Spaces:
Sleeping
Sleeping
| import os | |
| import tempfile | |
| # LLM Models | |
| from langchain_openai import ChatOpenAI | |
| from openai import OpenAI | |
| # Gradio | |
| import gradio as gr | |
| # Environment variables | |
| from dotenv import load_dotenv | |
| os.environ['USER_AGENT'] = 'JimYin88' | |
| load_dotenv() | |
| chat_model = ChatOpenAI(model="gpt-4o-mini-2024-07-18", | |
| max_completion_tokens=1024, | |
| api_key=os.getenv("OPENAI_API_KEY"), | |
| temperature=0.3) | |
| client = OpenAI() | |
| def english_chinese_translation(text_input: str = 'Hi, how are you?', | |
| voice: str = 'coral', | |
| tone: str = 'positive', | |
| speed: float = 1.0): | |
| response = chat_model.invoke(f"Translate the following words in English to traditional Chinese. " | |
| f"Give only the Chinese translation: {text_input}") | |
| audio = client.audio.speech.create(input=response.content, | |
| model="gpt-4o-mini-tts", | |
| voice=voice, | |
| instructions=f"Speak in a {tone} tone.", | |
| response_format="mp3", | |
| speed=speed) | |
| with tempfile.NamedTemporaryFile(delete=False, suffix=".mp3") as temp_audio_file: | |
| temp_audio_file.write(audio.content) | |
| temp_file_path = temp_audio_file.name | |
| return response.content, temp_file_path | |
| interface = gr.Interface(fn=english_chinese_translation, | |
| inputs=[gr.Text(value='Hi, how are you?', | |
| label="English", elem_classes="text_input", elem_id="label"), | |
| gr.Dropdown(choices=['alloy', 'ash', 'ballad', 'coral', 'echo', | |
| 'fable', 'nova', 'onyx', 'sage', 'shimmer', 'verse'], | |
| value='coral', | |
| multiselect=False, | |
| elem_classes="text_input", | |
| elem_id="textarea", | |
| label='VOICE'), | |
| gr.Text(value='positive', label='Tone', elem_classes="text-container"), | |
| gr.Slider(minimum=0.70, | |
| maximum=1.30, | |
| value=1.0, | |
| step=0.01, | |
| label="Speed", | |
| elem_classes="text_input")], | |
| outputs=[gr.Text(label="中文", | |
| # interactive=False, | |
| elem_classes="text_input", | |
| elem_id="label"), | |
| gr.Audio(label="Generated Audio", | |
| type="filepath", | |
| autoplay=True, | |
| show_share_button=False)], | |
| flagging_mode="never", | |
| # examples=[['Hi, how are you?', 'ash', 'cheerful and positive', 1.0]], | |
| # ['Do you speak Chinese?', 'coral', 'commanding', 1.0]], | |
| # live=False, | |
| title='English-Chinese Translator', | |
| description=""" | |
| Translate English words or phrases into traditional Chinese characters, | |
| and then pronounce them in Chinese. Type in the English words or phrases. | |
| Select the voice among 11 choices. Type in the tone like e.g. 'professional', | |
| 'friendly', 'cheerful', etc. Select the speed from 0.7 - 1.3. Then click on | |
| the Submit button to hear the pronunciation in Chinese. | |
| """, | |
| theme=gr.themes.Base(), | |
| css=""" | |
| .gradio-container {font-family: 'Arial'; | |
| # font-size: 54px; | |
| # font-weight: normal; | |
| # text-align: left; | |
| border: 1px solid black; | |
| border-radius: 0px; | |
| background-color: lightblue} | |
| textarea {font-size: 24px; | |
| font-family: 'Verdana'; | |
| font-weight: normal; | |
| # color: lightblue; | |
| # color: rgb(238, 130, 238); | |
| background-color: White; | |
| text-align: left; | |
| border: 1px solid black; | |
| border-radius: 0px | |
| } | |
| button {font-size: 16px; | |
| font-family: 'Arial'; | |
| font-weight: bold; | |
| # text-align: right; | |
| font-style: normal; | |
| # text-transform: uppercase; | |
| border: 1px solid black; | |
| border-radius: 0px; | |
| } | |
| label {font-family: 'Arial'; | |
| # font-size: 24px; | |
| font-weight: bold; | |
| text-align: center; | |
| font-style: normal; | |
| text-transform: uppercase; | |
| line-height: 1.0; | |
| letter-spacing: 2px} | |
| h1 { | |
| font-size: 28px; | |
| font-family: Verdana, Arial, Sans-Serif; | |
| font-style: normal; | |
| font-weight: bold; | |
| color: black; # rgb(238, 130, 238) | |
| # background-color: yellow; | |
| # text-align: right; | |
| text-transform: normal; | |
| letter-spacing: 2px; | |
| border: 1px solid lightblue; | |
| border-radius: 0px | |
| } | |
| p { | |
| font-size: 14px; | |
| font-family: Times New Roman, Times, Serif; | |
| font-style: normal; | |
| font-weight: normal; | |
| color: black; # rgb(238, 130, 238) | |
| # background-color: yellow; | |
| text-align: left; | |
| text-transform: normal; | |
| letter-spacing: 1px; | |
| border: 0px solid lightblue; | |
| border-radius: 0px | |
| } | |
| div { | |
| font-size: 14px; | |
| font-family: Verdana, Arial, Sans-Serif; | |
| font-style: normal; | |
| font-weight: normal; | |
| color: black; # rgb(238, 130, 238) | |
| # background-color: yellow; | |
| text-align: center; | |
| text-transform: normal; | |
| letter-spacing: 0px; | |
| border: 1px solid lightblue; | |
| border-radius: 0px | |
| } | |
| .text_input {font-family: 'Arial'; | |
| # font-size: 32px; | |
| # font-weight: normal; | |
| # text-align: left; | |
| background-color: lightblue} | |
| textarea {font-size: 20px; | |
| font-family: 'Verdana'; | |
| font-weight: normal; | |
| # color: DodgerBlue; | |
| text-align: left} | |
| label {font-family: 'Arial'; | |
| # font-size: 36px; | |
| # font-weight: normal; | |
| text-align: center; | |
| font-style: normal; | |
| text-transform: uppercase; | |
| line-height: 1.0; | |
| letter-spacing: 0px; | |
| border: 1px solid black; | |
| border-radius: 0px; | |
| } | |
| .text_container {font-family: 'Arial'; | |
| font-size: 48px; | |
| font-weight: bold; | |
| text-align: right; | |
| font-style: italic; | |
| text-transform: uppercase; | |
| line-height: 1.0; | |
| } | |
| """ | |
| ) | |
| interface.launch(share=False) | |