app.py

import streamlit as st
import requests
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics.pairwise import cosine_similarity
from transformers import MarianMTModel, MarianTokenizer

# Set up the Streamlit page
st.title("AI Opportunity Finder for Youth")
st.write("Find Scholarships, Internships, Online Courses, and more!")

# Language Translation Function
def translate_text(text, target_lang='de'):
    # Use Hugging Face's MarianMT for translation
    model_name = f'Helsinki-NLP/opus-mt-en-{target_lang}'
    model = MarianMTModel.from_pretrained(model_name)
    tokenizer = MarianTokenizer.from_pretrained(model_name)

    translated = model.generate(**tokenizer(text, return_tensors="pt", padding=True, truncation=True))
    translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
    return translated_text

# Mock function to get data from APIs (replace with actual API calls)
def get_scholarships(country, interests):
    url = f"https://jsonplaceholder.typicode.com/posts"  # Mock API (replace with real one)
    
    # Simulate API response based on country
    if country == "USA":
        return [{"title": f"USA Scholarship {i+1}", "description": f"Description for scholarship {i+1} in USA.", "eligibility": "Any student from USA."} for i in range(5)]
    elif country == "Germany":
        return [{"title": f"Germany Scholarship {i+1}", "description": f"Description for scholarship {i+1} in Germany.", "eligibility": "Any student from Germany."} for i in range(5)]
    else:
        return [{"title": f"Scholarship {i+1}", "description": f"Description for scholarship {i+1} in {country}.", "eligibility": "Any student from any background."} for i in range(5)]

def get_internships(country):
    url = f"https://jsonplaceholder.typicode.com/posts"  # Mock API for testing
    # Simulate internships data
    if country == "USA":
        return [{"jobtitle": f"Internship {i+1}", "company": "USA Company", "location": "USA", "snippet": "Description of internship in USA."} for i in range(5)]
    elif country == "Germany":
        return [{"jobtitle": f"Internship {i+1}", "company": "Germany Company", "location": "Germany", "snippet": "Description of internship in Germany."} for i in range(5)]
    else:
        return [{"jobtitle": f"Internship {i+1}", "company": "Sample Company", "location": "Remote", "snippet": "Description of internship."} for i in range(5)]

def recommend_opportunities(user_interests, user_skills, opportunities):
    user_profile = [f"{user_interests} {user_skills}"]
    opportunities_text = [f"{opportunity.get('description', 'No description available')} {opportunity.get('eligibility', 'No eligibility available')}" for opportunity in opportunities]
    
    # Vectorize the text using TF-IDF
    vectorizer = TfidfVectorizer(stop_words='english')
    tfidf_matrix = vectorizer.fit_transform(opportunities_text + user_profile)
    
    # Compute cosine similarity
    cosine_sim = cosine_similarity(tfidf_matrix[-1], tfidf_matrix[:-1])
    
    # Get the top 5 recommendations
    recommendations = cosine_sim[0].argsort()[-5:][::-1]
    
    return [opportunities[i] for i in recommendations]

# Form to gather user profile and country selection
with st.form(key='user_form'):
    st.sidebar.header("User Profile")
    location = st.selectbox("Select your Country", ["USA", "Germany", "UK", "India", "Australia", "Pakistan"])  # You can add more countries here
    skills = st.text_input("Skills (e.g., Python, Marketing)")
    interests = st.text_input("Interests (e.g., Technology, Science)")
    target_language = st.selectbox("Select target language", ['de', 'fr', 'es', 'it', 'pt'])  # Available language codes for translation

    submit_button = st.form_submit_button("Find Opportunities")

# Fetch data based on the user input
if submit_button:
    # Fetch scholarships and internships based on the selected country and profile
    scholarships = get_scholarships(location, interests)
    internships = get_internships(location)
    
    # Display Scholarships
    if scholarships:
        st.write("Scholarships found:")
        for scholarship in scholarships:
            title = translate_text(scholarship.get('title', 'No title available'), target_language)
            description = translate_text(scholarship.get('description', 'No description available'), target_language)
            eligibility = translate_text(scholarship.get('eligibility', 'No eligibility available'), target_language)
            
            st.write(f"Title: {title}")
            st.write(f"Description: {description}")
            st.write(f"Eligibility: {eligibility}")
            st.write("---")
    else:
        st.write("No scholarships found for the selected country.")
    
    # Display Internships
    if internships:
        st.write("Internships found:")
        for internship in internships:
            title = translate_text(internship.get('jobtitle', 'No title available'), target_language)
            company = translate_text(internship.get('company', 'No company available'), target_language)
            location = translate_text(internship.get('location', 'No location available'), target_language)
            snippet = translate_text(internship.get('snippet', 'No snippet available'), target_language)
            
            st.write(f"Title: {title}")
            st.write(f"Company: {company}")
            st.write(f"Location: {location}")
            st.write(f"Snippet: {snippet}")
            st.write("---")
    else:
        st.write("No internships found for the selected country.")
    
    # AI Recommendations based on interests and skills
    all_opportunities = scholarships + internships
    recommended_opportunities = recommend_opportunities(interests, skills, all_opportunities)
    
    st.write("AI-based Recommended Opportunities based on your profile:")
    for opportunity in recommended_opportunities:
        title = translate_text(opportunity.get('title', 'No title available'), target_language)
        description = translate_text(opportunity.get('description', 'No description available'), target_language)
        eligibility = translate_text(opportunity.get('eligibility', 'Not available'), target_language)
        
        st.write(f"Title: {title}")
        st.write(f"Description: {description}")
        st.write(f"Eligibility: {eligibility}")
        st.write("---")