vortex_array/arrays/bool/vtable/
mod.rs

1// SPDX-License-Identifier: Apache-2.0
2// SPDX-FileCopyrightText: Copyright the Vortex contributors
3
4use std::hash::Hash;
5use std::hash::Hasher;
6
7use kernel::PARENT_KERNELS;
8use prost::Message;
9use vortex_error::VortexExpect;
10use vortex_error::VortexResult;
11use vortex_error::vortex_bail;
12use vortex_error::vortex_ensure;
13use vortex_error::vortex_panic;
14use vortex_session::VortexSession;
15
16use crate::ArrayRef;
17use crate::ExecutionCtx;
18use crate::ExecutionResult;
19use crate::array::Array;
20use crate::array::ArrayView;
21use crate::array::VTable;
22use crate::arrays::bool::BoolData;
23use crate::arrays::bool::array::SLOT_NAMES;
24use crate::buffer::BufferHandle;
25use crate::dtype::DType;
26use crate::serde::ArrayChildren;
27use crate::validity::Validity;
28mod canonical;
29mod kernel;
30mod operations;
31mod validity;
32
33use vortex_session::registry::CachedId;
34
35use crate::Precision;
36use crate::array::ArrayId;
37use crate::arrays::bool::compute::rules::RULES;
38use crate::hash::ArrayEq;
39use crate::hash::ArrayHash;
40
41/// A [`Bool`]-encoded Vortex array.
42pub type BoolArray = Array<Bool>;
43
44#[derive(prost::Message)]
45pub struct BoolMetadata {
46    // The offset in bits must be <8
47    #[prost(uint32, tag = "1")]
48    pub offset: u32,
49}
50
51impl ArrayHash for BoolData {
52    fn array_hash<H: Hasher>(&self, state: &mut H, precision: Precision) {
53        self.bits.array_hash(state, precision);
54        self.offset.hash(state);
55    }
56}
57
58impl ArrayEq for BoolData {
59    fn array_eq(&self, other: &Self, precision: Precision) -> bool {
60        self.offset == other.offset && self.bits.array_eq(&other.bits, precision)
61    }
62}
63
64impl VTable for Bool {
65    type ArrayData = BoolData;
66
67    type OperationsVTable = Self;
68    type ValidityVTable = Self;
69
70    fn id(&self) -> ArrayId {
71        static ID: CachedId = CachedId::new("vortex.bool");
72        *ID
73    }
74
75    fn nbuffers(_array: ArrayView<'_, Self>) -> usize {
76        1
77    }
78
79    fn buffer(array: ArrayView<'_, Self>, idx: usize) -> BufferHandle {
80        match idx {
81            0 => array.bits.clone(),
82            _ => vortex_panic!("BoolArray buffer index {idx} out of bounds"),
83        }
84    }
85
86    fn buffer_name(_array: ArrayView<'_, Self>, idx: usize) -> Option<String> {
87        match idx {
88            0 => Some("bits".to_string()),
89            _ => None,
90        }
91    }
92
93    fn serialize(
94        array: ArrayView<'_, Self>,
95        _session: &VortexSession,
96    ) -> VortexResult<Option<Vec<u8>>> {
97        assert!(array.offset < 8, "Offset must be <8, got {}", array.offset);
98        Ok(Some(
99            BoolMetadata {
100                offset: u32::try_from(array.offset).vortex_expect("checked"),
101            }
102            .encode_to_vec(),
103        ))
104    }
105
106    fn validate(
107        &self,
108        data: &BoolData,
109        dtype: &DType,
110        len: usize,
111        slots: &[Option<ArrayRef>],
112    ) -> VortexResult<()> {
113        let DType::Bool(nullability) = dtype else {
114            vortex_bail!("Expected bool dtype, got {dtype:?}");
115        };
116        vortex_ensure!(
117            data.bits.len() * 8 >= data.offset + len,
118            "BoolArray buffer with offset {} cannot back outer length {} (buffer bits = {})",
119            data.offset,
120            len,
121            data.bits.len() * 8
122        );
123
124        let validity = crate::array::child_to_validity(&slots[0], *nullability);
125        if let Some(validity_len) = validity.maybe_len() {
126            vortex_ensure!(
127                validity_len == len,
128                "BoolArray validity len {} does not match outer length {}",
129                validity_len,
130                len
131            );
132        }
133
134        Ok(())
135    }
136
137    fn deserialize(
138        &self,
139        dtype: &DType,
140        len: usize,
141        metadata: &[u8],
142
143        buffers: &[BufferHandle],
144        children: &dyn ArrayChildren,
145        _session: &VortexSession,
146    ) -> VortexResult<crate::array::ArrayParts<Self>> {
147        let metadata = BoolMetadata::decode(metadata)?;
148        if buffers.len() != 1 {
149            vortex_bail!("Expected 1 buffer, got {}", buffers.len());
150        }
151
152        let validity = if children.is_empty() {
153            Validity::from(dtype.nullability())
154        } else if children.len() == 1 {
155            let validity = children.get(0, &Validity::DTYPE, len)?;
156            Validity::Array(validity)
157        } else {
158            vortex_bail!("Expected 0 or 1 child, got {}", children.len());
159        };
160
161        let buffer = buffers[0].clone();
162        let slots = BoolData::make_slots(&validity, len);
163        let data = BoolData::try_new_from_handle(buffer, metadata.offset as usize, len, validity)?;
164        Ok(crate::array::ArrayParts::new(self.clone(), dtype.clone(), len, data).with_slots(slots))
165    }
166
167    fn slot_name(_array: ArrayView<'_, Self>, idx: usize) -> String {
168        SLOT_NAMES[idx].to_string()
169    }
170
171    fn execute(array: Array<Self>, _ctx: &mut ExecutionCtx) -> VortexResult<ExecutionResult> {
172        Ok(ExecutionResult::done(array))
173    }
174
175    fn execute_parent(
176        array: ArrayView<'_, Self>,
177        parent: &ArrayRef,
178        child_idx: usize,
179        ctx: &mut ExecutionCtx,
180    ) -> VortexResult<Option<ArrayRef>> {
181        PARENT_KERNELS.execute(array, parent, child_idx, ctx)
182    }
183
184    fn reduce_parent(
185        array: ArrayView<'_, Self>,
186        parent: &ArrayRef,
187        child_idx: usize,
188    ) -> VortexResult<Option<ArrayRef>> {
189        RULES.evaluate(array, parent, child_idx)
190    }
191}
192
193#[derive(Clone, Debug)]
194pub struct Bool;
195
196#[cfg(test)]
197mod tests {
198    use vortex_buffer::ByteBufferMut;
199    use vortex_session::registry::ReadContext;
200
201    use crate::ArrayContext;
202    use crate::IntoArray;
203    use crate::LEGACY_SESSION;
204    use crate::arrays::BoolArray;
205    use crate::assert_arrays_eq;
206    use crate::serde::SerializeOptions;
207    use crate::serde::SerializedArray;
208
209    #[test]
210    fn test_nullable_bool_serde_roundtrip() {
211        let array = BoolArray::from_iter([Some(true), None, Some(false), None]);
212        let dtype = array.dtype().clone();
213        let len = array.len();
214
215        let ctx = ArrayContext::empty();
216        let serialized = array
217            .clone()
218            .into_array()
219            .serialize(&ctx, &LEGACY_SESSION, &SerializeOptions::default())
220            .unwrap();
221
222        let mut concat = ByteBufferMut::empty();
223        for buf in serialized {
224            concat.extend_from_slice(buf.as_ref());
225        }
226        let parts = SerializedArray::try_from(concat.freeze()).unwrap();
227        let decoded = parts
228            .decode(
229                &dtype,
230                len,
231                &ReadContext::new(ctx.to_ids()),
232                &LEGACY_SESSION,
233            )
234            .unwrap();
235
236        assert_arrays_eq!(decoded, array);
237    }
238}
vortex_array/arrays/bool/vtable/mod.rs

vortex_array/arrays/bool/vtable/
mod.rs
