Move aggregates out of vortex-array into vortex-aggregates

Cargo.toml

@@ -14,6 +14,7 @@ members = [
     "vortex-cloud",
     "vortex-array",
     "vortex-functions",
+    "vortex-aggregates",
     "vortex-expr",
     "vortex-arrow",
     "vortex-row",
@@ -300,6 +301,7 @@ zstd = { version = "0.13.3", default-features = false, features = [
 
 # BEGIN crates published by this project
 vortex = { version = "0.1.0", path = "./vortex" }
+vortex-aggregates = { version = "0.1.0", path = "./vortex-aggregates", default-features = false }
 vortex-alp = { version = "0.1.0", path = "./encodings/alp", default-features = false }
 vortex-array = { version = "0.1.0", path = "./vortex-array", default-features = false }
 vortex-array-macros = { version = "0.1.0", path = "./vortex-array-macros" }

benchmarks/string-bench/src/lib.rs

@@ -33,8 +33,8 @@ use parquet::arrow::ParquetRecordBatchStreamBuilder;
 use parquet::arrow::ProjectionMask;
 use tokio::fs::File;
 use vortex::VortexSessionDefault;
-use vortex::aggregate_fn::fns::uncompressed_size_in_bytes::UncompressedSizeOpts;
-use vortex::aggregate_fn::fns::uncompressed_size_in_bytes::uncompressed_size_in_bytes;
+use vortex::aggregates::uncompressed_size_in_bytes::UncompressedSizeOpts;
+use vortex::aggregates::uncompressed_size_in_bytes::uncompressed_size_in_bytes;
 use vortex::array::ArrayRef;
 use vortex::array::ExecutionCtx;
 use vortex::array::IntoArray;

encodings/alp/Cargo.toml

@@ -20,6 +20,7 @@ alp = { workspace = true }
 itertools = { workspace = true }
 num-traits = { workspace = true }
 prost = { workspace = true }
+vortex-aggregates = { workspace = true }
 vortex-array = { workspace = true }
 vortex-buffer = { workspace = true }
 vortex-error = { workspace = true }

encodings/alp/src/alp/compute/nan_count.rs

@@ -1,11 +1,11 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright the Vortex contributors
 
+use vortex_aggregates::nan_count::NanCount;
+use vortex_aggregates::nan_count::nan_count;
 use vortex_array::ArrayRef;
 use vortex_array::ExecutionCtx;
 use vortex_array::aggregate_fn::AggregateFnRef;
-use vortex_array::aggregate_fn::fns::nan_count::NanCount;
-use vortex_array::aggregate_fn::fns::nan_count::nan_count;
 use vortex_array::aggregate_fn::kernels::DynAggregateKernel;
 use vortex_array::scalar::Scalar;
 use vortex_error::VortexResult;

encodings/alp/src/lib.rs

@@ -16,9 +16,9 @@
 //!
 //! [paper]: https://ir.cwi.nl/pub/33334/33334.pdf
 
+use vortex_aggregates::nan_count::NanCount;
 use vortex_array::ArrayVTable;
 use vortex_array::aggregate_fn::AggregateFnVTable;
-use vortex_array::aggregate_fn::fns::nan_count::NanCount;
 use vortex_array::aggregate_fn::session::AggregateFnSessionExt;
 use vortex_array::arrays::patched::use_experimental_patches;
 use vortex_array::session::ArraySessionExt;

encodings/datetime-parts/Cargo.toml

@@ -18,6 +18,7 @@ workspace = true
 [dependencies]
 num-traits = { workspace = true }
 prost = { workspace = true }
+vortex-aggregates = { workspace = true }
 vortex-array = { workspace = true }
 vortex-buffer = { workspace = true }
 vortex-compute = { workspace = true }

encodings/datetime-parts/src/compute/compare.rs

@@ -288,10 +288,10 @@ mod test {
     use std::sync::LazyLock;
 
     use rstest::rstest;
+    use vortex_aggregates::sum::sum;
     use vortex_array::ArrayRef;
     use vortex_array::ExecutionCtx;
     use vortex_array::VortexSessionExecute;
-    use vortex_array::aggregate_fn::fns::sum::sum;
     use vortex_array::arrays::BoolArray;
     use vortex_array::arrays::ConstantArray;
     use vortex_array::arrays::PrimitiveArray;

encodings/datetime-parts/src/compute/is_constant.rs

@@ -1,11 +1,11 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright the Vortex contributors
 
+use vortex_aggregates::is_constant::IsConstant;
+use vortex_aggregates::is_constant::is_constant;
 use vortex_array::ArrayRef;
 use vortex_array::ExecutionCtx;
 use vortex_array::aggregate_fn::AggregateFnRef;
-use vortex_array::aggregate_fn::fns::is_constant::IsConstant;
-use vortex_array::aggregate_fn::fns::is_constant::is_constant;
 use vortex_array::aggregate_fn::kernels::DynAggregateKernel;
 use vortex_array::scalar::Scalar;
 use vortex_error::VortexResult;

encodings/datetime-parts/src/lib.rs

@@ -11,9 +11,9 @@ mod compute;
 mod ops;
 mod timestamp;
 
+use vortex_aggregates::is_constant::IsConstant;
 use vortex_array::ArrayVTable;
 use vortex_array::aggregate_fn::AggregateFnVTable;
-use vortex_array::aggregate_fn::fns::is_constant::IsConstant;
 use vortex_array::aggregate_fn::session::AggregateFnSessionExt;
 use vortex_array::session::ArraySessionExt;
 use vortex_session::VortexSession;

encodings/decimal-byte-parts/Cargo.toml

@@ -18,6 +18,7 @@ workspace = true
 [dependencies]
 num-traits = { workspace = true }
 prost = { workspace = true }
+vortex-aggregates = { workspace = true }
 vortex-array = { workspace = true }
 vortex-buffer = { workspace = true }
 vortex-error = { workspace = true }

encodings/decimal-byte-parts/src/decimal_byte_parts/compute/is_constant.rs

@@ -1,12 +1,12 @@
 // SPDX-License-Identifier: Apache-2.0
 // SPDX-FileCopyrightText: Copyright the Vortex contributors
 
+use vortex_aggregates::is_constant::IsConstant;
+use vortex_aggregates::is_constant::is_constant;
 use vortex_array::ArrayRef;
 use vortex_array::ArrayView;
 use vortex_array::ExecutionCtx;
 use vortex_array::aggregate_fn::AggregateFnRef;
-use vortex_array::aggregate_fn::fns::is_constant::IsConstant;
-use vortex_array::aggregate_fn::fns::is_constant::is_constant;
 use vortex_array::aggregate_fn::kernels::DynAggregateKernel;
 use vortex_array::scalar::Scalar;
 use vortex_error::VortexResult;

encodings/decimal-byte-parts/src/decimal_byte_parts/prop_tests.rs

@@ -33,6 +33,10 @@ const MAX_I256_HIGH: i128 = 10i128.pow(37);
 /// vectorized path is not trivially degenerate.
 const MAX_LEN: usize = 48;
 
+/// Bound on a drawn most significant part. Every part count's assembled value then stays inside
+/// its declared precision: `10^18 * 2^192` is about `6.3e75`, below the 76 digits of the widest.
+const MAX_MSP: i64 = 10i64.pow(18) - 1;
+
 fn ctx() -> ExecutionCtx {
     let session = VortexSession::empty();
     crate::initialize(&session);
@@ -104,7 +108,7 @@ fn canonicalize(array: ArrayRef, ctx: &mut ExecutionCtx) -> DecimalArray {
 fn draw_encoded(tc: &TestCase) -> (DecimalBytePartsArray, usize) {
     let lower_part_count = tc.draw(gs::integers::<usize>().min_value(0).max_value(3));
     let msp: Vec<i64> = tc.draw(
-        gs::vecs(gs::integers::<i64>())
+        gs::vecs(gs::integers::<i64>().min_value(-MAX_MSP).max_value(MAX_MSP))
             .min_size(1)
             .max_size(MAX_LEN),
     );

encodings/decimal-byte-parts/src/lib.rs

@@ -13,9 +13,9 @@ use decimal_byte_parts::compute::is_constant::DecimalBytePartsIsConstantKernel;
 /// an i128 decimal could be converted into a [i64, u64] with further narrowing applied to either
 /// value.
 pub use decimal_byte_parts::*;
+use vortex_aggregates::is_constant::IsConstant;
 use vortex_array::ArrayVTable;
 use vortex_array::aggregate_fn::AggregateFnVTable;
-use vortex_array::aggregate_fn::fns::is_constant::IsConstant;
 use vortex_array::aggregate_fn::session::AggregateFnSessionExt;
 use vortex_array::session::ArraySessionExt;
 use vortex_session::VortexSession;

encodings/fastlanes/Cargo.toml

@@ -25,6 +25,7 @@ lending-iterator = { workspace = true }
 num-traits = { workspace = true }
 prost = { workspace = true }
 rand = { workspace = true, optional = true }
+vortex-aggregates = { workspace = true }
 vortex-array = { workspace = true }
 vortex-buffer = { workspace = true }
 vortex-compute = { workspace = true }

encodings/fastlanes/src/bitpacking/array/bitpack_compress/mod.rs

@@ -19,14 +19,13 @@ pub use global::gather_patches;
 pub use global::test_harness;
 use vortex_array::ExecutionCtx;
 use vortex_array::arrays::PrimitiveArray;
-use vortex_array::match_each_integer_ptype;
+use vortex_array::arrays::primitive::integer_bounds;
 use vortex_error::VortexResult;
 use vortex_error::vortex_bail;
 use vortex_error::vortex_ensure;
 
 /// Return an error unless `array` holds integers that are all non-negative, which bit-packing
 /// requires.
-#[expect(unused_comparisons, clippy::absurd_extreme_comparisons)]
 fn ensure_non_negative_integers(
     array: &PrimitiveArray,
     ctx: &mut ExecutionCtx,
@@ -37,9 +36,8 @@ fn ensure_non_negative_integers(
         array.ptype()
     );
     if array.ptype().is_signed_int() {
-        let has_negative_values = match_each_integer_ptype!(array.ptype(), |P| {
-            array.statistics().compute_min::<P>(ctx).unwrap_or_default() < 0
-        });
+        let has_negative_values =
+            integer_bounds(array.as_ref(), ctx)?.is_some_and(|bounds| bounds.min < 0);
         if has_negative_values {
             vortex_bail!(InvalidArgument: "cannot bitpack array containing negative integers")
         }

encodings/fastlanes/src/bitpacking/compute/is_constant.rs

@@ -6,13 +6,13 @@ use std::ops::Range;
 
 use itertools::Itertools;
 use lending_iterator::LendingIterator;
+use vortex_aggregates::is_constant::IsConstant;
+use vortex_aggregates::is_constant::primitive::IS_CONST_LANE_WIDTH;
+use vortex_aggregates::is_constant::primitive::compute_is_constant;
 use vortex_array::ArrayRef;
 use vortex_array::ArrayView;
 use vortex_array::ExecutionCtx;
 use vortex_array::aggregate_fn::AggregateFnRef;
-use vortex_array::aggregate_fn::fns::is_constant::IsConstant;
-use vortex_array::aggregate_fn::fns::is_constant::primitive::IS_CONST_LANE_WIDTH;
-use vortex_array::aggregate_fn::fns::is_constant::primitive::compute_is_constant;
 use vortex_array::aggregate_fn::kernels::DynAggregateKernel;
 use vortex_array::arrays::PrimitiveArray;
 use vortex_array::dtype::IntegerPType;
@@ -193,9 +193,9 @@ fn apply_patches_idx_typed<T: BitPackedUnpack, I: IntegerPType>(
 mod tests {
     use std::sync::LazyLock;
 
+    use vortex_aggregates::is_constant::is_constant;
     use vortex_array::IntoArray;
     use vortex_array::VortexSessionExecute;
-    use vortex_array::aggregate_fn::fns::is_constant::is_constant;
     use vortex_buffer::buffer;
     use vortex_error::VortexResult;
     use vortex_session::VortexSession;