import gradio as gr
from transformers import pipeline

title = "Silly Ted-Talk snippet generator"
description = "Tap on the \"Submit\" button to generate a random text snippet."
article = "<p>Fine tuned <a href=\"https://huggingface.co/EleutherAI/gpt-neo-125M\">EleutherAI/gpt-neo-125M</a> upon a formatted <a href=\"https://www.kaggle.com/datasets/miguelcorraljr/ted-ultimate-dataset\"> TED – Ultimate Dataset</a> (English)</p>"

model_id = "./model"
text_generator = pipeline('text-generation', model=model_id, tokenizer=model_id)
max_length = 128
top_k = 40
top_p = 0.92
temperature = 1.0

def text_generation(input_text = None):
    if input_text == None or len(input_text) == 0:
        input_text = "\t\""
    else:
        input_text.replace("\"", "")
        if input_text.startswith("<|startoftext|>") == False:            
            input_text ="\t\"" + input_text
    generated_text = text_generator(input_text,
    max_length=max_length,
    top_k=top_k, 
    top_p=top_p,
    temperature=temperature,
    do_sample=True,
    repetition_penalty=2.0,
    num_return_sequences=1)
    parsed_text = generated_text[0]["generated_text"].replace("<|startoftext|>", "").replace("\r","").replace("\n\n", "\n").replace("\t", " ").replace("<|pad|>", " * ").replace("\"\"", "\"")
    return parsed_text

gr.Interface(
    text_generation,    
    [gr.inputs.Textbox(lines=1, label="Enter input text or leave blank")],
    outputs=[gr.outputs.Textbox(type="text", label="Generated Ted-Talk snippet")],
    title=title,
    description=description,
    article=article,
    theme="default",
    allow_flagging=False,
).launch()