17050663773

Committed 18 Aug 2025 07:38PM UTC coverage: 49.695%. First build

Build # 17050663773

Build Type

Pull #4177

github

Committed by

web-flow

Commit Message

Merge cdc94ae08 into 7eb8ac9fa

Pull Request Pull Request #4177: feat: ArrayOperations infallible, eager validation + new_unchecked

Run Details

654 of 1386 new or added lines in 156 files covered. (47.19%)

19893 of 40030 relevant lines covered (49.7%)

229974.76 hits per line

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

4.55

/encodings/bytebool/src/array.rs

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

use std::fmt::Debug;

use arrow_buffer::BooleanBuffer;
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::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 = BooleanBuffer::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) -> ArrayRef {
        ByteBoolArray::new(
            array.buffer().slice(start..stop),
            array.validity().slice(start, stop),
        )
        .into_array()
    }

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

vortex-data / vortex / 17050663773

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