import os
os.system("pip install -r requirements.txt")
os.system("pip freeze")
import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM, set_seed, pipeline


tokenizer = AutoTokenizer.from_pretrained("codeparrot/codeparrot-small-code-to-text")
model = AutoModelForCausalLM.from_pretrained("codeparrot/codeparrot-small-code-to-text")
pipe = pipeline("text-generation", model=model, tokenizer=tokenizer, num_return_sequences=1, device=-1)


def make_doctring(gen_prompt):
    return gen_prompt + f"\n\n\"\"\"\nExplanation:"


def code_generation(gen_prompts, max_tokens=8, temperature=0.6, seed=42):
    set_seed(seed)
    prompts = [make_doctring(p) for p in gen_prompts]
    generated_text = pipe(prompts, do_sample=True, top_p=0.95, temperature=temperature, max_length=max_tokens)[0]
    return generated_text["generated_text"]


title = "Code Explainer"
description = "This is a space to convert Python code into english text explaining what it does using [codeparrot-small-code-to-text](https://huggingface.co/codeparrot/codeparrot-small-code-to-text),\
            a code generation model for Python finetuned on [github-jupyter-code-to-text](https://huggingface.co/datasets/codeparrot/github-jupyter-code-to-text) a dataset of Python code followed by a docstring explaining it, the data was originally extracted from Jupyter notebooks."

EXAMPLES = [
    ["def sort_function(arr):\n    n = len(arr)\n \n    # Traverse through all array elements\n    for i in range(n):\n \n        # Last i elements are already in place\n        for j in range(0, n-i-1):\n \n            # traverse the array from 0 to n-i-1\n            # Swap if the element found is greater\n            # than the next element\n            if arr[j] > arr[j+1]:\n                arr[j], arr[j+1] = arr[j+1], arr[j]"],
    ["from sklearn import model_selection\nX_train, X_test, Y_train, Y_test = model_selection.train_test_split(X, Y, test_size=0.2)"],
    ["def load_text(filename):\n    with open(filename, 'r') as f:\n        text = f.read()\n    return text"]
]
 
iface = gr.Interface(
    fn=code_generation,
    inputs=[
        gr.inputs.Code(language="python", label="Python code snippet", lines=10),
        gr.inputs.Slider(minimum=8, maximum=256, step=1, default=256, label="Number of tokens to generate"),
        gr.inputs.Slider(minimum=0, maximum=2.5, step=0.1, default=0.1, label="Temperature"),
        gr.inputs.Slider(minimum=0, maximum=1000, step=1, default=42, label="Random seed")
    ],
    outputs=gr.outputs.Code(language="text", label="Generated explanation", lines=10),
    examples=EXAMPLES,
    layout="horizontal",
    theme="monochrome",
    description=description,
    title=title
)
iface.launch()