Add JavaScript API (WebAssembly) for Google MedAsr model (#2954)

.github/scripts/test-nodejs-npm.sh

@@ -9,6 +9,14 @@ git status
 ls -lh
 ls -lh node_modules
 
+curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+tar xvf sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+rm sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+
+node ./test-offline-medasr-ctc.js
+
+rm -rf sherpa-onnx-medasr-ctc-en-int8-2025-12-25
+
 curl -SL -O https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-omnilingual-asr-1600-languages-300M-ctc-int8-2025-11-12.tar.bz2
 tar xvf sherpa-onnx-omnilingual-asr-1600-languages-300M-ctc-int8-2025-11-12.tar.bz2
 rm sherpa-onnx-omnilingual-asr-1600-languages-300M-ctc-int8-2025-11-12.tar.bz2

nodejs-examples/README.md

@@ -203,6 +203,22 @@ rm sherpa-onnx-zipformer-ctc-zh-int8-2025-07-03.tar.bz2
 node ./test-offline-zipformer-ctc.js
 ```
 
+## ./test-offline-medasr-ctc.js
+
+[./test-offline-medasr-ctc.js](./test-offline-medasr-ctc.js) demonstrates
+how to decode a file with a Google MedASR CTC model. In the code we use
+[sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2](https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2).
+
+You can use the following command to run it:
+
+```bash
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+tar xvf sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+rm sherpa-onnx-medasr-ctc-en-int8-2025-12-25.tar.bz2
+
+node ./test-offline-medasr-ctc.js
+```
+
 ## ./test-offline-omnilingual-asr-ctc.js
 
 [./test-offline-omnilingual-asr-ctc.js](./test-offline-omnilingual-asr-ctc.js) demonstrates

nodejs-examples/test-offline-medasr-ctc.js

@@ -0,0 +1,35 @@
+// Copyright (c)  2025  Xiaomi Corporation (authors: Fangjun Kuang)
+//
+const fs = require('fs');
+const {Readable} = require('stream');
+const wav = require('wav');
+
+const sherpa_onnx = require('sherpa-onnx');
+
+function createOfflineRecognizer() {
+  let config = {
+    modelConfig: {
+      medasr: {
+        model: './sherpa-onnx-medasr-ctc-en-int8-2025-12-25/model.int8.onnx',
+      },
+      tokens: './sherpa-onnx-medasr-ctc-en-int8-2025-12-25/tokens.txt',
+    }
+  };
+
+  return sherpa_onnx.createOfflineRecognizer(config);
+}
+
+const recognizer = createOfflineRecognizer();
+const stream = recognizer.createStream();
+
+const waveFilename =
+    './sherpa-onnx-medasr-ctc-en-int8-2025-12-25/test_wavs/0.wav';
+const wave = sherpa_onnx.readWave(waveFilename);
+stream.acceptWaveform(wave.sampleRate, wave.samples);
+
+recognizer.decode(stream);
+const text = recognizer.getResult(stream).text;
+console.log(text);
+
+stream.free();
+recognizer.free();

nodejs-examples/test-offline-tts-matcha-en.js

@@ -6,7 +6,6 @@ function createOfflineTts() {
   let offlineTtsMatchaModelConfig = {
     acousticModel: './matcha-icefall-en_US-ljspeech/model-steps-3.onnx',
     vocoder: './vocos-22khz-univ.onnx',
-    lexicon: './matcha-icefall-en_US-ljspeech/lexicon.txt',
     tokens: './matcha-icefall-en_US-ljspeech/tokens.txt',
     dataDir: './matcha-icefall-en_US-ljspeech/espeak-ng-data',
 

wasm/asr/sherpa-onnx-asr.js

@@ -59,6 +59,10 @@ function freeConfig(config, Module) {
     freeConfig(config.omnilingual, Module)
   }
 
+  if ('medasr' in config) {
+    freeConfig(config.medasr, Module)
+  }
+
   if ('moonshine' in config) {
     freeConfig(config.moonshine, Module)
   }
@@ -776,6 +780,23 @@ function initSherpaOnnxOfflineOmnilingualAsrCtcModelConfig(config, Module) {
   }
 }
 
+function initSherpaOnnxOfflineMedAsrCtcModelConfig(config, Module) {
+  const n = Module.lengthBytesUTF8(config.model || '') + 1;
+
+  const buffer = Module._malloc(n);
+
+  const len = 1 * 4;  // 1 pointer
+  const ptr = Module._malloc(len);
+
+  Module.stringToUTF8(config.model || '', buffer, n);
+
+  Module.setValue(ptr, buffer, 'i8*');
+
+  return {
+    buffer: buffer, ptr: ptr, len: len,
+  }
+}
+
 function initSherpaOnnxOfflineWhisperModelConfig(config, Module) {
   const encoderLen = Module.lengthBytesUTF8(config.encoder || '') + 1;
   const decoderLen = Module.lengthBytesUTF8(config.decoder || '') + 1;
@@ -1052,6 +1073,12 @@ function initSherpaOnnxOfflineModelConfig(config, Module) {
     };
   }
 
+  if (!('medasr' in config)) {
+    config.medasr = {
+      model: '',
+    };
+  }
+
   if (!('whisper' in config)) {
     config.whisper = {
       encoder: '',
@@ -1139,10 +1166,13 @@ function initSherpaOnnxOfflineModelConfig(config, Module) {
   const omnilingual = initSherpaOnnxOfflineOmnilingualAsrCtcModelConfig(
       config.omnilingual, Module);
 
+  const medasr =
+      initSherpaOnnxOfflineMedAsrCtcModelConfig(config.medasr, Module);
+
   const len = transducer.len + paraformer.len + nemoCtc.len + whisper.len +
       tdnn.len + 8 * 4 + senseVoice.len + moonshine.len + fireRedAsr.len +
       dolphin.len + zipformerCtc.len + canary.len + wenetCtc.len +
-      omnilingual.len;
+      omnilingual.len + medasr.len;
 
   const ptr = Module._malloc(len);
 
@@ -1256,12 +1286,15 @@ function initSherpaOnnxOfflineModelConfig(config, Module) {
   Module._CopyHeap(omnilingual.ptr, omnilingual.len, ptr + offset);
   offset += omnilingual.len;
 
+  Module._CopyHeap(medasr.ptr, medasr.len, ptr + offset);
+  offset += medasr.len;
+
   return {
     buffer: buffer, ptr: ptr, len: len, transducer: transducer,
         paraformer: paraformer, nemoCtc: nemoCtc, whisper: whisper, tdnn: tdnn,
         senseVoice: senseVoice, moonshine: moonshine, fireRedAsr: fireRedAsr,
         dolphin: dolphin, zipformerCtc: zipformerCtc, canary: canary,
-        wenetCtc: wenetCtc, omnilingual: omnilingual
+        wenetCtc: wenetCtc, omnilingual: omnilingual, medasr: medasr
   }
 }
 

wasm/nodejs/sherpa-onnx-wasm-nodejs.cc

@@ -16,6 +16,7 @@ static_assert(sizeof(SherpaOnnxOfflineParaformerModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineZipformerCtcModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineWenetCtcModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineOmnilingualAsrCtcModelConfig) == 4, "");
+static_assert(sizeof(SherpaOnnxOfflineMedAsrCtcModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineDolphinModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineNemoEncDecCtcModelConfig) == 4, "");
 static_assert(sizeof(SherpaOnnxOfflineWhisperModelConfig) == 5 * 4, "");
@@ -39,7 +40,8 @@ static_assert(sizeof(SherpaOnnxOfflineModelConfig) ==
                       sizeof(SherpaOnnxOfflineZipformerCtcModelConfig) +
                       sizeof(SherpaOnnxOfflineCanaryModelConfig) +
                       sizeof(SherpaOnnxOfflineWenetCtcModelConfig) +
-                      sizeof(SherpaOnnxOfflineOmnilingualAsrCtcModelConfig),
+                      sizeof(SherpaOnnxOfflineOmnilingualAsrCtcModelConfig) +
+                      sizeof(SherpaOnnxOfflineMedAsrCtcModelConfig),
 
               "");
 static_assert(sizeof(SherpaOnnxFeatureConfig) == 2 * 4, "");
@@ -89,6 +91,7 @@ void PrintOfflineRecognizerConfig(SherpaOnnxOfflineRecognizerConfig *config) {
   auto canary = &model_config->canary;
   auto wenet_ctc = &model_config->wenet_ctc;
   auto omnilingual = &model_config->omnilingual;
+  auto medasr = &model_config->medasr;
 
   fprintf(stdout, "----------offline transducer model config----------\n");
   fprintf(stdout, "encoder: %s\n", transducer->encoder);
@@ -145,6 +148,9 @@ void PrintOfflineRecognizerConfig(SherpaOnnxOfflineRecognizerConfig *config) {
   fprintf(stdout, "----------offline Omnilingual ASR model config----------\n");
   fprintf(stdout, "model: %s\n", omnilingual->model);
 
+  fprintf(stdout, "----------offline MedASR model config----------\n");
+  fprintf(stdout, "model: %s\n", medasr->model);
+
   fprintf(stdout, "tokens: %s\n", model_config->tokens);
   fprintf(stdout, "num_threads: %d\n", model_config->num_threads);
   fprintf(stdout, "provider: %s\n", model_config->provider);