1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244

use crate::{
    approximate_element_count, approximate_false_positive_probability, bitset::Bitset,
    optimal_bit_count, optimal_number_of_hashers, BloomFilter,
};
use ahash::AHasher;
use std::fmt::Debug;
use std::hash::{Hash, Hasher};

/// A bloom filter that uses a single Hasher that can be seeded to simulate an arbitrary number
/// of hash functions.
///
/// Internally, the implementation uses *ahash::AHasher*.
pub struct SeededBloomFilter {
    number_of_hashers: usize,
    bitset: Bitset,
    bits_per_hasher: usize,
}

impl SeededBloomFilter {
    /// Initialize a new instance of SeededBloomFilter that guarantees that the false positive rate
    /// is less than *desired_false_positive_probability* for up to *desired_capacity*
    /// elements.
    ///
    /// SeededBloomFilter uses a single hash function that can be seeded to simulate an arbitrary
    /// number of hash functions.
    ///
    /// # Panics
    ///
    /// Panics if desired_capacity == 0
    ///
    /// # Examples
    /// ```
    /// use bloom_filter_simple::{BloomFilter,SeededBloomFilter};
    ///
    /// fn main() {
    ///     // We plan on storing at most 10 elements
    ///     let desired_capacity = 10;
    ///     // We want to assure that the chance of a false positive is less than 0.0001.
    ///     let desired_fp_probability = 0.0001;
    ///
    ///     // We initialize a new SeededBloomFilter by specifying the desired Hashers as type
    ///     // parameters
    ///     let mut filter = SeededBloomFilter::new(desired_capacity, desired_fp_probability);
    /// }
    /// ```
    pub fn new(desired_capacity: usize, desired_false_positive_probability: f64) -> Self {
        if desired_capacity == 0 {
            panic!("an empty bloom filter is not defined");
        }
        let bit_count = optimal_bit_count(desired_capacity, desired_false_positive_probability);
        let number_of_hashers = optimal_number_of_hashers(desired_capacity, bit_count);
        let bits_per_hasher = (bit_count as f64 / number_of_hashers as f64).ceil() as usize;
        Self {
            bitset: Bitset::new(bits_per_hasher * number_of_hashers),
            number_of_hashers,
            bits_per_hasher,
        }
    }

    /// Approximate number of elements stored.
    /// Approximation technique taken from Wikipedia:
    /// > Wikipedia, ["Bloom filter"](https://en.wikipedia.org/wiki/Bloom_filter#Approximating_the_number_of_items_in_a_Bloom_filter) [Accessed: 02.12.2020]
    pub fn approximate_element_count(&self) -> f64 {
        approximate_element_count(
            self.number_of_hashers,
            self.bits_per_hasher,
            self.bitset.count_ones(),
        )
    }

    /// Return the current approximate false positive probability which depends on the current
    /// number of elements in the filter.
    ///
    /// The probability is given as a value in the interval [0,1]
    /// Approximation technique taken from Sagi Kedmi:
    /// > S. Kedmi, ["Bloom Filters for the Perplexed"](https://sagi.io/bloom-filters-for-the-perplexed/), July 2017 [Accessed: 02.12.2020]
    pub fn approximate_current_false_positive_probability(&self) -> f64 {
        approximate_false_positive_probability(
            self.number_of_hashers,
            self.bits_per_hasher,
            self.approximate_element_count(),
        )
    }

    /// Creates a union of this bloom filter and 'other', which means 'contains' of the resulting
    /// bloom filter will always return true for elements inserted in either this bloom filter or in
    /// 'other' before creation.
    ///
    /// # Panics
    ///
    /// Panics if the desired capacity or desired false positive probability of 'self' and 'other'
    /// differ.
    /// 
    /// # Examples
    /// 
    /// Union of two bloom filters with the same configuration.
    /// ```
    /// use bloom_filter_simple::{BloomFilter,SeededBloomFilter};
    ///
    /// fn main() {
    ///     // The configuration of both bloom filters has to be the same
    ///     let desired_capacity = 10_000;
    ///     let desired_fp_probability = 0.0001;
    ///
    ///     // We initialize two new SeededBloomFilter 
    ///     let mut filter_one = SeededBloomFilter::new(desired_capacity, desired_fp_probability);
    ///     let mut filter_two = SeededBloomFilter::new(desired_capacity, desired_fp_probability);
    /// 
    ///     // Insert elements into the first filter
    ///     filter_one.insert(&0);
    ///     filter_one.insert(&1);
    /// 
    ///     // Insert elements into the second filter
    ///     filter_two.insert(&2);
    ///     filter_two.insert(&3);
    ///     
    ///     // Now we retrieve the union of both filters
    ///     let filter_union = filter_one.union(&filter_two);
    ///
    ///     // The union will return true for a 'contains' check for the elements inserted 
    ///     // previously into at least one of the constituent filters.
    ///     assert_eq!(true, filter_union.contains(&0));
    ///     assert_eq!(true, filter_union.contains(&1));
    ///     assert_eq!(true, filter_union.contains(&2));
    ///     assert_eq!(true, filter_union.contains(&3));
    /// }
    /// ```
    pub fn union(&self, other: &Self) -> Self {
        if !self.eq_configuration(other) {
            panic!("unable to union k-m bloom filters with different configurations");
        }
        Self {
            number_of_hashers: self.number_of_hashers,
            bitset: self.bitset.union(&other.bitset),
            bits_per_hasher: self.bits_per_hasher,
        }
    }

    /// Creates a intersection of this bloom filter and 'other', which means 'contains' of the resulting
    /// bloom filter will always return true for elements inserted both in this bloom filter and in
    /// 'other' before creation.
    /// The false positive probability of the resulting bloom filter is at most the false positive
    /// probability of 'other' or 'self'.
    /// The false positive probability of the resulting bloom filter may be bigger than the false
    /// positive probability of a new empty bloom filter with the intersecting elements inserted.
    /// The functions 'approximate_current_false_positive_probability' and 'approximate_element_count'
    /// called on the resulting bloom filter may return too big approximations.
    ///
    /// # Panics
    ///
    /// Panics if the desired capacity or desired false positive probability of 'self' and 'other'
    /// differ.
    ///
    /// # Examples
    /// 
    /// Intersection of two bloom filters with the same configuration.
    /// ```
    /// use bloom_filter_simple::{BloomFilter,SeededBloomFilter};
    ///
    /// fn main() {
    ///     // The configuration of both bloom filters has to be the same
    ///     let desired_capacity = 10_000;
    ///     let desired_fp_probability = 0.0001;
    ///
    ///     // We initialize two new SeededBloomFilter 
    ///     let mut filter_one = SeededBloomFilter::new(desired_capacity, desired_fp_probability);
    ///     let mut filter_two = SeededBloomFilter::new(desired_capacity, desired_fp_probability);
    /// 
    ///     // Insert elements into the first filter
    ///     filter_one.insert(&0);
    ///     filter_one.insert(&1);
    /// 
    ///     // Insert elements into the second filter
    ///     filter_two.insert(&1);
    ///     filter_two.insert(&2);
    ///     
    ///     // Now we retrieve the intersection of both filters
    ///     let filter_intersection = filter_one.intersect(&filter_two);
    ///
    ///     // The intersection will return true for a 'contains' check for the elements inserted 
    ///     // previously into both constituent filters.
    ///     assert_eq!(false, filter_intersection.contains(&0));
    ///     assert_eq!(true, filter_intersection.contains(&1));
    ///     assert_eq!(false, filter_intersection.contains(&2));
    /// }
    /// ```
    pub fn intersect(&self, other: &Self) -> Self {
        if !self.eq_configuration(other) {
            panic!("unable to intersect k-m bloom filters with different configurations");
        }
        Self {
            number_of_hashers: self.number_of_hashers,
            bitset: self.bitset.intersect(&other.bitset),
            bits_per_hasher: self.bits_per_hasher,
        }
    }

    /// Checks whether two bloom filters were created with the same desired capacity and desired false
    /// positive probability.
    pub fn eq_configuration(&self, other: &Self) -> bool {
        self.number_of_hashers == other.number_of_hashers
            && self.bits_per_hasher == other.bits_per_hasher
    }

    fn index<T>(i: usize, bits_per_hash: usize, data: &T) -> usize
    where
        T: Hash,
    {
        let mut hasher = AHasher::new_with_keys(i as u128, i as u128);
        data.hash(&mut hasher);
        i * bits_per_hash + hasher.finish() as usize % bits_per_hash
    }
}

impl Debug for SeededBloomFilter {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        write!(f, "SeededBloomFilter{{{:?}}}", self.bitset)
    }
}

impl BloomFilter for SeededBloomFilter {
    fn insert<T>(&mut self, data: &T)
    where
        T: Hash,
    {
        for i in 0..self.number_of_hashers {
            self.bitset
                .set(Self::index(i, self.bits_per_hasher, &data), true);
        }
    }

    fn contains<T>(&self, data: &T) -> bool
    where
        T: Hash,
    {
        for i in 0..self.number_of_hashers {
            if !self.bitset.get(Self::index(i, self.bits_per_hasher, &data)) {
                return false;
            }
        }

        return true;
    }
}