No more Borrow

CHANGELOG.md

@@ -1,5 +1,11 @@
 # Change Log
 
+## [0.2.0] - 2024-08-09
+
+### Improved
+
+* Argument specifying a thread pool are now simply references.
+
 ## [0.1.4] - 2024-08-09
 
 ### Fixed

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "webgraph"
-version = "0.1.4"
+version = "0.2.0"
 edition = "2021"
 description = "A Rust port of the WebGraph framework (http://webgraph.di.unimi.it/)."
 repository = "https://github.com/vigna/webgraph-rs/"

src/algo/llp/gap_cost.rs

@@ -5,11 +5,10 @@
  * SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later
  */
 
-use std::borrow::Borrow;
-
 use crate::traits::*;
 use dsi_progress_logger::prelude::*;
 use lender::prelude::*;
+use rayon::ThreadPool;
 use sux::prelude::*;
 
 /// Computes the gap cost, that is, the sum of the costs of the logarithms
@@ -21,7 +20,7 @@ pub(crate) fn compute_log_gap_cost<G: SequentialGraph + Sync>(
     graph: &G,
     arc_granularity: usize,
     deg_cumul: &(impl Succ<Input = usize, Output = usize> + Send + Sync),
-    thread_pool: impl Borrow<rayon::ThreadPool>,
+    thread_pool: &ThreadPool,
     pr: Option<&mut ProgressLogger>,
 ) -> f64 {
     graph.par_apply(

src/cli/from/arcs.rs

@@ -138,7 +138,7 @@ pub fn from_csv(args: CliArgs) -> Result<()> {
         &g,
         args.num_nodes,
         args.ca.into(),
-        thread_pool,
+        &thread_pool,
         dir,
         &target_endianness.unwrap_or_else(|| BE::NAME.into()),
     )

src/cli/transform/simplify.rs

@@ -199,7 +199,7 @@ where
                     &sorted,
                     graph.num_nodes(),
                     args.ca.into(),
-                    thread_pool,
+                    &thread_pool,
                     dir,
                     &target_endianness,
                 )?;
@@ -258,7 +258,7 @@ where
                     &sorted,
                     graph.num_nodes(),
                     args.ca.into(),
-                    thread_pool,
+                    &thread_pool,
                     dir,
                     &target_endianness,
                 )?;
@@ -282,7 +282,7 @@ where
                 &sorted,
                 seq_graph.num_nodes(),
                 args.ca.into(),
-                thread_pool,
+                &thread_pool,
                 dir,
                 &target_endianness,
             )?;

src/cli/transform/transpose.rs

@@ -78,7 +78,7 @@ where
         &sorted,
         sorted.num_nodes(),
         args.ca.into(),
-        thread_pool,
+        &thread_pool,
         dir,
         &target_endianness.unwrap_or_else(|| E::NAME.into()),
     )?;

src/graphs/bvgraph/comp/impls.rs

@@ -10,7 +10,8 @@ use anyhow::{ensure, Context, Result};
 use dsi_bitstream::prelude::*;
 use dsi_progress_logger::prelude::*;
 use lender::prelude::*;
-use std::borrow::Borrow;
+use rayon::ThreadPool;
+
 use std::fs::File;
 use std::io::{BufReader, BufWriter};
 use std::path::{Path, PathBuf};
@@ -198,7 +199,7 @@ impl BvComp<()> {
         graph: &G,
         num_nodes: usize,
         compression_flags: CompFlags,
-        threads: impl Borrow<rayon::ThreadPool>,
+        threads: &ThreadPool,
         tmp_dir: P,
         endianness: &str,
     ) -> Result<u64>
@@ -214,9 +215,7 @@ impl BvComp<()> {
                 // compress the transposed graph
                 Self::parallel_iter::<BigEndian, _>(
                     basename,
-                    graph
-                        .split_iter(threads.borrow().current_num_threads())
-                        .into_iter(),
+                    graph.split_iter(threads.current_num_threads()).into_iter(),
                     num_nodes,
                     compression_flags,
                     threads,
@@ -231,9 +230,7 @@ impl BvComp<()> {
                 // compress the transposed graph
                 Self::parallel_iter::<LittleEndian, _>(
                     basename,
-                    graph
-                        .split_iter(threads.borrow().current_num_threads())
-                        .into_iter(),
+                    graph.split_iter(threads.current_num_threads()).into_iter(),
                     num_nodes,
                     compression_flags,
                     threads,
@@ -249,7 +246,7 @@ impl BvComp<()> {
         basename: impl AsRef<Path> + Send + Sync,
         graph: &(impl SequentialGraph + SplitLabeling),
         compression_flags: CompFlags,
-        threads: impl Borrow<rayon::ThreadPool>,
+        threads: &ThreadPool,
         tmp_dir: impl AsRef<Path>,
     ) -> Result<u64>
     where
@@ -258,9 +255,7 @@ impl BvComp<()> {
     {
         Self::parallel_iter(
             basename,
-            graph
-                .split_iter(threads.borrow().current_num_threads())
-                .into_iter(),
+            graph.split_iter(threads.current_num_threads()).into_iter(),
             graph.num_nodes(),
             compression_flags,
             threads,
@@ -278,14 +273,13 @@ impl BvComp<()> {
         iter: impl Iterator<Item = L>,
         num_nodes: usize,
         compression_flags: CompFlags,
-        threads: impl Borrow<rayon::ThreadPool>,
+        threads: &ThreadPool,
         tmp_dir: impl AsRef<Path>,
     ) -> Result<u64>
     where
         BufBitWriter<E, WordAdapter<usize, BufWriter<std::fs::File>>>: CodeWrite<E>,
         BufBitReader<E, WordAdapter<u32, BufReader<std::fs::File>>>: BitRead<E>,
     {
-        let thread_pool = threads.borrow();
         let tmp_dir = tmp_dir.as_ref();
         let basename = basename.as_ref();
 
@@ -296,7 +290,7 @@ impl BvComp<()> {
 
         let thread_path = |thread_id: usize| tmp_dir.join(format!("{:016x}.bitstream", thread_id));
 
-        thread_pool.in_place_scope(|s| {
+        threads.in_place_scope(|s| {
             let cp_flags = &compression_flags;
 
             for (thread_id, mut thread_lender) in iter.enumerate() {

src/graphs/bvgraph/mod.rs

@@ -7,11 +7,11 @@
 
 //! An implementation of the Bv format.
 //!
-//! The format has been described by Paolo Boldi and Sebastiano Vigna in [“The
+//! The format has been described by Paolo Boldi and Sebastiano Vigna in “[The
 //!  WebGraph Framework I: Compression
-//!  Techniques”](http://vigna.di.unimi.it/papers.php#BoVWFI), in *Proc. of the
+//!  Techniques](http://vigna.di.unimi.it/papers.php#BoVWFI)”, in *Proc. of the
 //!  13th international conference on World Wide Web*, WWW 2004, pages 595-602,
-//!  ACM [DOI
+//!  ACM. [DOI
 //!  10.1145/988672.988752](https://dl.acm.org/doi/10.1145/988672.988752).
 //!
 //! The implementation is compatible with the [Java

src/traits/labels.rs

@@ -33,7 +33,8 @@ use core::{
 use dsi_progress_logger::prelude::*;
 use impl_tools::autoimpl;
 use lender::*;
-use std::borrow::Borrow;
+use rayon::ThreadPool;
+
 use sux::traits::Succ;
 
 /// A labeling that can be accessed sequentially.
@@ -105,7 +106,7 @@ pub trait SequentialLabeling {
         func: F,
         fold: R,
         node_granularity: usize,
-        thread_pool: impl Borrow<rayon::ThreadPool>,
+        thread_pool: &ThreadPool,
         pl: Option<&mut ProgressLogger>,
     ) -> A
     where
@@ -117,7 +118,6 @@ pub trait SequentialLabeling {
         let pl_lock = pl.map(std::sync::Mutex::new);
         let num_nodes = self.num_nodes();
         let num_scoped_threads = thread_pool
-            .borrow()
             .current_num_threads()
             .min(num_nodes / node_granularity)
             .max(1);
@@ -126,7 +126,7 @@ pub trait SequentialLabeling {
 
         // create a channel to receive the result
         let (tx, rx) = std::sync::mpsc::channel();
-        thread_pool.borrow().in_place_scope(|scope| {
+        thread_pool.in_place_scope(|scope| {
             for _ in 0..num_scoped_threads {
                 // create some references so that we can share them across threads
                 let pl_lock = &pl_lock;
@@ -187,7 +187,7 @@ pub trait SequentialLabeling {
         fold: R,
         arc_granularity: usize,
         deg_cumul: &(impl Succ<Input = usize, Output = usize> + Send + Sync),
-        thread_pool: impl Borrow<rayon::ThreadPool>,
+        thread_pool: &ThreadPool,
         pl: Option<&mut ProgressLogger>,
     ) -> A
     where
@@ -196,7 +196,6 @@ pub trait SequentialLabeling {
         T: Send,
         A: Default + Send,
     {
-        let thread_pool = thread_pool.borrow();
         let pl_lock = pl.map(std::sync::Mutex::new);
         let num_nodes = self.num_nodes();
         let num_arcs = self.num_arcs_hint().unwrap();

src/transform/perm.rs

@@ -4,14 +4,13 @@
  * SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later
  */
 
-use std::borrow::Borrow;
-
 use crate::graphs::arc_list_graph;
 use crate::prelude::sort_pairs::{BatchIterator, KMergeIters};
 use crate::prelude::*;
 use anyhow::{ensure, Context, Result};
 use dsi_progress_logger::prelude::*;
 use lender::*;
+use rayon::ThreadPool;
 use sux::traits::BitFieldSlice;
 use tempfile::Builder;
 
@@ -75,7 +74,7 @@ pub fn permute_split<S, P>(
     graph: &S,
     perm: &P,
     batch_size: usize,
-    threads: impl Borrow<rayon::ThreadPool>,
+    threads: &ThreadPool,
 ) -> Result<Left<arc_list_graph::ArcListGraph<KMergeIters<BatchIterator<()>, ()>>>>
 where
     S: SequentialGraph + SplitLabeling,
@@ -90,11 +89,10 @@ where
     // get a premuted view
     let pgraph = PermutedGraph { graph, perm };
 
-    let pool = threads.borrow();
-    let num_threads = pool.current_num_threads();
+    let num_threads = threads.current_num_threads();
     let mut dirs = vec![];
 
-    let edges = pool.in_place_scope(|scope| {
+    let edges = threads.in_place_scope(|scope| {
         let (tx, rx) = std::sync::mpsc::channel();
 
         for (thread_id, iter) in pgraph.split_iter(num_threads).enumerate() {

src/transform/simplify.rs

@@ -4,8 +4,6 @@
  * SPDX-License-Identifier: Apache-2.0 OR LGPL-2.1-or-later
  */
 
-use std::borrow::Borrow;
-
 use crate::graphs::{
     arc_list_graph, no_selfloops_graph::NoSelfLoopsGraph, union_graph::UnionGraph,
 };
@@ -16,6 +14,7 @@ use anyhow::{Context, Result};
 use dsi_progress_logger::prelude::*;
 use itertools::{Dedup, Itertools};
 use lender::*;
+use rayon::ThreadPool;
 use tempfile::Builder;
 
 use super::transpose;
@@ -111,18 +110,17 @@ pub fn simplify(
 pub fn simplify_split<S>(
     graph: &S,
     batch_size: usize,
-    threads: impl Borrow<rayon::ThreadPool>,
+    threads: &ThreadPool,
 ) -> Result<Left<arc_list_graph::ArcListGraph<itertools::Dedup<KMergeIters<BatchIterator<()>, ()>>>>>
 where
     S: SequentialGraph + SplitLabeling,
 {
-    let pool = threads.borrow();
-    let num_threads = pool.current_num_threads();
+    let num_threads = threads.current_num_threads();
     let (tx, rx) = std::sync::mpsc::channel();
 
     let mut dirs = vec![];
 
-    pool.in_place_scope(|scope| {
+    threads.in_place_scope(|scope| {
         let mut thread_id = 0;
         #[allow(clippy::explicit_counter_loop)] // enumerate requires some extra bounds here
         for iter in graph.split_iter(num_threads) {

tests/test_par_bvcomp.rs

@@ -49,7 +49,7 @@ fn _test_par_bvcomp(basename: &str) -> Result<()> {
             &tmp_basename,
             &graph,
             comp_flags,
-            rayon::ThreadPoolBuilder::new()
+            &rayon::ThreadPoolBuilder::new()
                 .num_threads(thread_num)
                 .build()
                 .expect("Failed to create thread pool"),