feat: kmeans1d

.github/workflows/psql.yml

@@ -80,6 +80,7 @@ jobs:
           wget --quiet -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
           sudo apt-get update
           sudo apt-get install -y clang-16
+          sudo update-alternatives --install /usr/bin/clang clang /usr/bin/clang-16 128
       - name: Set up Pgrx
         run: |
           # pg_config

.github/workflows/release.yml

@@ -77,6 +77,7 @@ jobs:
           wget --quiet -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
           sudo apt-get update
           sudo apt-get install -y clang-16
+          sudo update-alternatives --install /usr/bin/clang clang /usr/bin/clang-16 128
       - name: Set up Pgrx
         run: |
           # pg_config

.github/workflows/rust.yml

@@ -84,6 +84,7 @@ jobs:
           wget --quiet -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
           sudo apt-get update
           sudo apt-get install -y clang-16
+          sudo update-alternatives --install /usr/bin/clang clang /usr/bin/clang-16 128
       - name: Set up Pgrx
         run: |
           # pg_config
@@ -149,6 +150,7 @@ jobs:
           wget --quiet -O - https://apt.llvm.org/llvm-snapshot.gpg.key | sudo apt-key add -
           sudo apt-get update
           sudo apt-get install -y clang-16
+          sudo update-alternatives --install /usr/bin/clang clang /usr/bin/clang-16 128
       - name: Set up Pgrx
         run: |
           # pg_config

crates/base/src/scalar/mod.rs

@@ -2,6 +2,8 @@ mod f32;
 mod half_f16;
 mod i8;
 
+use std::iter::Sum;
+
 pub use f32::F32;
 pub use half_f16::F16;
 pub use i8::I8;
@@ -19,7 +21,9 @@ pub trait ScalarLike:
     + num_traits::Zero
     + num_traits::NumOps
     + num_traits::NumAssignOps
+    + Default
     + crate::pod::Pod
+    + Sum
 {
     fn from_f32(x: f32) -> Self;
     fn to_f32(self) -> f32;

crates/c/build.rs

@@ -2,7 +2,7 @@ fn main() {
     println!("cargo:rerun-if-changed=src/f16.h");
     println!("cargo:rerun-if-changed=src/f16.c");
     cc::Build::new()
-        .compiler("clang-16")
+        .compiler("clang")
         .file("./src/f16.c")
         .opt_level(3)
         .flag("-fassociative-math")

crates/c/src/f16.c

@@ -1,3 +1,7 @@
+#if !(__clang_major__ >= 16)
+#error "clang version must be >= 16"
+#endif
+
 #include "f16.h"
 #include <math.h>
 

crates/common/src/lib.rs

@@ -9,4 +9,3 @@ pub mod remap;
 pub mod sample;
 pub mod variants;
 pub mod vec2;
-pub mod vec3;

crates/common/src/sample.rs

@@ -10,10 +10,10 @@ pub fn sample<O: Operator>(vectors: &impl Vectors<O>) -> Vec2<Scalar<O>> {
     let n = vectors.len();
     let m = std::cmp::min(SAMPLES as u32, n);
     let f = super::rand::sample_u32(&mut rand::thread_rng(), n, m);
-    let mut samples = Vec2::new(vectors.dims(), m as usize);
+    let mut samples = Vec2::zeros((m as usize, vectors.dims() as usize));
     for i in 0..m {
         let v = vectors.vector(f[i as usize] as u32).to_vec();
-        samples[i as usize].copy_from_slice(&v);
+        samples[(i as usize,)].copy_from_slice(&v);
     }
     samples
 }
@@ -27,12 +27,12 @@ pub fn sample_subvector_transform<O: Operator>(
     let n = vectors.len();
     let m = std::cmp::min(SAMPLES as u32, n);
     let f = super::rand::sample_u32(&mut rand::thread_rng(), n, m);
-    let mut samples = Vec2::new((e - s) as u32, m as usize);
+    let mut samples = Vec2::zeros((m as usize, e - s));
     for i in 0..m {
         let v = transform(vectors.vector(f[i as usize] as u32))
             .as_borrowed()
             .to_vec();
-        samples[i as usize].copy_from_slice(&v[s..e]);
+        samples[(i as usize,)].copy_from_slice(&v[s..e]);
     }
     samples
 }

crates/common/src/vec2.rs

@@ -1,70 +1,75 @@
-use base::pod::Pod;
 use serde::{Deserialize, Serialize};
-use std::ops::{Deref, DerefMut, Index, IndexMut};
+use std::ops::{Index, IndexMut};
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct Vec2<T> {
-    dims: u32,
-    v: Vec<T>,
+    shape: (usize, usize),
+    base: Vec<T>,
 }
 
-impl<T: Pod + Ord> Vec2<T> {
-    pub fn new(dims: u32, n: usize) -> Self {
+impl<T: Default + Copy> Vec2<T> {
+    pub fn zeros(shape: (usize, usize)) -> Self {
         Self {
-            dims,
-            v: base::pod::zeroed_vec(dims as usize * n),
+            shape,
+            base: vec![T::default(); shape.0 * shape.1],
         }
     }
-    pub fn dims(&self) -> u32 {
-        self.dims
+    pub fn from_vec(shape: (usize, usize), base: Vec<T>) -> Self {
+        assert_eq!(shape.0 * shape.1, base.len());
+        Self { shape, base }
     }
-    pub fn len(&self) -> usize {
-        self.v.len() / self.dims as usize
+}
+
+impl<T: Copy> Vec2<T> {
+    pub fn copy_within(&mut self, (l_i,): (usize,), (r_i,): (usize,)) {
+        assert!(l_i < self.shape.0);
+        assert!(r_i < self.shape.0);
+        let src_from = l_i * self.shape.1;
+        let src_to = src_from + self.shape.1;
+        let dest = r_i * self.shape.1;
+        self.base.copy_within(src_from..src_to, dest);
+    }
+}
+
+impl<T> Vec2<T> {
+    pub fn shape_0(&self) -> usize {
+        self.shape.0
     }
-    pub fn is_empty(&self) -> bool {
-        self.len() == 0
+    pub fn shape_1(&self) -> usize {
+        self.shape.1
     }
-    pub fn argsort(&self) -> Vec<usize> {
-        let mut index: Vec<usize> = (0..self.len()).collect();
-        index.sort_by_key(|i| &self[*i]);
-        index
+    pub fn as_slice(&self) -> &[T] {
+        self.base.as_slice()
     }
-    pub fn copy_within(&mut self, i: usize, j: usize) {
-        assert!(i < self.len() && j < self.len());
-        unsafe {
-            if i != j {
-                let src = self.v.as_ptr().add(self.dims as usize * i);
-                let dst = self.v.as_mut_ptr().add(self.dims as usize * j);
-                std::ptr::copy_nonoverlapping(src, dst, self.dims as usize);
-            }
-        }
+    pub fn as_mut_slice(&mut self) -> &mut [T] {
+        self.base.as_mut_slice()
     }
 }
 
-impl<T> Index<usize> for Vec2<T> {
+impl<T> Index<(usize,)> for Vec2<T> {
     type Output = [T];
 
-    fn index(&self, index: usize) -> &Self::Output {
-        &self.v[self.dims as usize * index..][..self.dims as usize]
+    fn index(&self, (i,): (usize,)) -> &Self::Output {
+        &self.base[i * self.shape.1..][..self.shape.1]
     }
 }
 
-impl<T> IndexMut<usize> for Vec2<T> {
-    fn index_mut(&mut self, index: usize) -> &mut Self::Output {
-        &mut self.v[self.dims as usize * index..][..self.dims as usize]
+impl<T> IndexMut<(usize,)> for Vec2<T> {
+    fn index_mut(&mut self, (i,): (usize,)) -> &mut Self::Output {
+        &mut self.base[i * self.shape.1..][..self.shape.1]
     }
 }
 
-impl<T> Deref for Vec2<T> {
-    type Target = [T];
+impl<T> Index<(usize, usize)> for Vec2<T> {
+    type Output = T;
 
-    fn deref(&self) -> &Self::Target {
-        self.v.deref()
+    fn index(&self, (i, j): (usize, usize)) -> &Self::Output {
+        &self.base[i * self.shape.1..][j]
     }
 }
 
-impl<T> DerefMut for Vec2<T> {
-    fn deref_mut(&mut self) -> &mut Self::Target {
-        self.v.deref_mut()
+impl<T> IndexMut<(usize, usize)> for Vec2<T> {
+    fn index_mut(&mut self, (i, j): (usize, usize)) -> &mut Self::Output {
+        &mut self.base[i * self.shape.1..][j]
     }
 }

crates/ivf/src/ivf_naive.rs

@@ -98,8 +98,8 @@ fn from_nothing<O: Op>(
     rayon::check();
     let centroids = {
         let mut samples = samples;
-        for i in 0..samples.len() {
-            O::elkan_k_means_normalize(&mut samples[i]);
+        for i in 0..samples.shape_0() {
+            O::elkan_k_means_normalize(&mut samples[(i,)]);
         }
         k_means(nlist as usize, samples)
     };

crates/ivf/src/ivf_residual.rs

@@ -61,7 +61,7 @@ impl<O: Op> IvfResidual<O> {
         );
         let vectors = lists
             .iter()
-            .map(|&(_, i)| O::vector_sub(vector, &self.centroids[i]))
+            .map(|&(_, i)| O::vector_sub(vector, &self.centroids[(i,)]))
             .collect::<Vec<_>>();
         let mut reranker = self
             .quantization
@@ -104,8 +104,8 @@ fn from_nothing<O: Op>(
     rayon::check();
     let centroids = {
         let mut samples = samples;
-        for i in 0..samples.len() {
-            O::elkan_k_means_normalize(&mut samples[i]);
+        for i in 0..samples.shape_0() {
+            O::elkan_k_means_normalize(&mut samples[(i,)]);
         }
         k_means(nlist as usize, samples)
     };
@@ -141,7 +141,7 @@ fn from_nothing<O: Op>(
                 O::elkan_k_means_normalize(&mut vector);
                 k_means_lookup(&vector, &centroids)
             };
-            O::vector_sub(vector, &centroids[target])
+            O::vector_sub(vector, &centroids[(target,)])
         },
     );
     let payloads = MmapArray::create(

crates/k_means/Cargo.toml

@@ -9,6 +9,7 @@ rand.workspace = true
 
 base = { path = "../base" }
 common = { path = "../common" }
+smawk = "0.3.2"
 stoppable_rayon = { path = "../stoppable_rayon" }
 
 [lints]