simple String content |
||
| 261 | 9 | if (trim($cellContent) > '') { |
|
| 262 | // Only actually write it if there's content in the string |
||
| 263 | // Write to worksheet to be done here... |
||
| 264 | // ... we return the cell so we can mess about with styles more easily |
||
| 265 | 9 | $sheet->setCellValue($column . $row, $cellContent); |
|
| 266 | 9 | $this->dataArray[$row][$column] = $cellContent; |
|
| 267 | } |
||
| 268 | } else { |
||
| 269 | // We have a Rich Text run |
||
| 270 | // TODO |
||
| 271 | $this->dataArray[$row][$column] = 'RICH TEXT: ' . $cellContent; |
||
| 272 | } |
||
| 273 | 9 | $cellContent = (string) ''; |
|
| 274 | 9 | } |
|
| 275 | |||
| 276 | /** |
||
| 277 | * @param DOMNode $element |
||
| 278 | * @param Worksheet $sheet |
||
| 279 | * @param int $row |
||
| 280 | * @param string $column |
||
| 281 | * @param string $cellContent |
||
| 282 | */ |
||
| 283 | 9 | protected function processDomElement(DOMNode $element, Worksheet $sheet, &$row, &$column, &$cellContent) |
|
| 522 | } |
||
| 523 | } |
||
| 524 | } |
||
| 525 | 9 | } |
|
| 526 | |||
| 527 | /** |
||
| 528 | * Loads PhpSpreadsheet from file into PhpSpreadsheet instance. |
||
| 529 | * |
||
| 530 | * @param string $pFilename |
||
| 531 | * @param Spreadsheet $spreadsheet |
||
| 532 | * |
||
| 533 | * @throws Exception |
||
| 534 | * |
||
| 535 | * @return Spreadsheet |
||
| 536 | */ |
||
| 537 | 9 | public function loadIntoExisting($pFilename, Spreadsheet $spreadsheet) |
|
| 538 | { |
||
| 539 | // Validate |
||
| 540 | 9 | if (!$this->canRead($pFilename)) { |
|
| 541 | throw new Exception($pFilename . ' is an Invalid HTML file.'); |
||
| 542 | } |
||
| 543 | |||
| 544 | // Create new sheet |
||
| 545 | 9 | while ($spreadsheet->getSheetCount() <= $this->sheetIndex) { |
|
| 546 | $spreadsheet->createSheet(); |
||
| 547 | } |
||
| 548 | 9 | $spreadsheet->setActiveSheetIndex($this->sheetIndex); |
|
| 549 | |||
| 550 | // Create a new DOM object |
||
| 551 | 9 | $dom = new DOMDocument(); |
|
| 552 | // Reload the HTML file into the DOM object |
||
| 553 | 9 | $loaded = $dom->loadHTML(mb_convert_encoding($this->securityScanner->scanFile($pFilename), 'HTML-ENTITIES', 'UTF-8')); |
|
| 554 | 9 | if ($loaded === false) { |
|
| 555 | throw new Exception('Failed to load ' . $pFilename . ' as a DOM Document'); |
||
| 556 | } |
||
| 557 | |||
| 558 | // Discard white space |
||
| 559 | 9 | $dom->preserveWhiteSpace = false; |
|
| 560 | |||
| 561 | 9 | $row = 0; |
|
| 562 | 9 | $column = 'A'; |
|
| 563 | 9 | $content = ''; |
|
| 564 | 9 | $this->rowspan = []; |
|
| 565 | 9 | $this->processDomElement($dom, $spreadsheet->getActiveSheet(), $row, $column, $content); |
|
| 566 | |||
| 567 | // Return |
||
| 568 | 9 | return $spreadsheet; |
|
| 569 | } |
||
| 570 | |||
| 571 | /** |
||
| 572 | * Get sheet index. |
||
| 573 | * |
||
| 574 | * @return int |
||
| 575 | */ |
||
| 576 | public function getSheetIndex() |
||
| 579 | } |
||
| 580 | |||
| 581 | /** |
||
| 582 | * Set sheet index. |
||
| 583 | * |
||
| 584 | * @param int $pValue Sheet index |
||
| 585 | * |
||
| 586 | * @return HTML |
||
| 587 | */ |
||
| 588 | public function setSheetIndex($pValue) |
||
| 593 | } |
||
| 594 | |||
| 595 | /** |
||
| 596 | * Apply inline css inline style. |
||
| 597 | * |
||
| 598 | * NOTES : |
||
| 599 | * Currently only intended for td & th element, |
||
| 600 | * and only takes 'background-color' and 'color'; property with HEX color |
||
| 601 | * |
||
| 602 | * TODO : |
||
| 603 | * - Implement to other propertie, such as border |
||
| 604 | * |
||
| 605 | * @param Worksheet $sheet |
||
| 606 | * @param int $row |
||
| 607 | * @param string $column |
||
| 608 | * @param array $attributeArray |
||
| 609 | */ |
||
| 610 | 9 | private function applyInlineStyle(&$sheet, $row, $column, $attributeArray) |
|
| 645 | } |
||
| 646 | } |
||
| 647 | 1 | } |
|
| 648 | } |
||
| 649 |
Loading history...