+global\IcuCollation::fetchFirstLetterData()

Complexity

Conditions	13
Paths	97

Size

Total Lines	125
Code Lines	43

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	0

<?php
	private function fetchFirstLetterData() {
		// Generate data from serialized data file
		if ( isset( self::$tailoringFirstLetters[$this->locale] ) ) {
			$letters = wfGetPrecompiledData( 'first-letters-root.ser' );
			// Append additional characters
			$letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
			// Remove unnecessary ones, if any
			if ( isset( self::$tailoringFirstLetters['-' . $this->locale] ) ) {
				$letters = array_diff( $letters, self::$tailoringFirstLetters['-' . $this->locale] );
			}
			// Apply digit transforms
			$digits = [ '0', '1', '2', '3', '4', '5', '6', '7', '8', '9' ];
			$letters = array_diff( $letters, $digits );
			foreach ( $digits as $digit ) {
				$letters[] = $this->digitTransformLanguage->formatNum( $digit, true );
			}
		} else {
			$letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
			if ( $letters === false ) {
				throw new MWException( "MediaWiki does not support ICU locale " .
					"\"{$this->locale}\"" );
			}
		}

		/* Sort the letters.
		 *
		 * It's impossible to have the precompiled data file properly sorted,
		 * because the sort order changes depending on ICU version. If the
		 * array is not properly sorted, the binary search will return random
		 * results.
		 *
		 * We also take this opportunity to remove primary collisions.
		 */
		$letterMap = [];
		foreach ( $letters as $letter ) {
			$key = $this->getPrimarySortKey( $letter );
			if ( isset( $letterMap[$key] ) ) {
				// Primary collision
				// Keep whichever one sorts first in the main collator
				if ( $this->mainCollator->compare( $letter, $letterMap[$key] ) < 0 ) {
					$letterMap[$key] = $letter;
				}
			} else {
				$letterMap[$key] = $letter;
			}
		}
		ksort( $letterMap, SORT_STRING );

		/* Remove duplicate prefixes. Basically if something has a sortkey
		 * which is a prefix of some other sortkey, then it is an
		 * expansion and probably should not be considered a section
		 * header.
		 *
		 * For example 'þ' is sometimes sorted as if it is the letters
		 * 'th'. Other times it is its own primary element. Another
		 * example is '₨'. Sometimes its a currency symbol. Sometimes it
		 * is an 'R' followed by an 's'.
		 *
		 * Additionally an expanded element should always sort directly
		 * after its first element due to they way sortkeys work.
		 *
		 * UCA sortkey elements are of variable length but no collation
		 * element should be a prefix of some other element, so I think
		 * this is safe. See:
		 * - https://ssl.icu-project.org/repos/icu/icuhtml/trunk/design/collation/ICU_collation_design.htm
		 * - http://site.icu-project.org/design/collation/uca-weight-allocation
		 *
		 * Additionally, there is something called primary compression to
		 * worry about. Basically, if you have two primary elements that
		 * are more than one byte and both start with the same byte then
		 * the first byte is dropped on the second primary. Additionally
		 * either \x03 or \xFF may be added to mean that the next primary
		 * does not start with the first byte of the first primary.
		 *
		 * This shouldn't matter much, as the first primary is not
		 * changed, and that is what we are comparing against.
		 *
		 * tl;dr: This makes some assumptions about how icu implements
		 * collations. It seems incredibly unlikely these assumptions
		 * will change, but nonetheless they are assumptions.
		 */

		$prev = false;
		$duplicatePrefixes = [];
		foreach ( $letterMap as $key => $value ) {
			// Remove terminator byte. Otherwise the prefix
			// comparison will get hung up on that.
			$trimmedKey = rtrim( $key, "\0" );
			if ( $prev === false || $prev === '' ) {
				$prev = $trimmedKey;
				// We don't yet have a collation element
				// to compare against, so continue.
				continue;
			}

			// Due to the fact the array is sorted, we only have
			// to compare with the element directly previous
			// to the current element (skipping expansions).
			// An element "X" will always sort directly
			// before "XZ" (Unless we have "XY", but we
			// do not update $prev in that case).
			if ( substr( $trimmedKey, 0, strlen( $prev ) ) === $prev ) {
				$duplicatePrefixes[] = $key;
				// If this is an expansion, we don't want to
				// compare the next element to this element,
				// but to what is currently $prev
				continue;
			}
			$prev = $trimmedKey;
		}
		foreach ( $duplicatePrefixes as $badKey ) {
			wfDebug( "Removing '{$letterMap[$badKey]}' from first letters.\n" );
			unset( $letterMap[$badKey] );
			// This code assumes that unsetting does not change sort order.
		}
		$data = [
			'chars' => array_values( $letterMap ),
			'keys' => array_keys( $letterMap ),
		];

		// Reduce memory usage before caching
		unset( $letterMap );

		return $data;
	}