mirror of
https://github.com/sbrl/Pepperminty-Wiki.git
synced 2024-11-26 05:32:59 +00:00
Minor code formatting
This commit is contained in:
parent
5609506def
commit
7088990027
1 changed files with 7 additions and 12 deletions
|
@ -545,8 +545,7 @@ class search
|
||||||
$index = [];
|
$index = [];
|
||||||
|
|
||||||
$terms = self::tokenize($source, true);
|
$terms = self::tokenize($source, true);
|
||||||
foreach($terms as $term)
|
foreach($terms as $term) {
|
||||||
{
|
|
||||||
// Skip over stop words (see https://en.wikipedia.org/wiki/Stop_words)
|
// Skip over stop words (see https://en.wikipedia.org/wiki/Stop_words)
|
||||||
if(in_array($term[0], self::$stop_words)) continue;
|
if(in_array($term[0], self::$stop_words)) continue;
|
||||||
|
|
||||||
|
@ -566,8 +565,7 @@ class search
|
||||||
* @param bool $capture_offsets Whether to capture & return the character offsets of the tokens detected. If true, then each token returned will be an array in the form [ token, char_offset ].
|
* @param bool $capture_offsets Whether to capture & return the character offsets of the tokens detected. If true, then each token returned will be an array in the form [ token, char_offset ].
|
||||||
* @return array An array of raw tokens extracted from the specified source string.
|
* @return array An array of raw tokens extracted from the specified source string.
|
||||||
*/
|
*/
|
||||||
public static function tokenize($source, $capture_offsets = false)
|
public static function tokenize($source, $capture_offsets = false) {
|
||||||
{
|
|
||||||
/** Normalises input characters for searching & indexing */
|
/** Normalises input characters for searching & indexing */
|
||||||
static $literator; if($literator == null) $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD);
|
static $literator; if($literator == null) $literator = Transliterator::createFromRules(':: Any-Latin; :: Latin-ASCII; :: NFD; :: [:Nonspacing Mark:] Remove; :: Lower(); :: NFC;', Transliterator::FORWARD);
|
||||||
|
|
||||||
|
@ -664,15 +662,12 @@ class search
|
||||||
* @param array $changed An array to be filled with the nterms of all the changed entries.
|
* @param array $changed An array to be filled with the nterms of all the changed entries.
|
||||||
* @param array $removed An array to be filled with the nterms of all the removed entries.
|
* @param array $removed An array to be filled with the nterms of all the removed entries.
|
||||||
*/
|
*/
|
||||||
public static function compare_indexes($oldindex, $newindex, &$changed, &$removed)
|
public static function compare_indexes($oldindex, $newindex, &$changed, &$removed) {
|
||||||
{
|
foreach($oldindex as $nterm => $entry) {
|
||||||
foreach($oldindex as $nterm => $entry)
|
|
||||||
{
|
|
||||||
if(!isset($newindex[$nterm]))
|
if(!isset($newindex[$nterm]))
|
||||||
$removed[] = $nterm;
|
$removed[] = $nterm;
|
||||||
}
|
}
|
||||||
foreach($newindex as $nterm => $entry)
|
foreach($newindex as $nterm => $entry) {
|
||||||
{
|
|
||||||
if(!isset($oldindex[$nterm]) or // If this word is new
|
if(!isset($oldindex[$nterm]) or // If this word is new
|
||||||
$newindex[$nterm] !== $oldindex[$nterm]) // If this word has changed
|
$newindex[$nterm] !== $oldindex[$nterm]) // If this word has changed
|
||||||
$changed[$nterm] = $newindex[$nterm];
|
$changed[$nterm] = $newindex[$nterm];
|
||||||
|
@ -691,7 +686,7 @@ class search
|
||||||
/**
|
/**
|
||||||
* Reads in and parses an inverted index, measuring the time it takes to do so.
|
* Reads in and parses an inverted index, measuring the time it takes to do so.
|
||||||
* @param string $invindex_filename The path to the file inverted index to parse.
|
* @param string $invindex_filename The path to the file inverted index to parse.
|
||||||
* @return boolean Whether the measurement was actually able to take place. Usually this will be true, but it will return false if it can't find the specified index.
|
* @return bool Whether the measurement was actually able to take place. Usually this will be true, but it will return false if it can't find the specified index.
|
||||||
*/
|
*/
|
||||||
public static function measure_invindex_load_time($invindex_filename) {
|
public static function measure_invindex_load_time($invindex_filename) {
|
||||||
global $env;
|
global $env;
|
||||||
|
|
Loading…
Reference in a new issue