app.py

import subprocess

import gradio as gr


def run_scripts():
    try:
        # Execute datasets script
        result = subprocess.run(
            ["python", "hub_datasets_by_language.py"],
            capture_output=True,
            text=True,
            check=True,
        )

        # Uncomment this when models script is ready
        # models_result = subprocess.run(
        #     ["python", "hub_models_by_language.py"],
        #     capture_output=True,
        #     text=True,
        #     check=True,
        # )

        # Return success message and updated image paths
        return (
            "✅ Scripts executed successfully! All plots have been updated.",
            "plots/datasets_bar_plot_horizontal.png",
            "plots/datasets_stack_area_en_es.png",
            "plots/datasets_bar_plot_vertical.png",
            "plots/datasets_time_series.png",
        )
    except subprocess.CalledProcessError as e:
        error_msg = e.stderr if e.stderr else e.stdout if e.stdout else str(e)
        # Return error message and keep current images
        return (f"❌ Failed to execute scripts: {error_msg}", None, None, None, None)
    except Exception as e:
        # Return error message and keep current images
        return (f"❌ Unexpected error: {str(e)}", None, None, None, None)


def create_app():
    with gr.Blocks() as app:
        gr.Markdown(
            """
            # Visualizing The Language Gap In The Hugging Face Hub
                    
            The open-source community is creating more and more resources in languages other than English but there is still a huge gap. This Space showcases plots that can help visualize this gap in the case of Spanish and can easily be adapted to other languages.
        """
        )

        gr.Markdown(
            """
            ## English vs Spanish Monolingual Datasets
                    
            Note: We consider only **monolingual** resources in these plots, i.e. datasets and models that only contain data in one language. This is because *most* of the multilingual resources are usually machine-translated and we want to focus on original data.        
        """
        )

        with gr.Row():
            with gr.Column():
                img1 = gr.Image(
                    value="plots/datasets_bar_plot_horizontal.png",
                    label="Distribution of Datasets by Year (Horizontal)",
                    show_label=True,
                    show_download_button=True,
                    show_share_button=True,
                )
                img2 = gr.Image(
                    value="plots/datasets_stack_area_en_es.png",
                    label="Cumulative Growth of Datasets (Stacked)",
                    show_label=True,
                    show_download_button=True,
                    show_share_button=True,
                )
            with gr.Column():
                img3 = gr.Image(
                    value="plots/datasets_bar_plot_vertical.png",
                    label="Distribution of Datasets by Year (Vertical)",
                    show_label=True,
                    show_download_button=True,
                    show_share_button=True,
                )
                img4 = gr.Image(
                    value="plots/datasets_time_series.png",
                    label="Cumulative Growth of Datasets (Line)",
                    show_label=True,
                    show_download_button=True,
                    show_share_button=True,
                )

        # gr.Markdown(
        #     """
        #     ## English vs Spanish Models
        #     """
        # )

        # with gr.Row():
        #     gr.Image(
        #         value="plots/models_stack_area_en_es.png",
        #         label="Cumulative Growth of Models",
        #         show_label=True,
        #         show_download_button=True,
        #         show_share_button=True,
        #     )

        with gr.Row():
            update_button = gr.Button("Update plots with latest data (5 mins)")
            output_label = gr.Label()

        gr.Markdown(
            """
            ## Adapt to other languages

            This Space is WIP and more languages and visuals will be included shortly. Meanwhile, you can clone the Space, adapt the code in the scripts and run it to generate plots for other languages.
            """
        )

        gr.Markdown("## Citation")
        with gr.Accordion("Citation information", open=False):
            gr.Markdown(
                r"""
                If you use these plots or the code please cite:

                ```
                @misc{grandury2024gaphf,
                    author = {María Grandury},
                    title = {Visualizing The Language Gap In The Hugging Face Hub},
                    year = {2024},
                    publisher = {Hugging Face},
                    howpublished = {\url{https://huggingface.co/spaces/mariagrandury/language-gap-in-hf-hub}},
                }
                ```
                """
            )

        update_button.click(
            fn=run_scripts,
            outputs=[output_label, img1, img2, img3, img4],
        )

    return app


app = create_app()
app.launch()