Refactor online recognizer (#250)

* Refactor online recognizer.

Make it easier to support other streaming models.

Note that it is a breaking change for the Python API.
`sherpa_onnx.OnlineRecognizer()` used before should be
replaced by `sherpa_onnx.OnlineRecognizer.from_transducer()`.

sherpa-onnx/python/csrc/CMakeLists.txt

@@ -13,6 +13,7 @@ pybind11_add_module(_sherpa_onnx
   offline-transducer-model-config.cc
   offline-whisper-model-config.cc
   online-lm-config.cc
+  online-model-config.cc
   online-recognizer.cc
   online-stream.cc
   online-transducer-model-config.cc

sherpa-onnx/python/csrc/online-model-config.cc

@@ -0,0 +1,35 @@
+// sherpa-onnx/python/csrc/online-model-config.cc
+//
+// Copyright (c)  2023 by manyeyes
+
+#include "sherpa-onnx/python/csrc/online-model-config.h"
+
+#include <string>
+#include <vector>
+
+#include "sherpa-onnx/csrc/online-model-config.h"
+#include "sherpa-onnx/csrc/online-transducer-model-config.h"
+#include "sherpa-onnx/python/csrc/online-transducer-model-config.h"
+
+namespace sherpa_onnx {
+
+void PybindOnlineModelConfig(py::module *m) {
+  PybindOnlineTransducerModelConfig(m);
+
+  using PyClass = OnlineModelConfig;
+  py::class_<PyClass>(*m, "OnlineModelConfig")
+      .def(py::init<const OnlineTransducerModelConfig &, std::string &, int32_t,
+                    bool, const std::string &, const std::string &>(),
+           py::arg("transducer") = OnlineTransducerModelConfig(),
+           py::arg("tokens"), py::arg("num_threads"), py::arg("debug") = false,
+           py::arg("provider") = "cpu", py::arg("model_type") = "")
+      .def_readwrite("transducer", &PyClass::transducer)
+      .def_readwrite("tokens", &PyClass::tokens)
+      .def_readwrite("num_threads", &PyClass::num_threads)
+      .def_readwrite("debug", &PyClass::debug)
+      .def_readwrite("provider", &PyClass::provider)
+      .def_readwrite("model_type", &PyClass::model_type)
+      .def("__str__", &PyClass::ToString);
+}
+
+}  // namespace sherpa_onnx

sherpa-onnx/python/csrc/online-model-config.h

@@ -0,0 +1,16 @@
+// sherpa-onnx/python/csrc/online-model-config.h
+//
+// Copyright (c)  2023 by manyeyes
+
+#ifndef SHERPA_ONNX_PYTHON_CSRC_ONLINE_MODEL_CONFIG_H_
+#define SHERPA_ONNX_PYTHON_CSRC_ONLINE_MODEL_CONFIG_H_
+
+#include "sherpa-onnx/python/csrc/sherpa-onnx.h"
+
+namespace sherpa_onnx {
+
+void PybindOnlineModelConfig(py::module *m);
+
+}
+
+#endif  // SHERPA_ONNX_PYTHON_CSRC_ONLINE_MODEL_CONFIG_H_

sherpa-onnx/python/csrc/online-recognizer.cc

@@ -27,10 +27,9 @@ static void PybindOnlineRecognizerResult(py::module *m) {
 static void PybindOnlineRecognizerConfig(py::module *m) {
   using PyClass = OnlineRecognizerConfig;
   py::class_<PyClass>(*m, "OnlineRecognizerConfig")
-      .def(py::init<const FeatureExtractorConfig &,
-                    const OnlineTransducerModelConfig &, const OnlineLMConfig &,
-                    const EndpointConfig &, bool, const std::string &, int32_t,
-                    float>(),
+      .def(py::init<const FeatureExtractorConfig &, const OnlineModelConfig &,
+                    const OnlineLMConfig &, const EndpointConfig &, bool,
+                    const std::string &, int32_t, float>(),
            py::arg("feat_config"), py::arg("model_config"),
            py::arg("lm_config") = OnlineLMConfig(), py::arg("endpoint_config"),
            py::arg("enable_endpoint"), py::arg("decoding_method"),

sherpa-onnx/python/csrc/online-transducer-model-config.cc

@@ -14,20 +14,11 @@ void PybindOnlineTransducerModelConfig(py::module *m) {
   using PyClass = OnlineTransducerModelConfig;
   py::class_<PyClass>(*m, "OnlineTransducerModelConfig")
       .def(py::init<const std::string &, const std::string &,
-                    const std::string &, const std::string &, int32_t, bool,
-                    const std::string &, const std::string &>(),
-           py::arg("encoder_filename"), py::arg("decoder_filename"),
-           py::arg("joiner_filename"), py::arg("tokens"),
-           py::arg("num_threads"), py::arg("debug") = false,
-           py::arg("provider") = "cpu", py::arg("model_type") = "")
-      .def_readwrite("encoder_filename", &PyClass::encoder_filename)
-      .def_readwrite("decoder_filename", &PyClass::decoder_filename)
-      .def_readwrite("joiner_filename", &PyClass::joiner_filename)
-      .def_readwrite("tokens", &PyClass::tokens)
-      .def_readwrite("num_threads", &PyClass::num_threads)
-      .def_readwrite("debug", &PyClass::debug)
-      .def_readwrite("provider", &PyClass::provider)
-      .def_readwrite("model_type", &PyClass::model_type)
+                    const std::string &>(),
+           py::arg("encoder"), py::arg("decoder"), py::arg("joiner"))
+      .def_readwrite("encoder", &PyClass::encoder)
+      .def_readwrite("decoder", &PyClass::decoder)
+      .def_readwrite("joiner", &PyClass::joiner)
       .def("__str__", &PyClass::ToString);
 }
 

sherpa-onnx/python/csrc/sherpa-onnx.cc

@@ -12,9 +12,9 @@
 #include "sherpa-onnx/python/csrc/offline-recognizer.h"
 #include "sherpa-onnx/python/csrc/offline-stream.h"
 #include "sherpa-onnx/python/csrc/online-lm-config.h"
+#include "sherpa-onnx/python/csrc/online-model-config.h"
 #include "sherpa-onnx/python/csrc/online-recognizer.h"
 #include "sherpa-onnx/python/csrc/online-stream.h"
-#include "sherpa-onnx/python/csrc/online-transducer-model-config.h"
 
 namespace sherpa_onnx {
 
@@ -22,7 +22,7 @@ PYBIND11_MODULE(_sherpa_onnx, m) {
   m.doc() = "pybind11 binding of sherpa-onnx";
 
   PybindFeatures(&m);
-  PybindOnlineTransducerModelConfig(&m);
+  PybindOnlineModelConfig(&m);
   PybindOnlineLMConfig(&m);
   PybindOnlineStream(&m);
   PybindEndpoint(&m);