from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import joblib

tokenizer = AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-en-hi")
model = AutoModelForSeq2SeqLM.from_pretrained("Helsinki-NLP/opus-mt-en-hi")
joblib.dump(model, 'model.sav')
loaded_model = joblib.load('model.sav')
joblib.dump(tokenizer, 'tokenizer.sav')
loaded_tokenizer = joblib.load('tokenizer.sav')


def translator(text):
    # function to translate english text to hindi
    input_ids = loaded_tokenizer.encode(text, return_tensors="pt", padding=True)
    outputs = loaded_model.generate(input_ids)
    decoded_text = loaded_tokenizer.decode(outputs[0], skip_special_tokens=True)
    return decoded_text

texts = ["I spend a few hours a day maintaining my website.",
        "Where do random thoughts come from?",
        "I can't believe that she is older than my mother.",
        "My Mum tries to be cool by saying that she likes all the same things that I do",
        "A song can make or ruin a person’s day if they let it get to them."]

for text in texts:
  print("English Text: ", text)
  print("Hindi Translation: ", translator(text))
  print("*"*50,"\n")