Update http_server.py with the option of saving the generated speech to file, rather than streaming it back.

By adding a file argument to both the POST and GET requests, the synthesized audio can be stored to the requested filename. Maybe a startup audio_path argument can passed to http_server.py to provide a default location for the audio files, to both control location and not force the file argument to include path information.

Using the changes described above, I have modified http_server.py locally to process incoming requests and store the audio files in a common folder. Once they are available, another process picks up and plays back the audio files.