andreyp / libbloom

Yet another bloom filter realization

=== LIBBLOOM

by Andrey Mukhin <[e-mail]>

== What is libbloom?
libbloom is a simple classes set that provides bloom filter technology.
The Bloom filter is a space-efficient probabilistic data structure that is used
to test whether an element is a member of a set (see http://en.wikipedia.org/wiki/Bloom_filter).

== How it works
First of all you must define main parameters:
  - type of elements (must be C++ POD type)
  - type of return-value of the hash function
  - false positives
  - max number of elements
  - hash function which you want to use
  - initialization vector for the hash function
  - and maybe number of hash functions

You must set all of those parameters into the Bloom class constructor.
For example:
  Bloom<unsigned long, unsigned> b(1000000, 1., 0, &bloom::FNVHash);
  - type of elements -- unsigned long
  - type of return-value of the hash function -- unsigned
  - false positives -- 1.(percent)
  - max number of elements -- 1000000
  - hash function wich you want to use -- bloom::FNVHash
  - initialization vector for the hash function -- 0

So, when we have constructed Bloom object, we can get such values as:
  - min number of bits that our bloom filter needs -- b.getBitsNumber()
  - min number of bytes -- b.getBytesNumber()

Now we can create bloom filter storage, for example:
  unsigned char bloom_storage[b.getBytesNumber()];

Before starting to work with the filter we must specify a storage for it:
  b.setBitStorage(bloom_storage, b.getBytesNumber());

And in the end we can put any elements into the storage.
  unsigned long e = 128500;
  bool f_new_element = b.fillBitSet(e);

When this work is done we can specify a new storage, e.g.:
  b.setBitStorage(....);

Thus libbloom provides only bloom filter algorithm and doesn't include hash functions or storages.
Therefore we can use different combinations of functions and storages we need.