openai/openai-python
Publicmirrored from https://github.com/openai/openai-pythonAvailable
examples/audio.py
64lines · modeblame
9375d2ccStainless Bot2 years ago | 1 | #!/usr/bin/env rye run python |
e9c26d1eStainless Bot2 years ago | 2 | |
9375d2ccStainless Bot2 years ago | 3 | import time |
e9c26d1eStainless Bot2 years ago | 4 | from pathlib import Path |
| 5 | | |
| 6 | from openai import OpenAI | |
| 7 | | |
| 8 | # gets OPENAI_API_KEY from your environment variables | |
| 9 | openai = OpenAI() | |
| 10 | | |
| 11 | speech_file_path = Path(__file__).parent / "speech.mp3" | |
| 12 | | |
| 13 | | |
| 14 | def main() -> None: | |
9375d2ccStainless Bot2 years ago | 15 | stream_to_speakers() |
| 16 | | |
e9c26d1eStainless Bot2 years ago | 17 | # Create text-to-speech audio file |
86379b44Stainless Bot2 years ago | 18 | with openai.audio.speech.with_streaming_response.create( |
| 19 | model="tts-1", | |
| 20 | voice="alloy", | |
| 21 | input="the quick brown fox jumped over the lazy dogs", | |
| 22 | ) as response: | |
| 23 | response.stream_to_file(speech_file_path) | |
e9c26d1eStainless Bot2 years ago | 24 | |
| 25 | # Create transcription from audio file | |
86379b44Stainless Bot2 years ago | 26 | transcription = openai.audio.transcriptions.create( |
| 27 | model="whisper-1", | |
| 28 | file=speech_file_path, | |
| 29 | ) | |
e9c26d1eStainless Bot2 years ago | 30 | print(transcription.text) |
| 31 | | |
| 32 | # Create translation from audio file | |
| 33 | translation = openai.audio.translations.create( | |
| 34 | model="whisper-1", | |
| 35 | file=speech_file_path, | |
| 36 | ) | |
| 37 | print(translation.text) | |
| 38 | | |
| 39 | | |
9375d2ccStainless Bot2 years ago | 40 | def stream_to_speakers() -> None: |
| 41 | import pyaudio | |
| 42 | | |
| 43 | player_stream = pyaudio.PyAudio().open(format=pyaudio.paInt16, channels=1, rate=24000, output=True) | |
| 44 | | |
| 45 | start_time = time.time() | |
| 46 | | |
| 47 | with openai.audio.speech.with_streaming_response.create( | |
| 48 | model="tts-1", | |
| 49 | voice="alloy", | |
| 50 | response_format="pcm", # similar to WAV, but without a header chunk at the start. | |
| 51 | input="""I see skies of blue and clouds of white | |
| 52 | The bright blessed days, the dark sacred nights | |
| 53 | And I think to myself | |
| 54 | What a wonderful world""", | |
| 55 | ) as response: | |
| 56 | print(f"Time to first byte: {int((time.time() - start_time) * 1000)}ms") | |
| 57 | for chunk in response.iter_bytes(chunk_size=1024): | |
| 58 | player_stream.write(chunk) | |
| 59 | | |
| 60 | print(f"Done in {int((time.time() - start_time) * 1000)}ms.") | |
| 61 | | |
| 62 | | |
e9c26d1eStainless Bot2 years ago | 63 | if __name__ == "__main__": |
| 64 | main() |