Add benchmark for Python threaded word count

Add benchmark where the sequential Rust version of the word count is
executed twice to demonstrate parallelism with Python threads.  Also
slightly simplify the benchmark functions.

examples/word-count/src/lib.rs

@@ -7,21 +7,22 @@ use rayon::prelude::*;
 
 /// Searches for the word, parallelized by rayon
 #[pyfunction]
-fn search(py: Python<'_>, contents: &str, needle: &str) -> PyResult<usize> {
-    let count = py.allow_threads(move || {
-        contents
-            .par_lines()
-            .map(|line| count_line(line, needle))
-            .sum()
-    });
-    Ok(count)
+fn search(contents: &str, needle: &str) -> usize {
+    contents
+        .par_lines()
+        .map(|line| count_line(line, needle))
+        .sum()
 }
 
 /// Searches for a word in a classic sequential fashion
 #[pyfunction]
-fn search_sequential(contents: &str, needle: &str) -> PyResult<usize> {
-    let result = contents.lines().map(|line| count_line(line, needle)).sum();
-    Ok(result)
+fn search_sequential(contents: &str, needle: &str) -> usize {
+    contents.lines().map(|line| count_line(line, needle)).sum()
+}
+
+#[pyfunction]
+fn search_sequential_allow_threads(py: Python, contents: &str, needle: &str) -> usize {
+    py.allow_threads(|| search_sequential(contents, needle))
 }
 
 fn matches(word: &str, needle: &str) -> bool {
@@ -56,6 +57,7 @@ fn count_line(line: &str, needle: &str) -> usize {
 fn word_count(_py: Python<'_>, m: &PyModule) -> PyResult<()> {
     m.add_wrapped(wrap_pyfunction!(search))?;
     m.add_wrapped(wrap_pyfunction!(search_sequential))?;
+    m.add_wrapped(wrap_pyfunction!(search_sequential_allow_threads))?;
 
     Ok(())
 }

examples/word-count/tests/test_word_count.py

@@ -1,7 +1,8 @@
-# -*- coding: utf-8 -*-
 import os
+from concurrent.futures import ThreadPoolExecutor
 
 import pytest
+
 import word_count
 
 current_dir = os.path.abspath(os.path.dirname(__file__))
@@ -49,3 +50,23 @@ def test_word_count_rust_sequential(benchmark, contents):
 def test_word_count_python_sequential(benchmark, contents):
     count = benchmark(word_count.search_py, contents, "is")
     assert count == 10000
+
+
+def run_rust_sequential_twice(
+    executor: ThreadPoolExecutor, contents: str, needle: str
+) -> int:
+    future_1 = executor.submit(
+        word_count.search_sequential_allow_threads, contents, needle
+    )
+    future_2 = executor.submit(
+        word_count.search_sequential_allow_threads, contents, needle
+    )
+    result_1 = future_1.result()
+    result_2 = future_2.result()
+    return result_1 + result_2
+
+
+def test_word_count_rust_sequential_twice_with_threads(benchmark, contents):
+    executor = ThreadPoolExecutor(max_workers=2)
+    count = benchmark(run_rust_sequential_twice, executor, contents, "is")
+    assert count == 20000

examples/word-count/word_count/__init__.py

@@ -1,6 +1,11 @@
-from .word_count import search, search_sequential
+from .word_count import search, search_sequential, search_sequential_allow_threads
 
-__all__ = ["search_py", "search", "search_sequential"]
+__all__ = [
+    "search_py",
+    "search",
+    "search_sequential",
+    "search_sequential_allow_threads",
+]
 
 
 def search_py(contents, needle):