career_data.py

from transformers import AutoModel, AutoTokenizer
import torch

# Load model and tokenizer
model_name = "minishlab/M2V_base_output"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModel.from_pretrained(model_name)

# Career options with precomputed skills and interests
career_options = {
    "Software Engineer": {
        "skills": "programming, problem-solving",
        "interests": "technology, innovation"
    },
    "Graphic Designer": {
        "skills": "design, creativity",
        "interests": "art, visual communication"
    },
    "Project Manager": {
        "skills": "management, organization",
        "interests": "leadership, strategy"
    },
    # Add more careers as needed
}

# Generate embeddings for career options
def get_embedding(text):
    inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True)
    with torch.no_grad():
        embedding = model(**inputs).last_hidden_state.mean(dim=1).squeeze()
    return embedding

career_embeddings = {}
for career, attributes in career_options.items():
    combined_text = attributes["skills"] + ", " + attributes["interests"]
    career_embeddings[career] = get_embedding(combined_text)

# Function to recommend careers based on skills and interests
def get_career_recommendations(skills: str, interests: str):
    user_input = skills + ", " + interests
    user_embedding = get_embedding(user_input)
    
    recommendations = []
    for career, career_embedding in career_embeddings.items():
        similarity = torch.cosine_similarity(user_embedding, career_embedding, dim=0).item()
        recommendations.append((career, similarity))
    
    recommendations.sort(key=lambda x: x[1], reverse=True)
    
    return [f"{career} (Similarity: {similarity:.2f})" for career, similarity in recommendations[:5]]