fix: mixed Chinese and non-Chinese words

Brooooooklyn · Nov 3, 2020 · 40be9e7 · 40be9e7 · github-actions · Nov 3, 2020
1 parent a0c40e1
commit 40be9e7
Show file tree

Hide file tree

Showing 8 changed files with 384 additions and 125 deletions.
diff --git a/.github/workflows/CI.yml b/.github/workflows/CI.yml
@@ -82,6 +82,9 @@ jobs:
       - name: 'Install dependencies'
         run: yarn install --frozen-lockfile --registry https://registry.npmjs.org
 
+      - name: 'Cargo test'
+        run: cargo test
+
       - name: 'Build'
         if: matrix.os != 'macos-latest'
         run: yarn build

diff --git a/Cargo.toml b/Cargo.toml
@@ -15,6 +15,7 @@ napi = {git = "https://github.com/napi-rs/napi-rs"}
 napi-derive = {git = "https://github.com/napi-rs/napi-rs"}
 once_cell = "1.4"
 pinyin = "0.8"
+rayon = "1.5"
 
 [target.'cfg(all(unix, not(target_env = "musl")))'.dependencies]
 jemallocator = {version = "0.3", features = ["disable_initial_exec_tls"]}

diff --git a/README.md b/README.md
@@ -42,16 +42,14 @@ yarn add @napi-rs/pinyin
 Benchmark over `pinyin` package:
 
 ```bash
-yarn run v1.22.10
-$ node -r @swc-node/register benchmark/bench.ts
 Running "Short input without segment" suite...
 Progress: 100%
 
   @napi-rs/pinyin:
-    1 037 055 ops/s, ±0.62%   | fastest
+    962 035 ops/s, ±0.68%   | fastest
 
   node-pinyin:
-    461 335 ops/s, ±0.87%     | slowest, 55.51% slower
+    434 241 ops/s, ±0.66%   | slowest, 54.86% slower
 
 Finished 2 cases!
   Fastest: @napi-rs/pinyin
@@ -60,10 +58,10 @@ Running "Long input without segment" suite...
 Progress: 100%
 
   @napi-rs/pinyin:
-    325 ops/s, ±0.54%   | fastest
+    59 ops/s, ±0.83%   | fastest
 
   node-pinyin:
-    2 ops/s, ±3.44%     | slowest, 99.38% slower
+    2 ops/s, ±3.30%    | slowest, 96.61% slower
 
 Finished 2 cases!
   Fastest: @napi-rs/pinyin
@@ -72,10 +70,10 @@ Running "Short input with segment" suite...
 Progress: 100%
 
   @napi-rs/pinyin:
-    1 010 506 ops/s, ±2.59%   | fastest
+    530 228 ops/s, ±1.94%   | fastest
 
   node-pinyin:
-    315 076 ops/s, ±1.02%     | slowest, 68.82% slower
+    307 788 ops/s, ±0.83%   | slowest, 41.95% slower
 
 Finished 2 cases!
   Fastest: @napi-rs/pinyin
@@ -84,15 +82,15 @@ Running "Long input with segment" suite...
 Progress: 100%
 
   @napi-rs/pinyin:
-    330 ops/s, ±1.08%   | fastest
+    152 ops/s, ±1.09%   | fastest
 
   node-pinyin:
-    3 ops/s, ±2.44%     | slowest, 99.09% slower
+    3 ops/s, ±3.08%     | slowest, 98.03% slower
 
 Finished 2 cases!
   Fastest: @napi-rs/pinyin
   Slowest: node-pinyin
-✨  Done in 52.94s.
+✨  Done in 53.36s.
 ```
 
 ## 用法

diff --git a/__test__/index.spec.ts b/__test__/index.spec.ts
@@ -67,6 +67,19 @@ for (const fixture of fixtures) {
   }
 }
 
+test('mixed with segment', (t) => {
+  t.deepEqual(pinyin('特殊天-1', { style: PINYIN_STYLE.Plain, segment: true }), ['te', 'shu', 'tian', '-1'])
+})
+
+test('mixed with multi segment', (t) => {
+  t.deepEqual(pinyin('特殊天-1', { style: PINYIN_STYLE.Plain, segment: true, heteronym: true }), [
+    ['te'],
+    ['shu'],
+    ['tian'],
+    ['-1'],
+  ])
+})
+
 test('我,要,排,序 => 序,我,排,要', (t) => {
   const data = '我要排序'.split('')
   const sortedData = data.sort(compare)
Benchmark suite	Current: `40be9e7`	Previous: `b21547b`	Ratio
`Short input without segment#@napi-rs/pinyin`	`577606` ops/sec (`±0.91%`)	`488927` ops/sec (`±1.11%`)	`0.85`
`Short input without segment#node-pinyin`	`320882` ops/sec (`±0.77%`)	`268041` ops/sec (`±1.21%`)	`0.84`
`Long input without segment#@napi-rs/pinyin`	`45` ops/sec (`±1.48%`)	`174` ops/sec (`±1.03%`)	`3.87`
`Long input without segment#node-pinyin`	`1` ops/sec (`±1.74%`)	`1` ops/sec (`±7.73%`)	`1`
`Short input with segment#@napi-rs/pinyin`	`350149` ops/sec (`±1.51%`)	`501747` ops/sec (`±1.2%`)	`1.43`
`Short input with segment#node-pinyin`	`210479` ops/sec (`±1%`)	`180016` ops/sec (`±1.46%`)	`0.86`
`Long input with segment#@napi-rs/pinyin`	`99` ops/sec (`±0.95%`)	`169` ops/sec (`±1.84%`)	`1.71`
`Long input with segment#node-pinyin`	`2` ops/sec (`±1.62%`)	`2` ops/sec (`±4.25%`)	`1`