sweetrdf /
in-memory-store-sqlite
| Total Complexity | 55 |
| Total Lines | 306 |
| Duplicated Lines | 0 % |
| Coverage | 90.5% |
| Changes | 1 | ||
| Bugs | 0 | Features | 0 |
Complex classes like LoadQueryHandler 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 LoadQueryHandler, and based on these observations, apply Extract Interface, too.
| 1 | <?php |
||
| 20 | class LoadQueryHandler extends QueryHandler |
||
| 21 | { |
||
| 22 | private string $target_graph; |
||
| 23 | |||
| 24 | /** |
||
| 25 | * @todo required? |
||
| 26 | */ |
||
| 27 | private int $t_count; |
||
| 28 | |||
| 29 | private int $write_buffer_size = 2500; |
||
| 30 | |||
| 31 | 9 | public function runQuery($infos, $data = '', $keep_bnode_ids = 0) |
|
| 32 | { |
||
| 33 | 9 | $url = $infos['query']['url']; |
|
| 34 | 9 | $graph = $infos['query']['target_graph']; |
|
| 35 | 9 | $this->target_graph = $graph ? calcURI($graph) : calcURI($url); |
|
| 36 | 9 | $this->keep_bnode_ids = $keep_bnode_ids; |
|
|
Bug
Best Practice
introduced
by
|
|||
| 37 | |||
| 38 | // remove parameters |
||
| 39 | 9 | $parserLogger = $this->store->getLoggerPool()->createNewLogger('Turtle'); |
|
| 40 | 9 | $loader = new TurtleLoader($parserLogger); |
|
| 41 | 9 | $loader->setCaller($this); |
|
| 42 | |||
| 43 | /* logging */ |
||
| 44 | 9 | $this->t_count = 0; |
|
| 45 | 9 | $this->t_start = 0; |
|
| 46 | /* load and parse */ |
||
| 47 | 9 | $this->max_term_id = $this->getMaxTermID(); |
|
| 48 | 9 | $this->max_triple_id = $this->getMaxTripleID(); |
|
| 49 | |||
| 50 | 9 | $this->term_ids = []; |
|
| 51 | 9 | $this->triple_ids = []; |
|
| 52 | 9 | $this->sql_buffers = []; |
|
| 53 | 9 | $loader->parse($url, $data); |
|
| 54 | |||
| 55 | /* done */ |
||
| 56 | 9 | $this->checkSQLBuffers(1); |
|
| 57 | |||
| 58 | return [ |
||
| 59 | 9 | 't_count' => $this->t_count, |
|
| 60 | 9 | 'load_time' => 0, |
|
| 61 | ]; |
||
| 62 | } |
||
| 63 | |||
| 64 | 9 | public function addT($s, $p, $o, $s_type, $o_type, $o_dt = '', $o_lang = '') |
|
| 65 | { |
||
| 66 | 9 | $type_ids = ['uri' => '0', 'bnode' => '1', 'literal' => '2']; |
|
| 67 | 9 | $g = $this->getStoredTermID($this->target_graph, '0', 'id'); |
|
| 68 | 9 | $s = (('bnode' == $s_type) && !$this->keep_bnode_ids) ? '_:b'.abs(crc32($g.$s)).'_'.(\strlen($s) > 12 ? substr(substr($s, 2), -10) : substr($s, 2)) : $s; |
|
| 69 | 9 | $o = (('bnode' == $o_type) && !$this->keep_bnode_ids) ? '_:b'.abs(crc32($g.$o)).'_'.(\strlen($o) > 12 ? substr(substr($o, 2), -10) : substr($o, 2)) : $o; |
|
| 70 | /* triple */ |
||
| 71 | 9 | $t = [ |
|
| 72 | 9 | 's' => $this->getStoredTermID($s, $type_ids[$s_type], 's'), |
|
| 73 | 9 | 'p' => $this->getStoredTermID($p, '0', 'id'), |
|
| 74 | 9 | 'o' => $this->getStoredTermID($o, $type_ids[$o_type], 'o'), |
|
| 75 | 9 | 'o_lang_dt' => $this->getStoredTermID($o_dt.$o_lang, $o_dt ? '0' : '2', 'id'), |
|
| 76 | 9 | 'o_comp' => getNormalizedValue($o), |
|
| 77 | 9 | 's_type' => $type_ids[$s_type], |
|
| 78 | 9 | 'o_type' => $type_ids[$o_type], |
|
| 79 | ]; |
||
| 80 | 9 | $t['t'] = $this->getTripleID($t); |
|
| 81 | 9 | if (\is_array($t['t'])) {/* t exists already */ |
|
| 82 | $t['t'] = $t['t'][0]; |
||
| 83 | } else { |
||
| 84 | 9 | $this->bufferTripleSQL($t); |
|
| 85 | } |
||
| 86 | /* g2t */ |
||
| 87 | 9 | $g2t = ['g' => $g, 't' => $t['t']]; |
|
| 88 | 9 | $this->bufferGraphSQL($g2t); |
|
| 89 | 9 | ++$this->t_count; |
|
| 90 | /* check buffers */ |
||
| 91 | 9 | if (0 == ($this->t_count % $this->write_buffer_size)) { |
|
| 92 | $force_write = 1; |
||
| 93 | $reset_buffers = (0 == ($this->t_count % ($this->write_buffer_size * 2))); |
||
| 94 | $refresh_lock = (0 == ($this->t_count % 25000)); |
||
| 95 | $split_tables = (0 == ($this->t_count % ($this->write_buffer_size * 10))); |
||
| 96 | $this->checkSQLBuffers($force_write, $reset_buffers, $refresh_lock, $split_tables); |
||
| 97 | } |
||
| 98 | 9 | } |
|
| 99 | |||
| 100 | 9 | public function getMaxTermID(): int |
|
| 101 | { |
||
| 102 | 9 | $sql = ''; |
|
| 103 | 9 | foreach (['id2val', 's2val', 'o2val'] as $tbl) { |
|
| 104 | 9 | $sql .= $sql ? ' UNION ' : ''; |
|
| 105 | 9 | $sql .= 'SELECT MAX(id) as id FROM '.$tbl; |
|
| 106 | } |
||
| 107 | 9 | $r = 0; |
|
| 108 | |||
| 109 | 9 | $rows = $this->store->getDBObject()->fetchList($sql); |
|
| 110 | |||
| 111 | 9 | if (\is_array($rows)) { |
|
| 112 | 9 | foreach ($rows as $row) { |
|
| 113 | 9 | $r = ($r < $row['id']) ? $row['id'] : $r; |
|
| 114 | } |
||
| 115 | } |
||
| 116 | |||
| 117 | 9 | return $r + 1; |
|
| 118 | } |
||
| 119 | |||
| 120 | /** |
||
| 121 | * @todo change DB schema and avoid using this function because it does not protect against race conditions |
||
| 122 | * |
||
| 123 | * @return int |
||
| 124 | */ |
||
| 125 | 9 | public function getMaxTripleID() |
|
| 126 | { |
||
| 127 | 9 | $sql = 'SELECT MAX(t) AS `id` FROM triple'; |
|
| 128 | |||
| 129 | 9 | $row = $this->store->getDBObject()->fetchRow($sql); |
|
| 130 | 9 | if (isset($row['id'])) { |
|
| 131 | 5 | return $row['id'] + 1; |
|
| 132 | } |
||
| 133 | |||
| 134 | 9 | return 1; |
|
| 135 | } |
||
| 136 | |||
| 137 | 9 | public function getStoredTermID($val, $type_id, $tbl) |
|
| 138 | { |
||
| 139 | /* buffered */ |
||
| 140 | 9 | if (isset($this->term_ids[$val])) { |
|
| 141 | 9 | if (!isset($this->term_ids[$val][$tbl])) { |
|
| 142 | 8 | foreach (['id', 's', 'o'] as $other_tbl) { |
|
| 143 | 8 | if (isset($this->term_ids[$val][$other_tbl])) { |
|
| 144 | 8 | $this->term_ids[$val][$tbl] = $this->term_ids[$val][$other_tbl]; |
|
| 145 | 8 | $this->bufferIDSQL($tbl, $this->term_ids[$val][$tbl], $val, $type_id); |
|
| 146 | 8 | break; |
|
| 147 | } |
||
| 148 | } |
||
| 149 | } |
||
| 150 | |||
| 151 | 9 | return $this->term_ids[$val][$tbl]; |
|
| 152 | } |
||
| 153 | /* db */ |
||
| 154 | 9 | $sub_tbls = ('id' == $tbl) |
|
| 155 | 9 | ? ['id2val', 's2val', 'o2val'] |
|
| 156 | 9 | : ('s' == $tbl |
|
| 157 | 9 | ? ['s2val', 'id2val', 'o2val'] |
|
| 158 | 9 | : ['o2val', 'id2val', 's2val'] |
|
| 159 | ); |
||
| 160 | |||
| 161 | 9 | foreach ($sub_tbls as $sub_tbl) { |
|
| 162 | 9 | $id = 0; |
|
| 163 | /* via hash */ |
||
| 164 | 9 | if (preg_match('/^(s2val|o2val)$/', $sub_tbl)) { |
|
| 165 | 9 | $sql = 'SELECT id, val |
|
| 166 | 9 | FROM '.$sub_tbl.' |
|
| 167 | 9 | WHERE val_hash = "'.$this->getValueHash($val).'"'; |
|
| 168 | |||
| 169 | 9 | $rows = $this->store->getDBObject()->fetchList($sql); |
|
| 170 | 9 | if (\is_array($rows)) { |
|
| 171 | 9 | foreach ($rows as $row) { |
|
| 172 | if ($row['val'] == $val) { |
||
| 173 | $id = $row['id']; |
||
| 174 | break; |
||
| 175 | } |
||
| 176 | } |
||
| 177 | } |
||
| 178 | } else { |
||
| 179 | 9 | $binaryValue = $this->store->getDBObject()->escape($val); |
|
| 180 | 9 | if (false !== empty($binaryValue)) { |
|
| 181 | 9 | $sql = 'SELECT id FROM '.$sub_tbl." WHERE val = '".$binaryValue."'"; |
|
| 182 | |||
| 183 | 9 | $row = $this->store->getDBObject()->fetchRow($sql); |
|
| 184 | 9 | if (\is_array($row) && isset($row['id'])) { |
|
| 185 | 5 | $id = $row['id']; |
|
| 186 | } |
||
| 187 | } |
||
| 188 | } |
||
| 189 | 9 | if (0 < $id) { |
|
| 190 | 5 | $this->term_ids[$val] = [$tbl => $id]; |
|
| 191 | 5 | if ($sub_tbl != $tbl.'2val') { |
|
| 192 | $this->bufferIDSQL($tbl, $id, $val, $type_id); |
||
| 193 | } |
||
| 194 | 5 | break; |
|
| 195 | } |
||
| 196 | } |
||
| 197 | /* new */ |
||
| 198 | 9 | if (!isset($this->term_ids[$val])) { |
|
| 199 | 9 | $this->term_ids[$val] = [$tbl => $this->max_term_id]; |
|
| 200 | 9 | $this->bufferIDSQL($tbl, $this->max_term_id, $val, $type_id); |
|
| 201 | 9 | ++$this->max_term_id; |
|
| 202 | } |
||
| 203 | |||
| 204 | 9 | return $this->term_ids[$val][$tbl]; |
|
| 205 | } |
||
| 206 | |||
| 207 | 9 | public function getTripleID($t) |
|
| 208 | { |
||
| 209 | 9 | $val = serialize($t); |
|
| 210 | /* buffered */ |
||
| 211 | 9 | if (isset($this->triple_ids[$val])) { |
|
| 212 | /* hack for "don't insert this triple" */ |
||
| 213 | return [$this->triple_ids[$val]]; |
||
| 214 | } |
||
| 215 | /* db */ |
||
| 216 | 9 | $sql = 'SELECT t |
|
| 217 | FROM triple |
||
| 218 | 9 | WHERE s = '.$t['s'].' |
|
| 219 | 9 | AND p = '.$t['p'].' |
|
| 220 | 9 | AND o = '.$t['o'].' |
|
| 221 | 9 | AND o_lang_dt = '.$t['o_lang_dt'].' |
|
| 222 | 9 | AND s_type = '.$t['s_type'].' |
|
| 223 | 9 | AND o_type = '.$t['o_type'].' |
|
| 224 | LIMIT 1'; |
||
| 225 | 9 | $row = $this->store->getDBObject()->fetchRow($sql); |
|
| 226 | 9 | if (isset($row['t'])) { |
|
| 227 | /* hack for "don't insert this triple" */ |
||
| 228 | $this->triple_ids[$val] = $row['t']; |
||
| 229 | |||
| 230 | return [$row['t']]; |
||
| 231 | } else { |
||
| 232 | /* new */ |
||
| 233 | 9 | $this->triple_ids[$val] = $this->max_triple_id; |
|
| 234 | 9 | ++$this->max_triple_id; |
|
| 235 | |||
| 236 | 9 | return $this->triple_ids[$val]; |
|
| 237 | } |
||
| 238 | } |
||
| 239 | |||
| 240 | 9 | public function bufferTripleSQL($t) |
|
| 241 | { |
||
| 242 | 9 | $tbl = 'triple'; |
|
| 243 | 9 | $sql = ', '; |
|
| 244 | |||
| 245 | 9 | $sqlHead = 'INSERT OR IGNORE INTO '; |
|
| 246 | |||
| 247 | 9 | if (!isset($this->sql_buffers[$tbl])) { |
|
| 248 | 9 | $this->sql_buffers[$tbl] = $sqlHead; |
|
| 249 | 9 | $this->sql_buffers[$tbl] .= $tbl; |
|
| 250 | 9 | $this->sql_buffers[$tbl] .= ' (t, s, p, o, o_lang_dt, o_comp, s_type, o_type) VALUES'; |
|
| 251 | 9 | $sql = ' '; |
|
| 252 | } |
||
| 253 | |||
| 254 | 9 | $oCompEscaped = $this->store->getDBObject()->escape($t['o_comp']); |
|
| 255 | |||
| 256 | 9 | $this->sql_buffers[$tbl] .= $sql.'('.$t['t'].', '.$t['s'].', '.$t['p'].', '; |
|
| 257 | 9 | $this->sql_buffers[$tbl] .= $t['o'].', '.$t['o_lang_dt'].", '"; |
|
| 258 | 9 | $this->sql_buffers[$tbl] .= $oCompEscaped."', ".$t['s_type'].', '.$t['o_type'].')'; |
|
| 259 | 9 | } |
|
| 260 | |||
| 261 | 9 | public function bufferGraphSQL($g2t) |
|
| 262 | { |
||
| 263 | 9 | $tbl = 'g2t'; |
|
| 264 | 9 | $sql = ', '; |
|
| 265 | |||
| 266 | /* |
||
| 267 | * Use appropriate INSERT syntax, depending on the DBS. |
||
| 268 | */ |
||
| 269 | 9 | $sqlHead = 'INSERT OR IGNORE INTO '; |
|
| 270 | |||
| 271 | 9 | if (!isset($this->sql_buffers[$tbl])) { |
|
| 272 | 9 | $this->sql_buffers[$tbl] = $sqlHead.$tbl.' (g, t) VALUES'; |
|
| 273 | 9 | $sql = ' '; |
|
| 274 | } |
||
| 275 | 9 | $this->sql_buffers[$tbl] .= $sql.'('.$g2t['g'].', '.$g2t['t'].')'; |
|
| 276 | 9 | } |
|
| 277 | |||
| 278 | 9 | public function bufferIDSQL($tbl, $id, $val, $val_type) |
|
| 305 | 9 | } |
|
| 306 | |||
| 307 | 9 | public function checkSQLBuffers($force_write = 0, $reset_id_buffers = 0) |
|
| 308 | { |
||
| 309 | 9 | foreach (['triple', 'g2t', 'id2val', 's2val', 'o2val'] as $tbl) { |
|
| 310 | 9 | $buffer_size = isset($this->sql_buffers[$tbl]) ? 1 : 0; |
|
| 311 | 9 | if ($buffer_size && $force_write) { |
|
| 312 | 9 | $this->store->getDBObject()->exec($this->sql_buffers[$tbl]); |
|
| 313 | /* table error */ |
||
| 328 |
Loading history...