src-tauri/examples/asr_bench.rs

@@ -84,7 +84,13 @@ mod cputime {
 fn normalize(s: &str) -> Vec<String> {
     s.to_lowercase()
         .chars()
-        .map(|c| if c.is_alphanumeric() || c == '\'' { c } else { ' ' })
+        .map(|c| {
+            if c.is_alphanumeric() || c == '\'' {
+                c
+            } else {
+                ' '
+            }
+        })
         .collect::<String>()
         .split_whitespace()
         .map(str::to_string)
@@ -121,6 +127,8 @@ fn self_test() {
 
 // ---- engines ----
 
+// ponytail: one Engine per process for its whole lifetime — variant size skew is irrelevant.
+#[allow(clippy::large_enum_variant)]
 enum Engine {
     Whisper(WhisperTranscriber),
     Onnx(OnnxTranscriber),
@@ -282,12 +290,15 @@ fn run_corpus(contender: &str, engine: &Engine, load_ms: u128, corpus: &Path, ou
         (0usize, 0usize, 0u128, 0u64, 0f64);
     for wav in &wavs {
         let stem = wav.file_stem().unwrap().to_string_lossy().into_owned();
-        let reference =
+        let reference = std::fs::read_to_string(wav.with_extension("txt")).unwrap_or_else(|e| {
-            std::fs::read_to_string(wav.with_extension("txt")).unwrap_or_else(|e| {
+            panic!(
-                panic!("missing reference {}: {e}", wav.with_extension("txt").display())
+                "missing reference {}: {e}",
+                wav.with_extension("txt").display()
+            )
         });
-        let audio_secs =
+        let audio_secs = read_wav_mono_16k(wav)
-            read_wav_mono_16k(wav).map(|s| s.len() as f64 / SAMPLE_RATE as f64).unwrap_or(0.0);
+            .map(|s| s.len() as f64 / SAMPLE_RATE as f64)
+            .unwrap_or(0.0);
 
         let cpu0 = cputime::process_cpu_ms();
         let t0 = Instant::now();
@@ -324,9 +335,18 @@ fn run_corpus(contender: &str, engine: &Engine, load_ms: u128, corpus: &Path, ou
     }
 
     println!("---- {contender} summary ----");
-    println!("files: {}  audio: {audio_total:.1}s  load: {load_ms}ms", wavs.len());
+    println!(
-    println!("WER: {:.2}%  ({errors}/{words})", 100.0 * errors as f64 / words.max(1) as f64);
+        "files: {}  audio: {audio_total:.1}s  load: {load_ms}ms",
-    println!("RTF (wall): {:.3}", wall_total as f64 / 1000.0 / audio_total.max(0.001));
+        wavs.len()
+    );
+    println!(
+        "WER: {:.2}%  ({errors}/{words})",
+        100.0 * errors as f64 / words.max(1) as f64
+    );
+    println!(
+        "RTF (wall): {:.3}",
+        wall_total as f64 / 1000.0 / audio_total.max(0.001)
+    );
     println!(
         "CPU-sec per audio-sec: {:.3}",
         cpu_total as f64 / 1000.0 / audio_total.max(0.001)
@@ -349,7 +369,10 @@ fn run_live(contender: &str, engine: &Engine, wav: &Path) {
     }
     latencies.sort_unstable();
     let pct = |p: f64| latencies[(((latencies.len() - 1) as f64) * p) as usize];
-    println!("---- {contender} live ({WINDOW_SECS}s windows, n={}) ----", latencies.len());
+    println!(
+        "---- {contender} live ({WINDOW_SECS}s windows, n={}) ----",
+        latencies.len()
+    );
     println!(
         "p50={}ms p95={}ms max={}ms  (budget: {}ms)",
         pct(0.50),
@@ -365,7 +388,9 @@ fn main() {
         self_test();
         return;
     }
-    let contender = args.first().expect("usage: asr_bench <contender> [--corpus DIR|--live WAV]");
+    let contender = args
+        .first()
+        .expect("usage: asr_bench <contender> [--corpus DIR|--live WAV]");
     let flag = |name: &str| {
         args.iter()
             .position(|a| a == name)