simulstream.server.message_processor

Classes

MessageProcessor(client_id, speech_processor)

This class is responsible for processing the messages incoming from a client, which include control messages (e.g., configurations about languages to use, or signal that the stream is over).

class simulstream.server.message_processor.MessageProcessor(client_id: int, speech_processor: SpeechProcessor)

This class is responsible for processing the messages incoming from a client, which include control messages (e.g., configurations about languages to use, or signal that the stream is over).

clear()

Clear the internal states to be ready for a new input stream.

end_of_stream() → IncrementalOutput

Performs the last operations to conclude the processing of the stream of audio by the speech processor and cleans up everything to be ready for the next stream.

Returns:

last output at the end of the stream.

Return type:

IncrementalOutput

process_metadata(metadata: dict)

Takes a dictionary of metadata regarding the incoming speech and desired output, and interacts with the configured speech_processor to set it up accordingly.

Parameters:

metadata (dict) – Dictionary of metadata regarding the incoming speech.

process_speech(speech_data: bytes) → IncrementalOutput | None

Process an audio chunk and return incremental transcription/translation. Namely, it receives and buffers raw audio chunks (bytes) and processes audio audio incrementally with the configured SpeechProcessor.

Parameters:

speech_data (bytes) – Raw PCM audio bytes (16-bit little endian).

Returns:

incremental processing results, if any. None otherwise.

Return type:

IncrementalOutput