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("---")