compactly 0.1.6

Compactly encode data types using adaptive arithmetic coding
Documentation 
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

use super::Encode;

/// A unsigned integer with a value less than `N`.
///
/// This type is unseful if you want to compactly encode a value like the
/// variant of an enum with a range that is known at compil time to be limited,
/// with a range that may not be an integer number of bits.
///
/// Some values will take fewer bits than others, and *all* values will be
/// adapted independently so if e.g. any two variants are exclusively present
/// and equally likely, they will each take only a tiny bit more than a single
/// bit to encode (eventually).
#[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord)]
pub struct ULessThan<const N: usize>(usize);

impl<const N: usize> ULessThan<N> {
    /// Construct a new `ULessThan<N>`.
    ///
    /// Panics if the value is not less than `N`.
    #[inline]
    pub const fn new(value: usize) -> Self {
        if value < N {
            ULessThan(value)
        } else {
            panic!("Invalid value in compactly::ULessThan")
        }
    }
}

impl<const N: usize> From<ULessThan<N>> for usize {
    #[inline]
    fn from(value: ULessThan<N>) -> Self {
        value.0
    }
}

impl<const N: usize> TryFrom<usize> for ULessThan<N> {
    type Error = ();
    #[inline]
    fn try_from(value: usize) -> Result<Self, Self::Error> {
        if value < N {
            Ok(ULessThan(value))
        } else {
            Err(())
        }
    }
}

#[derive(Debug, Default, Clone, PartialEq, Eq)]
pub struct ULessThanContext<const N: usize> {
    /// This uses way more context than is needed, because I couldn't find an
    /// elegant way to map the N needed context to the possible bit sequences.  :(
    bits: Vec<<bool as Encode>::Context>,
}

impl<const N: usize> ULessThanContext<N> {
    #[inline]
    fn index_mut(&mut self, index: usize) -> &mut <bool as Encode>::Context {
        if self.bits.len() <= index {
            self.bits.reserve(index - self.bits.len());
            while self.bits.len() <= index {
                self.bits.push(Default::default());
            }
        }
        &mut self.bits[index]
    }
}

#[inline]
fn half(i: usize) -> usize {
    let half = i / 2;
    if half > 1 {
        1 << half.ilog(2)
    } else {
        half
    }
}

impl<const N: usize> Encode for ULessThan<N> {
    type Context = ULessThanContext<N>;
    #[inline]
    fn encode<E: super::EntropyCoder>(&self, writer: &mut E, ctx: &mut Self::Context) {
        let mut filled_up = 0;
        let mut accumulated_value = 0;
        let mut bits_chosen = 0;
        // println!("N={N} and value {}", self.0);
        let mut possible_values_left = N;
        let mut value_considered = half(possible_values_left); // big endian bits, splitting values by half each time
        let mut i = 1;
        while accumulated_value + value_considered < N && possible_values_left > 1 {
            let bit = self.0 >= accumulated_value + value_considered;
            // println!(
            //     "bit {i} is {bit:?} == {} >= {}",
            //     self.0,
            //     accumulated_value + value_considered
            // );
            // println!(
            //     "{}: bit {i} is {bit:?} with context {} considering {value_considered} with {possible_values_left} values left to consider",
            //     self.0,
            //     filled_up + bits_chosen
            // );
            let ctx = ctx.index_mut(filled_up + bits_chosen);
            bit.encode(writer, ctx);
            filled_up += i;
            if bit {
                bits_chosen += 1 << i;
                accumulated_value += value_considered;
                possible_values_left -= value_considered;
            } else {
                possible_values_left = value_considered;
            }
            // println!("E {i} ==> {filled_up} -> {accumulated_value}");
            value_considered = half(possible_values_left);
            i += 1;
        }
    }
    #[inline]
    fn decode<D: super::EntropyDecoder>(
        reader: &mut D,
        ctx: &mut Self::Context,
    ) -> Result<Self, std::io::Error> {
        let mut filled_up = 0;
        let mut accumulated_value = 0;
        let mut bits_chosen = 0;
        let mut possible_values_left = N;
        let mut value_considered = half(possible_values_left); // big endian bits, splitting values by half each time
        let mut i = 1;
        while accumulated_value + value_considered < N && possible_values_left > 1 {
            let ctx = ctx.index_mut(filled_up + bits_chosen);
            let bit = bool::decode(reader, ctx)?;
            filled_up += i;
            if bit {
                bits_chosen += 1 << i;
                accumulated_value += value_considered;
                possible_values_left -= value_considered;
            } else {
                possible_values_left = value_considered;
            }
            // println!("D {i} ==> {filled_up} -> {accumulated_value}");
            value_considered = half(possible_values_left);
            i += 1;
        }
        Ok(Self(accumulated_value))
    }
}

#[test]
fn size() {
    use super::assert_bits;
    fn test_urange<const N: usize>() {
        for i in 0..N {
            let v = ULessThan::<N>::new(i);
            println!("Testing ULessThan::<{N}>::new({i})");
            let encoded = super::encode(&v);
            let decoded = super::decode::<ULessThan<N>>(&encoded).unwrap();
            assert_eq!(decoded, v);
        }
    }
    test_urange::<1>();
    test_urange::<2>();
    test_urange::<3>();
    test_urange::<4>();
    test_urange::<5>();
    test_urange::<6>();
    test_urange::<7>();
    test_urange::<8>();
    test_urange::<9>();
    test_urange::<10>();
    test_urange::<255>();
    test_urange::<256>();
    test_urange::<257>();

    assert_bits!(ULessThan::<3>::try_from(0).unwrap(), 1);
    assert_bits!(ULessThan::<3>::try_from(1).unwrap(), 2);
    assert_bits!(ULessThan::<3>::try_from(2).unwrap(), 1);

    assert_bits!(ULessThan::<5>::try_from(0).unwrap(), 2);
    assert_bits!(ULessThan::<5>::try_from(1).unwrap(), 2);
    assert_bits!(ULessThan::<5>::try_from(2).unwrap(), 2);
    assert_bits!(ULessThan::<5>::try_from(3).unwrap(), 3);
    assert_bits!(ULessThan::<5>::try_from(4).unwrap(), 1);

    assert_bits!(ULessThan::<6>::try_from(0).unwrap(), 2);
    assert_bits!(ULessThan::<6>::try_from(1).unwrap(), 2);
    assert_bits!(ULessThan::<6>::try_from(2).unwrap(), 3);
    assert_bits!(ULessThan::<6>::try_from(3).unwrap(), 3);
    assert_bits!(ULessThan::<6>::try_from(4).unwrap(), 3);
    assert_bits!(ULessThan::<6>::try_from(5).unwrap(), 1);

    assert_bits!(ULessThan::<128>::try_from(0).unwrap(), 7);
    assert_bits!(ULessThan::<128>::try_from(1).unwrap(), 7);
    assert_bits!(ULessThan::<128>::try_from(127).unwrap(), 3);

    assert_bits!(ULessThan::<256>::try_from(0).unwrap(), 8);
    assert_bits!(ULessThan::<256>::try_from(1).unwrap(), 8);
    assert_bits!(ULessThan::<256>::try_from(255).unwrap(), 3);
}