app.py

import streamlit as st
from streamlit_option_menu import option_menu
import requests
import os

API_TOKEN = os.environ.get("API_TOKEN")


#changement du logo et du titre de mon application en anglais
st.set_page_config(page_title="TSC", page_icon="books", layout="centered", menu_items=None)


col1, col2, col3 = st.columns(3)
xx = col1
ar = col2.image("keyce.jpg", width=180)
yy = col3


# menu horizontal
selected = option_menu(None, ["Home", "Translate", "Summarization", "Chatbot"], 
    icons=['house', 'globe', "pencil", "chat"], 
    menu_icon="cast", default_index=0, orientation="horizontal",
    styles={
            "container": {"padding":"5!important", "background-color":"#000E38", "color":"white"},
            "icon": {"color":"#fff", "font-size":"16px"}, 
            "nav-link": {"font-size":"16px", "text-align":"left", "margin":"0px", "--hover-color":"#f9d1ac", "color":"white"},
            "nav-link-selected": {"background-color": "#FF9633", "color": "white"},
    })


if selected == "Home":
    st.subheader("Welcome to our app !!")
    st.info(":blue[We are the present of tomorrow! Please like it after use!]")
    #Insertion d'une image 
    st.image("nlp1.png", width=None)  
    st.write("Hello to everybody! In this Artificial Intelligence app, we offer you a THREE-IN-ONE package. First of all, we've got the (Translate) menu, which lets you do a good job of translating into a few foreign languages, then we've got the (Summary) option, which lets you enter your text in order to summarize it, and finally, as a bonus, we've integrated a chatbot model in the (Chatbot) option, which lets you try out our chatbot, which we're currently perfecting.") 

elif selected == 'Translate':
    # Choose the translation language
    translated_model = {
        "english - spanish": "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-es",
        "english - french": "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-fr",
        "english - chinese": "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-zh",
        "english - german": "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-de"
    }
    languages = st.selectbox("", list(translated_model.keys()))
    API_TRANSLATED = translated_model[languages]
    col3, col4 = st.columns(2)
    text_in = col3.text_area("", height=200, placeholder='type or paste the text you wish to translate.')
    #Boutton pour effectuer la traduction
    btn_translate = st.button("🌐 Translate")
    if btn_translate:
        headers_spanish = {"Authorization": API_TOKEN}
        if languages == "english - spanish" and text_in:
            def query(payload):
                response = requests.post(API_TRANSLATED, headers=headers_spanish, json=payload)
                return response.json() 
            output = query_spanish({"inputs": text_in})
            if not output[0]["translation_text"]:
                error_message = output[0]["error"]
                st.error(f"Le texte n'a pas pu être traduit: {error_message}")
            else:
                translated_text = output[0]["translation_text"]
                col5.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ translated_text +'''</p>''',  unsafe_allow_html=True)
                
            # text_out_spanish = col4.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ output_spanish[0] ["translation_text"] +'''</p>''',  unsafe_allow_html=True)
            st.info(":blue[Please feel fit to like it after use! 😉🙏]")
        elif languages == "english - french" and text_in:
            API_URL_french = "https://api-inference.huggingface.co/models/Helsinki-NLP/opus-mt-en-fr"
            headers_french = {"Authorization": API_TOKEN}
            if btn_translate:
                def query(payload):
                    response = requests.post(API_TRANSLATED, headers=headers_french, json=payload)
                    return response.json() 
                output = query({"inputs": text_in})

                if not output[0]["translation_text"]:
                    error_message = output[0]["error"]
                    st.error(f"Le texte n'a pas pu être traduit: {error_message}")
                else:
                    translated_text = output_translate[0]["translation_text"]
                    col5.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ translated_text +'''</p>''',  unsafe_allow_html=True)
                
                # text_out_french = col4.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ output_french[0] ["translation_text"] +'''</p>''',  unsafe_allow_html=True)
                # st.info(":blue[Please feel fit to like it after use! 😉🙏]")   
        elif languages == 'english - chinese' and text_in:
            headers_chinese = {"Authorization": API_TOKEN}       
            if btn_translate:
                def query(payload):
                    response = requests.post(API_TRANSLATED, headers=headers_chinese, json=payload)
                    return response.json() 
                output = query({"inputs": text_in})
                if not output[0]["translation_text"]:
                    error_message = output[0]["error"]
                    st.error(f"Le texte n'a pas pu être traduit: {error_message}")
                else:
                    translated_text = output[0]["translation_text"]
                    col5.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ translated_text +'''</p>''',  unsafe_allow_html=True)
                
                # text_out_chinese = col4.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ output_chinese[0] ["translation_text"] +'''</p>''',  unsafe_allow_html=True)
                st.info(":blue[Please feel fit to like it after use! 😉🙏]")
        elif languages == "english - german" and text_in:
            headers_german = {"Authorization": API_TOKEN}
            if btn_translate:
                def query(payload):
                    response = requests.post(API_TRANSLATED, headers=headers_german, json=payload)
                    return response.json() 
                output = query({"inputs": text_in})
                if not output[0]["translation_text"]:
                    error_message = output[0]["error"]
                    st.error(f"Le texte n'a pas pu être traduit: {error_message}")
                else:
                    translated_text = output[0]["translation_text"]
                    col5.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ translated_text +'''</p>''',  unsafe_allow_html=True)
                
                # text_out_german = col4.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ output_german[0] ["translation_text"] +'''</p>''',  unsafe_allow_html=True)
                st.info(":blue[Please feel fit to like it after use! 😉🙏]")

elif selected == 'Summarization':
    text_sum = st.text_area("", height=250, placeholder='Please type or paste the text you wish to summarise here.')
    #Boutton pour effectuer la prédiction
    btn_sum = st.button("📝 Summary")
    if btn_sum:
        if text_sum:
            API_URL_summary = "https://api-inference.huggingface.co/models/Falconsai/text_summarization"
            headers_summary = {"Authorization": API_TOKEN}
            def query(payload):
                response = requests.post(API_URL_summary, headers=headers_summary, json=payload)
                return response.json()
            st.markdown('''<h4 style='text-align:left;color:black;'> Answer </h4>''',  unsafe_allow_html=True) 
            output = query({"inputs": text_sum,})
            summary_text = output[0]["summary_text"]
            st.success(f"Response: {summary_text}")
            # text_out_sum = st.markdown('''<p style='text-align:left;font-size:16px;color:orange;font-weight:bold'>'''+ output_sum[0]["summary_text"] +'''</p>''',  unsafe_allow_html=True)
            st.info(":blue[Please feel fit to like it after use! 😉🙏]")

elif selected == 'Chatbot':
    text_chat = st.text_input("Enter your question")
    #Boutton pour effectuer le chatbot
    btn_chat = st.button("💬 Submit")
    API_URL_chatbot = "https://api-inference.huggingface.co/models/tiiuae/falcon-7b-instruct"
    headers_chatbot = {"Authorization": API_TOKEN}
    if btn_chat:
        if text_chat:
            def query(payload):
                response = requests.post(API_URL_chatbot, headers=headers_chatbot, json=payload)
                return response.json()
            output = query({"inputs": text_chat,})
            generated_text = output[0]["generated_text"]
            st.success(f"Response: {generated_text}")
            # text_out_chat = st.info(':green['+ output_chat[0]["generated_text"] +']')
            st.info(":blue[Please feel fit to like it after use! 😉🙏]")

st.markdown('''<br/> <h6 style='text-align:center;color:grey;font-size:11px;'> © 2024 - TOUNDE ZOGO - All rights reserved </h6>''',  unsafe_allow_html=True)

# for key in range(5):
#     try:
#         text_out = col4.markdown('''<br/> <p style='text-align:left;font-size:16px;'>'''+ output[0]["translation_text"] +'''</p>''',  unsafe_allow_html=True)
#     except KeyError:
#         print("Couldn't find a match for the key:", key)