import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM, set_seed, pipeline


title = "Python Code Generator"
description = "This is a space to convert English text to Python code using the [codeparrot-small-text-to-code](https://huggingface.co/codeparrot/codeparrot-small-text-to-code) model, a pre-trained Python code generation model trained on a dataset of docstrings and Python code extracted from Jupyter notebooks available at [github-jupyter-text](https://huggingface.co/datasets/codeparrot/github-jupyter-text)."
example = [
    ["Utility function to calculate the precision of predictions using sklearn metrics", 65, 0.6, 42],
    ["Let's implement a function that calculates the size of a file called filepath", 60, 0.6, 42],
    ["Let's implement the Bubble Sort sorting algorithm in an auxiliary function:", 87, 0.6, 42],
    ]

# Change the model to the pre-trained model
tokenizer = AutoTokenizer.from_pretrained("codeparrot/codeparrot-small-text-to-code")
model = AutoModelForCausalLM.from_pretrained("codeparrot/codeparrot-small-text-to-code")

def create_docstring(gen_prompt):
    return "\"\"\"\n" + gen_prompt + "\n\"\"\"\n\n"

def generate_code(gen_prompt, max_tokens, temperature=0.6, seed=42):
    set_seed(seed)
    pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
    prompt = create_docstring(gen_prompt)
    generated_text = pipe(prompt, do_sample=True, top_p=0.95, temperature=temperature, max_new_tokens=max_tokens)[0]['generated_text']
    return generated_text


iface = gr.Interface(
    fn=generate_code, 
    inputs=[
        gr.Textbox(label="English instructions", placeholder="Enter English instructions..."),
        gr.inputs.Slider(
            minimum=8,
            maximum=256,
            step=1,
            default=8,
            label="Number of tokens to generate",
        ),
        gr.inputs.Slider(
            minimum=0,
            maximum=2.5,
            step=0.1,
            default=0.6,
            label="Temperature",
        ),
        gr.inputs.Slider(
            minimum=0,
            maximum=1000,
            step=1,
            default=42,
            label="Random seed for generation"
        )
    ],
    outputs=gr.Code(label="Generated Python code", language="python", lines=10),
    examples=example,
    layout="horizontal",
    theme="peach",
    description=description,
    title=title
)
iface.launch()