src/distilabel_dataset_generator/apps/sft.py

import multiprocessing
import time

import gradio as gr
import pandas as pd
from distilabel.distiset import Distiset
from huggingface_hub import whoami

from src.distilabel_dataset_generator.pipelines.sft import (
    DEFAULT_DATASET,
    DEFAULT_DATASET_DESCRIPTION,
    DEFAULT_SYSTEM_PROMPT,
    MODEL,
    PROMPT_CREATION_PROMPT,
    get_pipeline,
    get_prompt_generation_step,
)
from src.distilabel_dataset_generator.utils import (
    get_login_button,
    get_org_dropdown,
    get_token,
)


def _run_pipeline(result_queue, num_turns, num_rows, system_prompt):
    pipeline = get_pipeline(
        num_turns,
        num_rows,
        system_prompt,
    )
    distiset: Distiset = pipeline.run(use_cache=False)
    result_queue.put(distiset)


def generate_system_prompt(dataset_description, progress=gr.Progress()):
    progress(0.1, desc="Initializing text generation")
    generate_description = get_prompt_generation_step()
    progress(0.4, desc="Loading model")
    generate_description.load()
    progress(0.7, desc="Generating system prompt")
    result = next(
        generate_description.process(
            [
                {
                    "system_prompt": PROMPT_CREATION_PROMPT,
                    "instruction": dataset_description,
                }
            ]
        )
    )[0]["generation"]
    progress(1.0, desc="System prompt generated")
    return result


def generate_sample_dataset(system_prompt, progress=gr.Progress()):
    progress(0.1, desc="Initializing sample dataset generation")
    result = generate_dataset(system_prompt, num_turns=1, num_rows=2, progress=progress)
    progress(1.0, desc="Sample dataset generated")
    return result


def generate_dataset(
    system_prompt,
    num_turns=1,
    num_rows=5,
    private=True,
    org_name=None,
    repo_name=None,
    token=None,
    progress=gr.Progress(),
):
    repo_id = (
        f"{org_name}/{repo_name}"
        if repo_name is not None and org_name is not None
        else None
    )
    if repo_id is not None:
        if not repo_id:
            raise gr.Error(
                "Please provide a repo_name and org_name to push the dataset to."
            )
        try:
            whoami(token=token)
        except Exception:
            raise gr.Error(
                "Provide a Hugging Face token with write access to the organization you want to push the dataset to."
            )

    if num_turns > 4:
        num_turns = 4
        gr.Info("You can only generate a dataset with 4 or fewer turns. Setting to 4.")
    if num_rows > 5000:
        num_rows = 5000
        gr.Info(
            "You can only generate a dataset with 5000 or fewer rows. Setting to 5000."
        )

    if num_rows < 50:
        duration = 60
    elif num_rows < 250:
        duration = 300
    elif num_rows < 1000:
        duration = 500
    else:
        duration = 1000

    gr.Info(
        "Dataset generation started. This might take a while. Don't close the page.",
        duration=duration,
    )
    result_queue = multiprocessing.Queue()
    p = multiprocessing.Process(
        target=_run_pipeline,
        args=(result_queue, num_turns, num_rows, system_prompt),
    )

    try:
        p.start()
        total_steps = 100
        for step in range(total_steps):
            if not p.is_alive() or p._popen.poll() is not None:
                break
            progress(
                (step + 1) / total_steps,
                desc=f"Generating dataset with {num_rows} rows",
            )
            time.sleep(duration / total_steps)  # Adjust this value based on your needs
        p.join()
    except Exception as e:
        raise gr.Error(f"An error occurred during dataset generation: {str(e)}")

    distiset = result_queue.get()

    if repo_id is not None:
        progress(0.95, desc="Pushing dataset to Hugging Face Hub.")
        distiset.push_to_hub(
            repo_id=repo_id,
            private=private,
            include_script=False,
            token=token,
        )

    # If not pushing to hub generate the dataset directly
    distiset = distiset["default"]["train"]
    if num_turns == 1:
        outputs = distiset.to_pandas()[["prompt", "completion"]]
    else:
        outputs = distiset.to_pandas()[["messages"]]

    progress(1.0, desc="Dataset generation completed")
    return pd.DataFrame(outputs)


def generate_pipeline_code(system_prompt):
    code = f"""
from distilabel.pipeline import Pipeline
from distilabel.steps import KeepColumns
from distilabel.steps.tasks import MagpieGenerator
from distilabel.llms import InferenceEndpointsLLM

MODEL = "{MODEL}"
SYSTEM_PROMPT = "{system_prompt}"
# increase this to generate multi-turn conversations
NUM_TURNS = 1
# increase this to generate a larger dataset
NUM_ROWS = 100

with Pipeline(name="sft") as pipeline:
    magpie = MagpieGenerator(
        llm=InferenceEndpointsLLM(
            model_id=MODEL,
            tokenizer_id=MODEL,
            magpie_pre_query_template="llama3",
            generation_kwargs={{
                "temperature": 0.8,
                "do_sample": True,
                "max_new_tokens": 2048,
                "stop_sequences": [
                    "<|eot_id|>",
                    "<|end_of_text|>",
                    "<|start_header_id|>",
                    "<|end_header_id|>",
                    "assistant",
                ],
            }}
        ),
        n_turns=NUM_TURNS,
        num_rows=NUM_ROWS,
        system_prompt=SYSTEM_PROMPT,
    )

if __name__ == "__main__":
    distiset = pipeline.run()
"""
    return code


def update_pipeline_code(system_prompt):
    return generate_pipeline_code(system_prompt)


with gr.Blocks(
    title="⚗️ Distilabel Dataset Generator",
    head="⚗️ Distilabel Dataset Generator",
) as app:
    get_login_button()
    gr.Markdown("## Iterate on a sample dataset")
    with gr.Column() as main_ui:
        dataset_description = gr.TextArea(
            label="Provide a description of the dataset",
            value=DEFAULT_DATASET_DESCRIPTION,
        )
        with gr.Row():
            gr.Column(scale=1)
            btn_generate_system_prompt = gr.Button(value="Generate sample dataset")
            gr.Column(scale=1)

        system_prompt = gr.TextArea(
            label="If you want to improve the dataset, you can tune the system prompt and regenerate the sample",
            value=DEFAULT_SYSTEM_PROMPT,
        )

        with gr.Row():
            gr.Column(scale=1)
            btn_generate_sample_dataset = gr.Button(
                value="Regenerate sample dataset",
            )
            gr.Column(scale=1)

        with gr.Row():
            table = gr.DataFrame(
                value=DEFAULT_DATASET,
                interactive=False,
                wrap=True,
            )

        result = btn_generate_system_prompt.click(
            fn=generate_system_prompt,
            inputs=[dataset_description],
            outputs=[system_prompt],
            show_progress=True,
        ).then(
            fn=generate_sample_dataset,
            inputs=[system_prompt],
            outputs=[table],
            show_progress=True,
        )

        btn_generate_sample_dataset.click(
            fn=generate_sample_dataset,
            inputs=[system_prompt],
            outputs=[table],
            show_progress=True,
        )

        # Add a header for the full dataset generation section
        gr.Markdown("## Generate full dataset")
        gr.Markdown(
            "Once you're satisfied with the sample, generate a larger dataset and push it to the hub. Get <a href='https://huggingface.co/settings/tokens' target='_blank'>a Hugging Face token</a> with write access to the organization you want to push the dataset to."
        )

        with gr.Column() as push_to_hub_ui:
            with gr.Row(variant="panel"):
                num_turns = gr.Number(
                    value=1,
                    label="Number of turns in the conversation",
                    minimum=1,
                    maximum=4,
                    step=1,
                    info="Choose between 1 (single turn with 'instruction-response' columns) and 2-4 (multi-turn conversation with a 'conversation' column).",
                )
                num_rows = gr.Number(
                    value=100,
                    label="Number of rows in the dataset",
                    minimum=1,
                    maximum=5000,
                    info="The number of rows in the dataset. Note that you are able to generate more rows at once but that this will take time.",
                )

            with gr.Row(variant="panel"):
                hf_token = gr.Textbox(label="HF token", type="password")
                org_name = get_org_dropdown()
                repo_name = gr.Textbox(label="Repo name", placeholder="dataset_name")
                private = gr.Checkbox(
                    label="Private dataset", value=True, interactive=True, scale=0.5
                )

            btn_generate_full_dataset = gr.Button(
                value="⚗️ Generate Full Dataset", variant="primary"
            )

            success_message = gr.Markdown(visible=False)

    def show_success_message(org_name, repo_name):
        return gr.Markdown(
            value=f"""
            <div style="padding: 1em; background-color: #e6f3e6; border-radius: 5px; margin-top: 1em;">
                <h3 style="color: #2e7d32; margin: 0;">Dataset Published Successfully!</h3>
                <p style="margin-top: 0.5em;">
                    Your dataset is now available at:
                    <a href="https://huggingface.co/datasets/{org_name}/{repo_name}" target="_blank" style="color: #1565c0; text-decoration: none;">
                        https://huggingface.co/datasets/{org_name}/{repo_name}
                    </a>
                </p>
            </div>
        """,
            visible=True,
        )

    btn_generate_full_dataset.click(
        fn=generate_dataset,
        inputs=[
            system_prompt,
            num_turns,
            num_rows,
            private,
            org_name,
            repo_name,
            hf_token,
        ],
        outputs=[table],
        show_progress=True,
    ).then(
        fn=show_success_message,
        inputs=[org_name, repo_name],
        outputs=[success_message],
    )

    gr.Markdown("## Or run this pipeline locally with distilabel")

    with gr.Accordion("Run this pipeline on Distilabel", open=False):
        pipeline_code = gr.Code(language="python", label="Distilabel Pipeline Code")

    system_prompt.change(
        fn=update_pipeline_code,
        inputs=[system_prompt],
        outputs=[pipeline_code],
    )

    app.load(get_token, outputs=[hf_token])
    app.load(get_org_dropdown, outputs=[org_name])