# predict.py

import torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import pickle

model_path = 'shirleylqs/mistral-snomed-classification'
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForSequenceClassification.from_pretrained(model_path)

with open(f'{model_path}/label_encoder.pkl', 'rb') as f:
    label_encoder = pickle.load(f)

def predict_class(text):
    inputs = tokenizer(text, return_tensors='pt', truncation=True, max_length=128)
    with torch.no_grad():
        outputs = model(**inputs)
    logits = outputs.logits
    predict_class_id = logits.argmax(-1).item()
    predict_label = label_encoder.inverse_transform([predict_class_id])[0]
    return predict_label

if __name__ == "__main__":
    text = "purulent discharge"
    predicted_label = predict_class(text)
    print(predicted_label)