Spaces:
Build error
Build error
| """ | |
| 🗣️ Translator - Translate text from one language to another. | |
| Application file made with Streamlit. | |
| Author: | |
| - @ChainYo | |
| """ | |
| import re | |
| import streamlit as st | |
| from datetime import datetime | |
| from transformers import pipeline | |
| from available_models import MODELS | |
| st.set_page_config(page_title="Translator", page_icon="🗣️") | |
| st.title("🗣️ Translator") | |
| st.subheader("Translation made fast and easy.") | |
| st.markdown(""" | |
| [](https://github.com/ChainYo) | |
| [](https://huggingface.co/ChainYo) | |
| [](https://www.linkedin.com/in/thomas-chaigneau-dev/) | |
| [](https://discord.gg/) | |
| """) | |
| st.write("To add a new model, hit me up! ⬆️") | |
| with st.expander(label="❓ How does it work", expanded=True): | |
| st.markdown(""" | |
| **Translator** is a **simple tool** that allows you to **translate text** from one language to another. | |
| **Translator** is powered by the [Transformers library](https://huggingface.co/transformers) and uses the | |
| [Helsinki-NLP](https://huggingface.co/Helsinki-NLP) models. | |
| Choose the **source language**, the **target language** and add some **text to translate**. | |
| **Translator** will translate the text and **save the output in a text file**. It cuts the sentences by following | |
| the punctuation marks. | |
| The output file content will also be displayed in the browser to help you understand the translation and choose | |
| if you want to download it. | |
| There is **no limit to the number of characters** that can be translated. | |
| The only limit is the time you are ready to wait! 🤗 | |
| *P.S. I have built this tool to help me start writing blog posts in different languages. I am a French native speaker | |
| and I will use it to translate my potential future blog posts in English.* | |
| *P.P.S. I am a **Junior ML Engineer** passionate about **machine learning** and **data science**. Reach out to me by | |
| clicking on the socials badges above.* | |
| """) | |
| lang1, lang2 = st.columns(2) | |
| lang1.selectbox( | |
| "Source Language", ["🇬🇧 English", "🇫🇷 French", "🇩🇪 German", "🇪🇸 Spanish", "🇷🇺 Russian"], | |
| key="input_lang", index=1, | |
| ) | |
| lang2.selectbox( | |
| "Target Language", ["🇬🇧 English", "🇫🇷 French", "🇩🇪 German", "🇪🇸 Spanish", "🇷🇺 Russian"], | |
| key="output_lang", index=0, | |
| ) | |
| selected_model = MODELS[f"{st.session_state['input_lang']}->{st.session_state['output_lang']}"] | |
| if selected_model[0] == None: | |
| st.write("No model available for this pair.") | |
| elif selected_model[0] == 0: | |
| st.write("No translation necessary.") | |
| else: | |
| st.markdown(f""" | |
| **Selected model:** [{selected_model[0]}]({selected_model[1]}) | |
| """) | |
| input_text = st.text_area("Enter text to translate:", height=400, key="input") | |
| translate_text = st.button("Translate") | |
| if translate_text: | |
| with st.spinner(text="⚙️ Model loading..."): | |
| task = pipeline( | |
| "translation", | |
| model=selected_model[0], | |
| tokenizer=selected_model[0], | |
| ) | |
| progress_bar = st.progress(0) | |
| with st.spinner(text="🔄 Translating..."): | |
| text_to_translate = re.split('(?<=[.!?]) +', input_text) | |
| total_progress = len(text_to_translate) | |
| for i, text in enumerate(text_to_translate): | |
| translation = task(text) | |
| text_to_translate[i] = translation[0]["translation_text"] | |
| progress_bar.progress((i + 1) / total_progress) | |
| st.success("🗣️ Translated!") | |
| st.write(f"**Translation:** {' '.join(text_to_translate)}") | |
| st.download_button( | |
| label="Download translated text", | |
| data="\n".join(text_to_translate), | |
| file_name=f"{st.session_state['input_lang']}-{st.session_state['output_lang']}-{datetime.now().strftime('%Y-%m-%d-%H-%M-%S')}.txt", | |
| mime="text/plain" | |
| ) | |