Dump output seglst file for speaker diarization eval (#123)

* Add speaker diarization to streaming ASR clients (#116)

* enable speaker diarization for streaming_asr_client

* add: speaker diarization to transcribe_file

* fix: print confidence when word offsets are disabled

* Fix output format of transcribe_file.py client

* update: print_streaming to write seglst file for diarization