try_openai_tts.py

try_openai_tts.py

import os
from io import BytesIO

import ipywidgets as widgets
from openai import OpenAI

os.environ["OPENAI_API_KEY"] = "sk-..."  # Your OpenAI API Key here.

client = OpenAI()

response = client.audio.speech.create(
    model="tts-1",
    input="こんにちは",
    voice="alloy",
)

class OpenAITTS:
    def __init__(self):
        self.client = OpenAI()

        self.text = "こんにちは！"
        self.voice = "alloy"
        self.model = "tts-1"
        self.speed = 1.0
        self.audio = BytesIO().read()

        self.text_input = widgets.Textarea(
            value=self.text,
            description="Text:",
            layout=widgets.Layout(height='100px'),
        )
        self.text_input.observe(self.on_text_input_change, names="value")

        self.voice_selector = widgets.Dropdown(
            options=[
                ("alloy", "alloy"),
                ("echo", "echo"),
                ("fable", "fable"),
                ("onyx", "onyx"),
                ("nova", "nova"),
                ("shimmer", "shimmer"),
            ],
            value=self.voice,
            description="Voice:",
        )
        self.voice_selector.observe(self.on_voice_selector_change, names="value")

        self.model_selector = widgets.Dropdown(
            options=[
                ("tts-1", "tts-1"),
                ("tts-1-hd", "tts-1-hd"),
            ],
            value=self.model,
            description="Model:",
        )
        self.model_selector.observe(self.on_model_selector_change, names="value")

        self.speed_slider = widgets.FloatSlider(
            value=self.speed,
            min=0.25,
            max=4.0,
            step=0.05,
            description="Speed:",
        )
        self.speed_slider.observe(self.on_speed_slider_change, names="value")

        self.generate_button = widgets.Button(description="Generate")
        self.generate_button.on_click(self.on_generate_button_clicked)

        self.audio_player = widgets.Audio(
            value=self.audio,
            loop=False,
            autoplay=False,
        )

    def on_text_input_change(self, change):
        self.text = change["new"]

    def on_voice_selector_change(self, change):
        self.voice = change["new"]

    def on_model_selector_change(self, change):
        self.model = change["new"]

    def on_speed_slider_change(self, change):
        self.speed = change["new"]

    def on_generate_button_clicked(self, b):
        global client, params
        response = self.client.audio.speech.create(
            model=self.model,
            input=self.text,
            voice=self.voice,
            speed=self.speed,
        )
        self.audio = BytesIO(response.content)
        self.audio.seek(0)
        self.audio_player.value = self.audio.read()

    def display(self):
        display(
            widgets.VBox(
                [
                    self.text_input,
                    self.voice_selector,
                    self.model_selector,
                    self.speed_slider,
                    self.generate_button,
                    self.audio_player,
                ],
                layout=widgets.Layout(
                    display="flex",
                    flex_flow="column",
                    align_items="center",
                    width="100%",
                )
            )
        )

tts = OpenAITTS()
tts.display()