added method 'recognize_whispercpp' to support whisper.cpp

[eliranwong]

added method 'recognize_whispercpp' to support whisper.cpp as backend for recognizing speech

        Adapted from code: https://github.com/eliranwong/freegenius/blob/96d2fd7751ca26f2c7adaa63082a3cb79681f3ed/package/freegenius/utils/prompts.py#L118

        Performs speech recognition on ``audio_data`` (an ``AudioData`` instance), using Whisper.

        ``whispercpp_main`` is the local path of the main file of whisper.cpp, it depends on how users set up their local copies of whisper.cpp

        e.g., with the following setup, set '~/whisper.cpp/main' as ``whispercpp_main``:

        > cd ~

        > git clone https://github.com/ggerganov/whisper.cpp.git

        > cd whisper.cpp

        > make

        ``model_path`` is the local file path of any of *.bin files downloaded from https://huggingface.co/ggerganov/whisper.cpp/tree/main.

        e.g. download 'ggml-large-v3-q5_0.bin' to home directory, then ``model_path`` is '~/ggml-large-v3-q5_0.bin'

        The recognition language is determined by ``language``, an uncapitalized language code like "en" or "zh". 'auto' for auto-detect. See the full language list at https://github.com/openai/whisper/blob/main/whisper/tokenizer.py

        e.g. set 'en' as ``language`` for English

        e.g. set 'auto' as ``language`` for non-English languages

        ``additional_options`` are additional options that are passed directly to whisper.cpp. See https://github.com/ggerganov/whisper.cpp/tree/master/examples/main for all options

        e.g. set '-t 12' as ``additional_options``, to use 12 threads during computation

        e.g. set '-tr' as ``additional_options``, to translate from the speech to english