16979224669

Committed 14 Aug 2025 11:42PM UTC coverage: 23.728%. First build

Build # 16979224669

Build Type

Pull #2456

github

Committed by

web-flow

Commit Message

Merge 30049dfa7 into aaf3e36ad

Pull Request Pull Request #2456: feat: basic BoolBuffer / BoolBufferMut

Run Details

68 of 1065 new or added lines in 82 files covered. (6.38%)

8616 of 36312 relevant lines covered (23.73%)

146.37 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

4.41

/encodings/bytebool/src/array.rs

// SPDX-License-Identifier: Apache-2.0
// SPDX-FileCopyrightText: Copyright the Vortex contributors

use std::fmt::Debug;

use vortex_array::arrays::BoolArray;
use vortex_array::stats::{ArrayStats, StatsSetRef};
use vortex_array::validity::Validity;
use vortex_array::vtable::{
    ArrayVTable, CanonicalVTable, NotSupported, OperationsVTable, VTable, ValidityHelper,
    ValidityVTableFromValidityHelper,
};
use vortex_array::{ArrayRef, Canonical, EncodingId, EncodingRef, IntoArray, vtable};
use vortex_buffer::{BitBuffer, ByteBuffer};
use vortex_dtype::DType;
use vortex_error::{VortexResult, vortex_panic};
use vortex_scalar::Scalar;

vtable!(ByteBool);

impl VTable for ByteBoolVTable {
    type Array = ByteBoolArray;
    type Encoding = ByteBoolEncoding;

    type ArrayVTable = Self;
    type CanonicalVTable = Self;
    type OperationsVTable = Self;
    type ValidityVTable = ValidityVTableFromValidityHelper;
    type VisitorVTable = Self;
    type ComputeVTable = NotSupported;
    type EncodeVTable = NotSupported;
    type SerdeVTable = Self;

    fn id(_encoding: &Self::Encoding) -> EncodingId {
        EncodingId::new_ref("vortex.bytebool")
    }

    fn encoding(_array: &Self::Array) -> EncodingRef {
        EncodingRef::new_ref(ByteBoolEncoding.as_ref())
    }
}

#[derive(Clone, Debug)]
pub struct ByteBoolArray {
    dtype: DType,
    buffer: ByteBuffer,
    validity: Validity,
    stats_set: ArrayStats,
}

#[derive(Clone, Debug)]
pub struct ByteBoolEncoding;

impl ByteBoolArray {
    pub fn new(buffer: ByteBuffer, validity: Validity) -> Self {
        let length = buffer.len();
        if let Some(vlen) = validity.maybe_len()
            && length != vlen
        {
            vortex_panic!(
                "Buffer length ({}) does not match validity length ({})",
                length,
                vlen
            );
        }
        Self {
            dtype: DType::Bool(validity.nullability()),
            buffer,
            validity,
            stats_set: Default::default(),
        }
    }

    // TODO(ngates): deprecate construction from vec
    pub fn from_vec<V: Into<Validity>>(data: Vec<bool>, validity: V) -> Self {
        let validity = validity.into();
        // SAFETY: we are transmuting a Vec<bool> into a Vec<u8>
        let data: Vec<u8> = unsafe { std::mem::transmute(data) };
        Self::new(ByteBuffer::from(data), validity)
    }

    pub fn buffer(&self) -> &ByteBuffer {
        &self.buffer
    }

    pub fn as_slice(&self) -> &[bool] {
        // Safety: The internal buffer contains byte-sized bools
        unsafe { std::mem::transmute(self.buffer().as_slice()) }
    }
}

impl ValidityHelper for ByteBoolArray {
    fn validity(&self) -> &Validity {
        &self.validity
    }
}

impl ArrayVTable<ByteBoolVTable> for ByteBoolVTable {
    fn len(array: &ByteBoolArray) -> usize {
        array.buffer.len()
    }

    fn dtype(array: &ByteBoolArray) -> &DType {
        &array.dtype
    }

    fn stats(array: &ByteBoolArray) -> StatsSetRef<'_> {
        array.stats_set.to_ref(array.as_ref())
    }
}

impl CanonicalVTable<ByteBoolVTable> for ByteBoolVTable {
    fn canonicalize(array: &ByteBoolArray) -> VortexResult<Canonical> {
        let boolean_buffer = BitBuffer::from(array.as_slice());
        let validity = array.validity().clone();
        Ok(Canonical::Bool(BoolArray::new(boolean_buffer, validity)))
    }
}

impl OperationsVTable<ByteBoolVTable> for ByteBoolVTable {
    fn slice(array: &ByteBoolArray, start: usize, stop: usize) -> VortexResult<ArrayRef> {
        Ok(ByteBoolArray::new(
            array.buffer().slice(start..stop),
            array.validity().slice(start, stop)?,
        )
        .into_array())
    }

    fn scalar_at(array: &ByteBoolArray, index: usize) -> VortexResult<Scalar> {
        Ok(Scalar::bool(
            array.buffer()[index] == 1,
            array.dtype().nullability(),
        ))
    }
}

impl From<Vec<bool>> for ByteBoolArray {
    fn from(value: Vec<bool>) -> Self {
        Self::from_vec(value, Validity::AllValid)
    }
}

impl From<Vec<Option<bool>>> for ByteBoolArray {
    fn from(value: Vec<Option<bool>>) -> Self {
        let validity = Validity::from_iter(value.iter().map(|v| v.is_some()));

        // This doesn't reallocate, and the compiler even vectorizes it
        let data = value.into_iter().map(Option::unwrap_or_default).collect();

        Self::from_vec(data, validity)
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    // #[cfg_attr(miri, ignore)]
    // #[test]
    // fn test_bytebool_metadata() {
    //     check_metadata(
    //         "bytebool.metadata",
    //         SerdeMetadata(ByteBoolMetadata {
    //             validity: ValidityMetadata::AllValid,
    //         }),
    //     );
    // }

    #[test]
    fn test_validity_construction() {
        let v = vec![true, false];
        let v_len = v.len();

        let arr = ByteBoolArray::from(v);
        assert_eq!(v_len, arr.len());

        for idx in 0..arr.len() {
            assert!(arr.is_valid(idx).unwrap());
        }

        let v = vec![Some(true), None, Some(false)];
        let arr = ByteBoolArray::from(v);
        assert!(arr.is_valid(0).unwrap());
        assert!(!arr.is_valid(1).unwrap());
        assert!(arr.is_valid(2).unwrap());
        assert_eq!(arr.len(), 3);

        let v: Vec<Option<bool>> = vec![None, None];
        let v_len = v.len();

        let arr = ByteBoolArray::from(v);
        assert_eq!(v_len, arr.len());

        for idx in 0..arr.len() {
            assert!(!arr.is_valid(idx).unwrap());
        }
        assert_eq!(arr.len(), 2);
    }
}

1	// SPDX-License-Identifier: Apache-2.0
2	// SPDX-FileCopyrightText: Copyright the Vortex contributors
3
4	use std::fmt::Debug;
5
6	use vortex_array::arrays::BoolArray;
7	use vortex_array::stats::{ArrayStats, StatsSetRef};
8	use vortex_array::validity::Validity;
9	use vortex_array::vtable::{
10	ArrayVTable, CanonicalVTable, NotSupported, OperationsVTable, VTable, ValidityHelper,
11	ValidityVTableFromValidityHelper,
12	};
13	use vortex_array::{ArrayRef, Canonical, EncodingId, EncodingRef, IntoArray, vtable};
14	use vortex_buffer::{BitBuffer, ByteBuffer};
15	use vortex_dtype::DType;
16	use vortex_error::{VortexResult, vortex_panic};
17	use vortex_scalar::Scalar;
18
19	vtable!(ByteBool);
20
21	impl VTable for ByteBoolVTable {
22	type Array = ByteBoolArray;
23	type Encoding = ByteBoolEncoding;
24
25	type ArrayVTable = Self;
26	type CanonicalVTable = Self;
27	type OperationsVTable = Self;
28	type ValidityVTable = ValidityVTableFromValidityHelper;
29	type VisitorVTable = Self;
30	type ComputeVTable = NotSupported;
31	type EncodeVTable = NotSupported;
32	type SerdeVTable = Self;
33
34	fn id(_encoding: &Self::Encoding) -> EncodingId {	2✔
35	EncodingId::new_ref("vortex.bytebool")	2✔
36	}	2✔
37
38	fn encoding(_array: &Self::Array) -> EncodingRef {	×
39	EncodingRef::new_ref(ByteBoolEncoding.as_ref())	×
40	}	×
41	}
42
43	#[derive(Clone, Debug)]
44	pub struct ByteBoolArray {
45	dtype: DType,
46	buffer: ByteBuffer,
47	validity: Validity,
48	stats_set: ArrayStats,
49	}
50
51	#[derive(Clone, Debug)]
52	pub struct ByteBoolEncoding;
53
54	impl ByteBoolArray {
55	pub fn new(buffer: ByteBuffer, validity: Validity) -> Self {	×
56	let length = buffer.len();	×
57	if let Some(vlen) = validity.maybe_len()	×
58	&& length != vlen	×
59	{
60	vortex_panic!(	×
61	"Buffer length ({}) does not match validity length ({})",	×
62	length,
63	vlen
64	);
65	}	×
66	Self {	×
67	dtype: DType::Bool(validity.nullability()),	×
68	buffer,	×
69	validity,	×
70	stats_set: Default::default(),	×
71	}	×
72	}	×
73
74	// TODO(ngates): deprecate construction from vec
75	pub fn from_vec<V: Into<Validity>>(data: Vec<bool>, validity: V) -> Self {	×
76	let validity = validity.into();	×
77	// SAFETY: we are transmuting a Vec<bool> into a Vec<u8>
78	let data: Vec<u8> = unsafe { std::mem::transmute(data) };	×
79	Self::new(ByteBuffer::from(data), validity)	×
80	}	×
81
82	pub fn buffer(&self) -> &ByteBuffer {	×
83	&self.buffer	×
84	}	×
85
86	pub fn as_slice(&self) -> &[bool] {	×
87	// Safety: The internal buffer contains byte-sized bools
88	unsafe { std::mem::transmute(self.buffer().as_slice()) }	×
89	}	×
90	}
91
92	impl ValidityHelper for ByteBoolArray {
93	fn validity(&self) -> &Validity {	×
94	&self.validity	×
95	}	×
96	}
97
98	impl ArrayVTable<ByteBoolVTable> for ByteBoolVTable {
99	fn len(array: &ByteBoolArray) -> usize {	×
100	array.buffer.len()	×
101	}	×
102
103	fn dtype(array: &ByteBoolArray) -> &DType {	×
104	&array.dtype	×
105	}	×
106
107	fn stats(array: &ByteBoolArray) -> StatsSetRef<'_> {	×
108	array.stats_set.to_ref(array.as_ref())	×
109	}	×
110	}
111
112	impl CanonicalVTable<ByteBoolVTable> for ByteBoolVTable {
113	fn canonicalize(array: &ByteBoolArray) -> VortexResult<Canonical> {	×
NEW 114	let boolean_buffer = BitBuffer::from(array.as_slice());	×
115	let validity = array.validity().clone();	×
116	Ok(Canonical::Bool(BoolArray::new(boolean_buffer, validity)))	×
117	}	×
118	}
119
120	impl OperationsVTable<ByteBoolVTable> for ByteBoolVTable {
121	fn slice(array: &ByteBoolArray, start: usize, stop: usize) -> VortexResult<ArrayRef> {	×
122	Ok(ByteBoolArray::new(	×
123	array.buffer().slice(start..stop),	×
124	array.validity().slice(start, stop)?,	×
125	)
126	.into_array())	×
127	}	×
128
129	fn scalar_at(array: &ByteBoolArray, index: usize) -> VortexResult<Scalar> {	×
130	Ok(Scalar::bool(	×
131	array.buffer()[index] == 1,	×
132	array.dtype().nullability(),	×
133	))	×
134	}	×
135	}
136
137	impl From<Vec<bool>> for ByteBoolArray {
138	fn from(value: Vec<bool>) -> Self {	×
139	Self::from_vec(value, Validity::AllValid)	×
140	}	×
141	}
142
143	impl From<Vec<Option<bool>>> for ByteBoolArray {
144	fn from(value: Vec<Option<bool>>) -> Self {	×
145	let validity = Validity::from_iter(value.iter().map(\|v\| v.is_some()));	×
146
147	// This doesn't reallocate, and the compiler even vectorizes it
148	let data = value.into_iter().map(Option::unwrap_or_default).collect();	×
149
150	Self::from_vec(data, validity)	×
151	}	×
152	}
153
154	#[cfg(test)]
155	mod tests {
156	use super::*;
157
158	// #[cfg_attr(miri, ignore)]
159	// #[test]
160	// fn test_bytebool_metadata() {
161	// check_metadata(
162	// "bytebool.metadata",
163	// SerdeMetadata(ByteBoolMetadata {
164	// validity: ValidityMetadata::AllValid,
165	// }),
166	// );
167	// }
168
169	#[test]
170	fn test_validity_construction() {
171	let v = vec![true, false];
172	let v_len = v.len();
173
174	let arr = ByteBoolArray::from(v);
175	assert_eq!(v_len, arr.len());
176
177	for idx in 0..arr.len() {
178	assert!(arr.is_valid(idx).unwrap());
179	}
180
181	let v = vec![Some(true), None, Some(false)];
182	let arr = ByteBoolArray::from(v);
183	assert!(arr.is_valid(0).unwrap());
184	assert!(!arr.is_valid(1).unwrap());
185	assert!(arr.is_valid(2).unwrap());
186	assert_eq!(arr.len(), 3);
187
188	let v: Vec<Option<bool>> = vec![None, None];
189	let v_len = v.len();
190
191	let arr = ByteBoolArray::from(v);
192	assert_eq!(v_len, arr.len());
193
194	for idx in 0..arr.len() {
195	assert!(!arr.is_valid(idx).unwrap());
196	}
197	assert_eq!(arr.len(), 2);
198	}
199	}

vortex-data / vortex / 16979224669

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous