php.bloom.filter

Bloom filter - is the best way to check object existance and save memory and disk usage.

Usage

Examples and benchmarks are avialable in examples directory.

Basic parameters:

• entries_max (int) Maximal number of entries objects. Default: 100.
• error_chance (float) (0;1) Chance of error check existance. Used for automatic setup next 2 parameters. Default: 0.001.
• set_size (int) Size of set object. Default: calculated.
• hash_count (int) Count of unique hash Objects. Default: calculated.
• counter (boolean) Uses only if ypu want to remove elements too. Slows down. Default: false.
• hash (array) Parameters for Hash objects.
• strtolower (boolean) Put string to lower case or not. Default: true;

Simple code example:

include 'bloom.class.php';

$parameters = array(
  'entries_max' => 2
);
$bloom = new Bloom($parameters);

/**
* Set value to Bloom filter
* You can put single string or array of infinite deep of strings
*
* @param mixed
* @return BloomObject
*/
$bloom->set('Some string');
echo $bloom->has('Some string'); //true
echo $bloom->has('Some string', false); //0

/**
* Test set with given array or string, to check it existance
* You can put single string or array of infinite deep of strings
*
* @param mixed (array) or (string)
* @param boolean return boolean or float number of tests
* @return mixed (array) or (boolean) or (float)
*/
echo $bloom->has('Other string'); //false

/**
* Unset value from Bloom filter
* You can put single string or array of infinite deep of strings
* Only works with counter parameter
* @param mixed
* @return mixed (boolean) or (array)
*/
$bloom->delete('Some string');
echo $bloom->has('Some string'); //false

Caching

You can also cache the hole object, just using serialize() and unserialize() functions

When to use

• Get/Set operations are more than 0.001
• You need a fast check with big set, more than 100 000 entries

Use cases:

• to cooldown HDD usage
• to speedup checking of object excistance (with counter 40% slower, but still more faster than native)
• to save memory