* 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()`.
* Don't reset model state and feature extractor on endpointing
* support passing decoding_method from commandline
* Add modified_beam_search to Python API
* fix C API example
* Fix style issues
