Current File : /www/varak.net/wiki.varak.net/extensions/VisualEditor/lib/ve/src/dm/ve.dm.IndexValueStore.js
/*!
* VisualEditor IndexValueStore class.
*
* @copyright 2011-2016 VisualEditor Team and others; see http://ve.mit-license.org
*/
/* global SparkMD5 */
/**
* Ordered append-only hash store, whose values once inserted are immutable
*
* Values are objects, strings or Arrays, and are hashed using an algorithm with low collision
* probability: values with the same hash can be assumed equal.
*
* Values are stored in insertion order, and the store can be sliced to get a subset of values
* inserted consecutively.
*
* Two stores can be merged even if they have differently computed hashes, so long as two values
* will (with high probability) have the same hash only if equal. In this case, equivalent
* values can have two different hashes.
*
* TODO: rename the class to reflect that it is no longer an index-value store
*
* @class
* @constructor
*/
ve.dm.IndexValueStore = function VeDmIndexValueStore() {
// Maps hashes to values
this.hashStore = {};
// Hashes in order of insertion (used for slicing)
this.hashes = [];
};
/* Methods */
/**
* Get the number of values in the store
*
* @return {number} Number of values in the store
*/
ve.dm.IndexValueStore.prototype.getLength = function () {
return this.hashes.length;
};
/**
* Return a new store containing a slice of the values in insertion order
*
* @param {number} [start] Include values from position start onwards (default: 0)
* @param {number} [end] Include values to position end exclusive (default: slice to end)
* @return {ve.dm.IndexValueStore} Slice of the current store (with non-cloned value references)
*/
ve.dm.IndexValueStore.prototype.slice = function ( start, end ) {
var i, len, hash,
sliced = new this.constructor();
sliced.hashes = this.hashes.slice( start, end );
for ( i = 0, len = sliced.hashes.length; i < len; i++ ) {
hash = sliced.hashes[ i ];
sliced.hashStore[ hash ] = this.hashStore[ hash ];
}
return sliced;
};
/**
* Clone a store.
*
* @deprecated Use #slice with no arguments.
* @return {ve.dm.IndexValueStore} New store with the same contents as this one
*/
ve.dm.IndexValueStore.prototype.clone = function () {
return this.slice();
};
/**
* Insert a value into the store
*
* @method
* @param {Object|string|Array} value Value to store
* @param {string} [stringified] Stringified version of value; default OO.getHash( value )
* @return {string} Hash value with low collision probability
*/
ve.dm.IndexValueStore.prototype.index = function ( value, stringified ) {
var hash = this.indexOfValue( value, stringified );
if ( !this.hashStore[ hash ] ) {
if ( Array.isArray( value ) ) {
this.hashStore[ hash ] = ve.copy( value );
} else if ( typeof value === 'object' ) {
this.hashStore[ hash ] = ve.cloneObject( value );
} else {
this.hashStore[ hash ] = value;
}
this.hashes.push( hash );
}
return hash;
};
/**
* Get the hash of a value without inserting it in the store
*
* @method
* @param {Object|string|Array} value Value to hash
* @param {string} [stringified] Stringified version of value; default OO.getHash( value )
* @return {string} Hash value with low collision probability
*/
ve.dm.IndexValueStore.prototype.indexOfValue = function ( value, stringified ) {
if ( typeof stringified !== 'string' ) {
stringified = OO.getHash( value );
}
// We don't need cryptographically strong hashes, just low collision probability. Given
// effectively random hash distribution, for n values hashed into a space of m hash
// strings, the probability of a collision is roughly n^2 / (2m). We use 16 hex digits
// of MD5 i.e. 2^64 possible hash strings, so given 2^16 stored values the collision
// probability is about 2^-33 =~ 0.0000000001 , i.e. negligible.
//
// Prefix with a letter to prevent all numeric hashes, and to constrain the space of
// possible object property values.
return 'h' + SparkMD5.hash( stringified ).slice( 0, 16 );
};
/**
* Get the hashes of values in the store
*
* Same as index but with arrays.
*
* @method
* @param {Object[]} values Values to lookup or store
* @return {string[]} The hashes of the values in the store
*/
ve.dm.IndexValueStore.prototype.indexes = function ( values ) {
var i, length, hashes = [];
for ( i = 0, length = values.length; i < length; i++ ) {
hashes.push( this.index( values[ i ] ) );
}
return hashes;
};
/**
* Get the value stored for a particular hash
*
* @method
* @param {string} hash Hash to look up
* @return {Object|undefined} Value stored for this hash if present, else undefined
*/
ve.dm.IndexValueStore.prototype.value = function ( hash ) {
return this.hashStore[ hash ];
};
/**
* Get the values stored for a list of hashes
*
* Same as value but with arrays.
*
* @method
* @param {string[]} hashes Hashes to lookup
* @return {Array} Values for these hashes (undefined for any not present)
*/
ve.dm.IndexValueStore.prototype.values = function ( hashes ) {
var i, length, values = [];
for ( i = 0, length = hashes.length; i < length; i++ ) {
values.push( this.value( hashes[ i ] ) );
}
return values;
};
/**
* Merge another store into this store.
*
* It is allowed for the two stores to have differently computed hashes, so long as two values
* will (with high probability) have the same hash only if equal. In this case, equivalent
* values can have two different hashes.
*
* Values are added in the order they appear in the other store. Objects added to the store are
* added by reference, not cloned, unlike in .index()
*
* @param {ve.dm.IndexValueStore} other Store to merge into this one
*/
ve.dm.IndexValueStore.prototype.merge = function ( other ) {
var i, len, hash;
for ( i = 0, len = other.hashes.length; i < len; i++ ) {
hash = other.hashes[ i ];
if ( !Object.prototype.hasOwnProperty.call( this.hashStore, hash ) ) {
this.hashStore[ hash ] = other.hashStore[ hash ];
this.hashes.push( hash );
}
}
};
Mini Shell