Realtimetts

- resolved an issue where feeding `str` content while still playing caused problems.
Huge shoutout to [Yokopops](https://github.com/Yokopops) for pointing me to this bug, which was completely under the radar for me.
- fixed 75 (muted flag did not work for the ElevenlabsEngine)
- adjusted `on_playback_start` callback so that it's called before the first chunk plays instead of after

- fixed 64

New parameters for play and play_async methods:
- added sentence_fragment_delimiters (A string of characters that are considered sentence delimiters. Default is ".?!;:,\n…)]}。-")
- force_first_fragment_after_words (int): The number of words after which the first sentence fragment is forced to be yielded. Default is 15 words.

Fixed versions for dependencies. Elevenlabs just upgraded to 1.0.0 which broke stream generation so we temporarily set a fixed version 0.2.27 until RealtimeTTS is adjusted, otherwise this engine would not work anymore.

- added CoquiVoice & OpenAIVoice classes to be consistent with all engines
- routing coquiengine synthesize process output back to main process stdout to see log
- added comma_silence_duration and sentence_silence_duration parameters to CoquiEngine constructor
- added set_model(self, checkpoint: str) method to CoquiEngine to allow switching of basic xtts model at runtime

- merged 22 & fixed 23 (thank you)
- change CoquiEngine cloning_reference_wav parameter to "voice"
- added coqui predefined [speaker voices](https://github.com/KoljaB/RealtimeTTS/tree/master/tests/coqui_voices)
Submit name or parts of the name as voice (case-insensitive):
python
engine = CoquiEngine(voice="aaron")

RealtimeTTS will first look for wav or json files, then for the predefined speaker voices

- support for multiple speaker cloning reference wave files of coqui engine (submit them as list of filenames as you normally would with a single filename string)
- default voice now male (female voice [here](https://github.com/KoljaB/RealtimeTTS/blob/master/tests/female.json))