BE-629: Add spherical k-means entity clustering endpoint via /entities/embeddings/clusters

.github/workflows/codspeed.yml

@@ -70,6 +70,9 @@ jobs:
         with:
           scope: ${{ matrix.name }}
 
+      - name: Warm up repository
+        uses: ./.github/actions/warm-up-repo
+
       - name: Build the benchmark target
         run: turbo run build:codspeed --filter=${{ matrix.name }}
 

Cargo.toml

@@ -222,6 +222,7 @@ quote                              = { version = "1.0.41", default-features = fa
 rand                               = { version = "0.10.0", default-features = false }
 rand_core                          = { version = "0.10.0", default-features = false }
 rand_distr                         = { version = "0.6.0", default-features = false }
+rand_xoshiro                       = { version = "0.8.1" }
 rapidfuzz                          = { version = "0.5.0", default-features = false }
 ratatui                            = { version = "0.30.0" }
 rayon                              = { version = "1.11.0", default-features = false }

apps/hash-graph/Cargo.toml

@@ -33,6 +33,7 @@ futures = { workspace = true }
 jsonwebtoken = { workspace = true }
 mimalloc = { workspace = true }
 multiaddr = { workspace = true }
+rayon = { workspace = true }
 regex = { workspace = true }
 reqwest = { workspace = true, features = ["rustls"] }
 simple-mermaid = { workspace = true }

apps/hash-graph/src/subcommand/mod.rs

@@ -7,7 +7,7 @@ mod snapshot;
 mod type_fetcher;
 
 use core::time::Duration;
-use std::time::Instant;
+use std::{thread::available_parallelism, time::Instant};
 
 use clap::Parser;
 use error_stack::{Report, ensure};
@@ -141,11 +141,22 @@ pub enum Subcommand {
     ReindexCache(Box<ReindexCacheArgs>),
 }
 
+#[expect(clippy::integer_division, clippy::integer_division_remainder_used)]
 fn block_on(
     future: impl Future<Output = Result<(), Report<GraphError>>>,
     service_name: &'static str,
     tracing_config: TracingConfig,
 ) -> Result<(), Report<GraphError>> {
+    rayon::ThreadPoolBuilder::new()
+        .num_threads(
+            available_parallelism()
+                .map_or(1, |cores| cores.get() / 2)
+                .max(1),
+        )
+        .thread_name(|index| format!("rayon-{index}"))
+        .build_global()
+        .expect("rayon pool should be initialized exactly once");
+
     tokio::runtime::Builder::new_multi_thread()
         .enable_all()
         .build()