vortex_sampling_compressor/compressors/
zigzag.rs

1use vortex_array::aliases::hash_set::HashSet;
2use vortex_array::array::PrimitiveArray;
3use vortex_array::stats::Stat;
4use vortex_array::variants::PrimitiveArrayTrait;
5use vortex_array::{Array, Encoding, EncodingId, IntoArray};
6use vortex_error::VortexResult;
7use vortex_zigzag::{zigzag_encode, ZigZagArray, ZigZagEncoding};
8
9use crate::compressors::{CompressedArray, CompressionTree, EncodingCompressor};
10use crate::{constants, SamplingCompressor};
11
12#[derive(Debug)]
13pub struct ZigZagCompressor;
14
15impl EncodingCompressor for ZigZagCompressor {
16    fn id(&self) -> &str {
17        ZigZagEncoding::ID.as_ref()
18    }
19
20    fn cost(&self) -> u8 {
21        constants::ZIGZAG_COST
22    }
23
24    fn can_compress(&self, array: &Array) -> Option<&dyn EncodingCompressor> {
25        // Only support primitive arrays
26        let parray = PrimitiveArray::maybe_from(array)?;
27
28        // Only supports signed integers
29        if !parray.ptype().is_signed_int() {
30            return None;
31        }
32
33        // Only compress if the array has negative values
34        // TODO(ngates): also check that Stat::Max is less than half the max value of the type
35        parray
36            .statistics()
37            .compute_as::<i64>(Stat::Min)
38            .filter(|&min| min < 0)
39            .map(|_| self as &dyn EncodingCompressor)
40    }
41
42    fn compress<'a>(
43        &'a self,
44        array: &Array,
45        like: Option<CompressionTree<'a>>,
46        ctx: SamplingCompressor<'a>,
47    ) -> VortexResult<CompressedArray<'a>> {
48        let encoded = zigzag_encode(PrimitiveArray::try_from(array.clone())?)?;
49        let compressed =
50            ctx.compress(&encoded.encoded(), like.as_ref().and_then(|l| l.child(0)))?;
51        Ok(CompressedArray::compressed(
52            ZigZagArray::try_new(compressed.array)?.into_array(),
53            Some(CompressionTree::new(self, vec![compressed.path])),
54            array,
55        ))
56    }
57
58    fn used_encodings(&self) -> HashSet<EncodingId> {
59        HashSet::from([ZigZagEncoding::ID])
60    }
61}
vortex_sampling_compressor/compressors/zigzag.rs

vortex_sampling_compressor/compressors/
zigzag.rs
