Add C/CXX/JavaScript API for NeMo Canary models (#2357)

This PR introduces support for NeMo Canary models across C, C++, and JavaScript APIs 
by adding new Canary configuration structures, updating bindings, extending examples,
and enhancing CI workflows.

- Add OfflineCanaryModelConfig to all language bindings (C, C++, JS, ETS).
- Implement SetConfig methods and NAPI wrappers for updating recognizer config at runtime.
- Update examples and CI scripts to demonstrate and test NeMo Canary model usage.

nodejs-addon-examples/README.md

@@ -123,6 +123,7 @@ The following tables list the examples in this folder.
 |[./test_asr_non_streaming_moonshine.js](./test_asr_non_streaming_moonshine.js)|Non-streaming speech recognition from a file using [Moonshine](https://github.com/usefulsensors/moonshine)|
 |[./test_vad_with_non_streaming_asr_moonshine.js](./test_vad_with_non_streaming_asr_moonshine.js)| Non-streaming speech recognition from a file using [Moonshine](https://github.com/usefulsensors/moonshine) + [Silero VAD](https://github.com/snakers4/silero-vad)|
 |[./test_asr_non_streaming_nemo_ctc.js](./test_asr_non_streaming_nemo_ctc.js)|Non-streaming speech recognition from a file using a [NeMo](https://github.com/NVIDIA/NeMo) CTC model with greedy search|
+|[./test_asr_non_streaming_nemo_canary.js](./test_asr_non_streaming_nemo_canary.js)|Non-streaming speech recognition from a file using a [NeMo](https://github.com/NVIDIA/NeMo) [Canary](https://k2-fsa.github.io/sherpa/onnx/nemo/canary.html#sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8-english-spanish-german-french) model|
 |[./test_asr_non_streaming_zipformer_ctc.js](./test_asr_non_streaming_zipformer_ctc.js)|Non-streaming speech recognition from a file using a Zipformer CTC model with greedy search|
 |[./test_asr_non_streaming_nemo_parakeet_tdt_v2.js](./test_asr_non_streaming_nemo_parakeet_tdt_v2.js)|Non-streaming speech recognition from a file using a [NeMo](https://github.com/NVIDIA/NeMo) [parakeet-tdt-0.6b-v2](https://k2-fsa.github.io/sherpa/onnx/pretrained_models/offline-transducer/nemo-transducer-models.html#sherpa-onnx-nemo-parakeet-tdt-0-6b-v2-int8-english) model with greedy search|
 |[./test_asr_non_streaming_dolphin_ctc.js](./test_asr_non_streaming_dolphin_ctc.js)|Non-streaming speech recognition from a file using a [Dolphinhttps://github.com/DataoceanAI/Dolphin]) CTC model with greedy search|
@@ -389,6 +390,16 @@ npm install naudiodon2
 node ./test_vad_asr_non_streaming_zipformer_ctc_microphone.js
 ```
 
+### Non-streaming speech recognition with NeMo Canary models
+
+```bash
+wget https://github.com/k2-fsa/sherpa-onnx/releases/download/asr-models/sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8.tar.bz2
+tar xvf sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8.tar.bz2
+rm sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8.tar.bz2
+
+node ./test_asr_non_streaming_nemo_canary.js
+```
+
 ### Non-streaming speech recognition with NeMo CTC models
 
 ```bash

nodejs-addon-examples/test_asr_non_streaming_nemo_canary.js

@@ -0,0 +1,62 @@
+// Copyright (c)  2024  Xiaomi Corporation
+const sherpa_onnx = require('sherpa-onnx-node');
+
+// Please download test files from
+// https://github.com/k2-fsa/sherpa-onnx/releases/tag/asr-models
+const config = {
+  'featConfig': {
+    'sampleRate': 16000,
+    'featureDim': 80,
+  },
+  'modelConfig': {
+    'canary': {
+      'encoder':
+          './sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8/encoder.int8.onnx',
+      'decoder':
+          './sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8/decoder.int8.onnx',
+      'srcLang': 'en',
+      'tgtLang': 'en',
+      'usePnc': 1,
+    },
+    'tokens':
+        './sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8/tokens.txt',
+    'numThreads': 2,
+    'provider': 'cpu',
+    'debug': 0,
+  }
+};
+
+const waveFilename =
+    './sherpa-onnx-nemo-canary-180m-flash-en-es-de-fr-int8/test_wavs/en.wav';
+
+const recognizer = new sherpa_onnx.OfflineRecognizer(config);
+console.log('Started')
+let start = Date.now();
+let stream = recognizer.createStream();
+const wave = sherpa_onnx.readWave(waveFilename);
+stream.acceptWaveform({sampleRate: wave.sampleRate, samples: wave.samples});
+
+recognizer.decode(stream);
+let result = recognizer.getResult(stream)
+let stop = Date.now();
+console.log('Done')
+
+const elapsed_seconds = (stop - start) / 1000;
+const duration = wave.samples.length / wave.sampleRate;
+const real_time_factor = elapsed_seconds / duration;
+console.log('Wave duration', duration.toFixed(3), 'seconds')
+console.log('Elapsed', elapsed_seconds.toFixed(3), 'seconds')
+console.log(
+    `RTF = ${elapsed_seconds.toFixed(3)}/${duration.toFixed(3)} =`,
+    real_time_factor.toFixed(3))
+console.log(waveFilename)
+console.log('result (English)\n', result)
+
+stream = recognizer.createStream();
+stream.acceptWaveform({sampleRate: wave.sampleRate, samples: wave.samples});
+recognizer.config.modelConfig.canary.tgtLang = 'de';
+recognizer.setConfig(recognizer.config);
+
+recognizer.decode(stream);
+result = recognizer.getResult(stream)
+console.log('result (German)\n', result)