HuasoFoundries /
math_histogram
| Total Complexity | 70 |
| Total Lines | 356 |
| Duplicated Lines | 10.39 % |
| Changes | 0 | ||
Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like Histogram3D 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 Histogram3D, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 15 | class Histogram3D extends AbstractHistogram |
||
| 16 | { |
||
| 17 | |||
| 18 | /** |
||
| 19 | * Constructor for Math_Histogram3D |
||
| 20 | * |
||
| 21 | * @access public |
||
| 22 | * @param optional int $type one of HISTOGRAM_SIMPLE or HISTOGRAM_CUMMULATIVE |
||
| 23 | * @param optional array $binOptions an array of options for binning the data |
||
| 24 | * @return object Math_Histogram3D |
||
| 25 | * |
||
| 26 | * @see setBinOptions() |
||
| 27 | * @see Math_AbstractHistogram::setType() |
||
| 28 | * @see Math_AbstractHistogram |
||
| 29 | */ |
||
| 30 | |||
| 31 | public function __construct($type = self::HISTOGRAM_SIMPLE, $binOptions = "") |
||
| 37 | // Falling back to default options |
||
| 38 | } |
||
| 39 | } |
||
| 40 | |||
| 41 | /** |
||
| 42 | * Sets the binning options. Overrides parent's method. |
||
| 43 | * |
||
| 44 | * @access public |
||
| 45 | * @param array $binOptions an array of options for binning the data |
||
| 46 | * @return void |
||
| 47 | */ |
||
| 48 | |||
| 49 | public function setBinOptions($binOptions) |
||
| 50 | { |
||
| 51 | if ($this->_validBinOptions($binOptions)) { |
||
| 52 | return parent::setBinOptions($binOptions); |
||
|
Bug
Best Practice
introduced
by
|
|||
| 53 | } else { |
||
| 54 | throw new \PEAR_Exception("incorrect options array"); |
||
| 55 | } |
||
| 56 | } |
||
| 57 | |||
| 58 | /** |
||
| 59 | * Sets the data to be processed. The data will be validated to |
||
| 60 | * be a simple bi-dimensional numerical array |
||
| 61 | * |
||
| 62 | * @access public |
||
| 63 | * @param array $data the numeric array |
||
| 64 | * @return mixed boolean true on success, a \PEAR_Error object otherwise |
||
| 65 | * |
||
| 66 | * @see _clear() |
||
| 67 | * @see Math_AbstractHistogram::getData() |
||
| 68 | * @see Math_AbstractHistogram |
||
| 69 | * @see getHistogramData() |
||
| 70 | */ |
||
| 71 | public function setData($data) |
||
| 72 | { |
||
| 73 | $this->_clear(); |
||
| 74 | if (!$this->_validData($data)) { |
||
| 75 | throw new \PEAR_Exception("array of numeric coordinates expected"); |
||
| 76 | } |
||
| 77 | |||
| 78 | $this->_data = $data; |
||
| 79 | list($xMin, $xMax) = $this->_getMinMax('x'); |
||
| 80 | list($yMin, $yMax) = $this->_getMinMax('y'); |
||
| 81 | if (is_null($this->_rangeLow)) { |
||
| 82 | $this->_rangeLow = array('x' => $xMin, 'y' => $yMin); |
||
| 83 | } |
||
| 84 | |||
| 85 | if (is_null($this->_rangeHigh)) { |
||
| 86 | $this->_rangeHigh = array('x' => $xMax, 'y' => $yMax); |
||
| 87 | } |
||
| 88 | |||
| 89 | if (is_null($this->_nbins)) { |
||
| 90 | $this->_nbins = array('x' => 10, 'y' => 10); |
||
| 91 | } |
||
| 92 | |||
| 93 | return true; |
||
| 94 | } |
||
| 95 | |||
| 96 | /** |
||
| 97 | * Calculates the histogram bins and frequencies |
||
| 98 | * |
||
| 99 | * @access public |
||
| 100 | * @param optional $statsMode calculate basic statistics (STATS_BASIC) or full (STATS_FULL) |
||
| 101 | * @return mixed boolean true on success, a \PEAR_Error object otherwise |
||
| 102 | * |
||
| 103 | * @see Math_Stats |
||
| 104 | */ |
||
| 105 | public function calculate($statsMode = \HuasoFoundries\Math\Stats::STATS_BASIC) |
||
| 106 | { |
||
| 107 | $this->_bins = array(); |
||
| 108 | $this->_stats = array('x' => new \HuasoFoundries\Math\Stats(), 'y' => new \HuasoFoundries\Math\Stats()); |
||
| 109 | $this->_statsMode = $statsMode; |
||
| 110 | $deltaX = ($this->_rangeHigh['x'] - $this->_rangeLow['x']) / $this->_nbins['x']; |
||
| 111 | $deltaY = ($this->_rangeHigh['y'] - $this->_rangeLow['y']) / $this->_nbins['y']; |
||
| 112 | $data = $this->_histogramData(); |
||
| 113 | //$dataX = $this->_data['x']; |
||
| 114 | //$dataY = $this->_data['y']; |
||
| 115 | $dataX = $data['x']; |
||
| 116 | $dataY = $data['y']; |
||
| 117 | $ignoreList = array(); |
||
| 118 | $cumm = 0; |
||
| 119 | $nData = count($dataX); |
||
| 120 | for ($i = 0; $i < $this->_nbins['x']; $i++) { |
||
| 121 | $loXBin = $this->_rangeLow['x'] + $i * $deltaX; |
||
| 122 | $hiXBin = $loXBin + $deltaX; |
||
| 123 | $xBin = array('low' => $loXBin, 'high' => $hiXBin, |
||
| 124 | 'mid' => ($hiXBin + $loXBin) / 2); |
||
| 125 | for ($j = 0; $j < $this->_nbins['y']; $j++) { |
||
| 126 | $loYBin = $this->_rangeLow['y'] + $j * $deltaY; |
||
| 127 | $hiYBin = $loYBin + $deltaY; |
||
| 128 | $yBin = array('low' => $loYBin, 'high' => $hiYBin, |
||
| 129 | 'mid' => ($hiYBin + $loYBin) / 2); |
||
| 130 | $bin = array('x' => $xBin, 'y' => $yBin); |
||
| 131 | $freq = 0; |
||
| 132 | for ($k = 0; $k < $nData; $k++) { |
||
| 133 | if (!empty($ignoreList) && in_array($k, $ignoreList)) { |
||
| 134 | continue; |
||
| 135 | } |
||
| 136 | |||
| 137 | $valueX = $dataX[$k]; |
||
| 138 | $valueY = $dataY[$k]; |
||
| 139 | $inRangeX = $inRangeY = false; |
||
| 140 | View Code Duplication | if ($i == 0) { |
|
| 141 | $inRangeX = ($loXBin <= $valueX && $hiXBin >= $valueX); |
||
| 142 | } else { |
||
| 143 | $inRangeX = ($loXBin < $valueX && $hiXBin >= $valueX); |
||
| 144 | } |
||
| 145 | |||
| 146 | View Code Duplication | if ($j == 0) { |
|
| 147 | $inRangeY = ($loYBin <= $valueY && $hiYBin >= $valueY); |
||
| 148 | } else { |
||
| 149 | $inRangeY = ($loYBin < $valueY && $hiYBin >= $valueY); |
||
| 150 | } |
||
| 151 | |||
| 152 | if ($inRangeX && $inRangeY) { |
||
| 153 | $freq++; |
||
| 154 | $cumm++; |
||
| 155 | $ignoreList[] = $k; |
||
| 156 | } |
||
| 157 | } |
||
| 158 | View Code Duplication | if ($this->_type == self::HISTOGRAM_CUMMULATIVE) { |
|
| 159 | if ($freq > 0) { |
||
| 160 | $bin['count'] = $freq + $cumm - 1; |
||
| 161 | } else { |
||
| 162 | $bin['count'] = 0; |
||
| 163 | } |
||
| 164 | } else { |
||
| 165 | $bin['count'] = $freq; |
||
| 166 | } |
||
| 167 | $bin['xbin'] = $i; |
||
| 168 | $bin['ybin'] = $j; |
||
| 169 | $this->_bins[] = $bin; |
||
| 170 | } |
||
| 171 | } |
||
| 172 | } |
||
| 173 | |||
| 174 | /** |
||
| 175 | * Returns the statistics for the data set |
||
| 176 | * |
||
| 177 | * @access public |
||
| 178 | * @return mixed an associative array on success, a \PEAR_Error object otherwise |
||
| 179 | */ |
||
| 180 | public function getDataStats() |
||
| 181 | { |
||
| 182 | if (empty($this->_bins)) { |
||
| 183 | throw new \PEAR_Exception("histogram has not been calculated"); |
||
| 184 | } |
||
| 185 | |||
| 186 | $this->_stats['x']->setData($this->_data['x']); |
||
| 187 | $this->_stats['y']->setData($this->_data['y']); |
||
| 188 | return array('x' => $this->_stats['x']->calc($this->_statsMode), |
||
| 189 | 'y' => $this->_stats['y']->calc($this->_statsMode)); |
||
| 190 | } |
||
| 191 | |||
| 192 | /** |
||
| 193 | * Returns the statistics for the data set, filtered using the bin ranges |
||
| 194 | * |
||
| 195 | * @access public |
||
| 196 | * @return mixed an associative array on success, a \PEAR_Error object otherwise |
||
| 197 | */ |
||
| 198 | public function getHistogramDataStats() |
||
| 199 | { |
||
| 200 | if (empty($this->_bins)) { |
||
| 201 | throw new \PEAR_Exception("histogram has not been calculated"); |
||
| 202 | } |
||
| 203 | |||
| 204 | $data = $this->_histogramData(); |
||
| 205 | $this->_stats['x']->setData($data['x']); |
||
| 206 | $this->_stats['y']->setData($data['y']); |
||
| 207 | return array('x' => $this->_stats['x']->calc($this->_statsMode), |
||
| 208 | 'y' => $this->_stats['y']->calc($this->_statsMode)); |
||
| 209 | } |
||
| 210 | |||
| 211 | /** |
||
| 212 | * Returns the bins and frequencies calculated using the given |
||
| 213 | * bin mode and separator |
||
| 214 | * |
||
| 215 | * @access public |
||
| 216 | * @param int $mode one of HISTOGRAM_LO_BINS, HISTOGRAM_MID_BINS (default), or HISTOGRAM_HI_BINS |
||
| 217 | * @param string $separator the separator, default ", " |
||
| 218 | * @return mixed a string on success, a \PEAR_Error object otherwise |
||
| 219 | */ |
||
| 220 | View Code Duplication | public function toSeparated($mode = self::HISTOGRAM_MID_BINS, $separator = ", ") |
|
| 221 | { |
||
| 222 | try { |
||
| 223 | $bins = $this->getBins($mode); |
||
| 224 | } catch (\PEAR_Exception $e) { |
||
| 225 | return $bins; |
||
| 226 | } |
||
| 227 | |||
| 228 | |||
| 229 | $nbins = count($bins); |
||
| 230 | $out = array("# x_bin{$separator}y_bin{$separator}frequency"); |
||
| 231 | for ($i = 0; $i < $nbins; $i++) { |
||
| 232 | $out[] = implode($separator, $bins[$i]); |
||
| 233 | } |
||
| 234 | |||
| 235 | return implode("\n", $out) . "\n"; |
||
| 236 | } |
||
| 237 | |||
| 238 | /** |
||
| 239 | * Returns the minimum and maximum of the given unidimensional numeric |
||
| 240 | * array |
||
| 241 | * |
||
| 242 | * @access private |
||
| 243 | * @param array $elem |
||
| 244 | * @return array of values: array(min, max) |
||
| 245 | */ |
||
| 246 | public function _getMinMax($elem) |
||
| 247 | { |
||
| 248 | return array(min($this->_data[$elem]), max($this->_data[$elem])); |
||
| 249 | } |
||
| 250 | |||
| 251 | /** |
||
| 252 | * Returns a subset of the bins array by bin value type |
||
| 253 | * |
||
| 254 | * @access private |
||
| 255 | * @param int $mode one of HISTOGRAM_MID_BINS, HISTOGRAM_LO_BINS or HISTOGRAM_HI_BINS |
||
| 256 | * @return array |
||
| 257 | */ |
||
| 258 | public function _filterBins($mode) |
||
| 259 | { |
||
| 260 | $map = array( |
||
| 261 | self::HISTOGRAM_MID_BINS => "mid", |
||
| 262 | self::HISTOGRAM_LO_BINS => "low", |
||
| 263 | self::HISTOGRAM_HI_BINS => "high", |
||
| 264 | ); |
||
| 265 | $filtered = array(); |
||
| 266 | foreach ($this->_bins as $bin) { |
||
| 267 | $tmp['x'] = $bin['x'][$map[$mode]]; |
||
| 268 | $tmp['y'] = $bin['y'][$map[$mode]]; |
||
| 269 | $tmp['count'] = $bin['count']; |
||
| 270 | $filtered[] = $tmp; |
||
| 271 | } |
||
| 272 | return $filtered; |
||
| 273 | } |
||
| 274 | |||
| 275 | /** |
||
| 276 | * Checks that the array of options passed is valid |
||
| 277 | * Options array should have the form: |
||
| 278 | * |
||
| 279 | * $opt = array ('low'=>array('x'=>10, 'y'=>10), |
||
| 280 | * 'high'=>array(...), |
||
| 281 | * 'nbins'=>array(...)); |
||
| 282 | * |
||
| 283 | * @access private |
||
| 284 | * @return boolean |
||
| 285 | */ |
||
| 286 | public function _validBinOptions($binOptions) |
||
| 287 | { |
||
| 288 | $barray = (is_array($binOptions) |
||
| 289 | && is_array($binOptions['low']) |
||
| 290 | && is_array($binOptions['high']) |
||
| 291 | && is_array($binOptions['nbins'])); |
||
| 292 | if (!$barray) { |
||
| 293 | return false; |
||
| 294 | } |
||
| 295 | $low = $binOptions['low']; |
||
| 296 | $high = $binOptions['high']; |
||
| 297 | $nbins = $binOptions['nbins']; |
||
| 298 | $blow = (isset($low['x']) && isset($low['y']) |
||
| 299 | && is_numeric($low['x']) && is_numeric($low['y'])); |
||
| 300 | $bhigh = (isset($high['x']) && isset($high['y']) |
||
| 301 | && is_numeric($high['x']) && is_numeric($high['y'])); |
||
| 302 | $bnbins = (isset($nbins['x']) && isset($nbins['y']) |
||
| 303 | && is_numeric($nbins['x']) && is_numeric($nbins['y'])); |
||
| 304 | return ($blow && $bhigh && $bnbins); |
||
| 305 | } |
||
| 306 | |||
| 307 | /** |
||
| 308 | * Checks that the data passed is bi-dimensional numeric array |
||
| 309 | * of the form: |
||
| 310 | * |
||
| 311 | * $data = array ('x'=>array(...), 'y'=>array(...)); |
||
| 312 | * |
||
| 313 | * It also checks that: count($data['x']) == count($data['y']) |
||
| 314 | * |
||
| 315 | * @access private |
||
| 316 | * @return boolean |
||
| 317 | */ |
||
| 318 | public function _validData($data) |
||
| 319 | { |
||
| 320 | if (is_array($data) && is_array($data['x']) && is_array($data['y'])) { |
||
| 321 | $n = count($data['x']); |
||
| 322 | if (count($data) == 2 && $n == count($data['y'])) { |
||
| 323 | for ($i = 0; $i < $n; $i++) { |
||
| 324 | if (!is_numeric($data['x'][$i]) || !is_numeric($data['y'][$i])) { |
||
| 325 | return false; |
||
| 326 | } |
||
| 327 | } |
||
| 328 | |||
| 329 | // if everything checks out |
||
| 330 | return true; |
||
| 331 | } else { |
||
| 332 | return false; |
||
| 333 | } |
||
| 334 | } else { |
||
| 335 | return false; |
||
| 336 | } |
||
| 337 | } |
||
| 338 | |||
| 339 | /** |
||
| 340 | * Returns an array of data contained within the ranges for the |
||
| 341 | * histogram calculation. Overrides the empty implementation in |
||
| 342 | * Math_AbstractHistogram::_histogramData() |
||
| 343 | * |
||
| 344 | * @access private |
||
| 345 | * @return array |
||
| 346 | */ |
||
| 347 | public function _histogramData() |
||
| 371 | } |
||
| 372 | } |
||
| 373 | |||
| 374 | // vim: ts=4:sw=4:et: |
||
| 375 | // vim6: fdl=1: |
||
| 376 |
Loading history...
