app.py

import json

import gradio as gr
import spaces

from data_utils import (
    dataset_overview,
    get_class_names,
    get_images_for_gallery,
)
from train_utils import (
    train_model,
    list_saved_models,
    model_meta_path,
    evaluate_saved_model,
)
from predict_utils import (
    predict_uploaded_image,
    test_random_sample,
)


def load_dataset_overview_callback():
    try:
        summary, distribution_df = dataset_overview()
        class_names = ["Toutes les classes"] + get_class_names()

        return (
            summary,
            distribution_df,
            gr.update(choices=class_names, value="Toutes les classes"),
        )

    except Exception as e:
        return (
            {"Erreur": str(e)},
            None,
            gr.update(),
        )


def refresh_gallery_callback(split_name, class_name, max_images):
    try:
        gallery = get_images_for_gallery(
            split_name=split_name,
            class_name=class_name,
            max_images=int(max_images),
        )
        return gallery
    except Exception as e:
        return [(None, f"Erreur : {str(e)}")]


@spaces.GPU(duration=300)
def train_callback(
    dropout,
    fc_dim,
    learning_rate,
    weight_decay,
    batch_size,
    epochs,
    fine_tune_mode,
    model_tag,
):
    try:
        result = train_model(
            dropout=float(dropout),
            fc_dim=int(fc_dim),
            learning_rate=float(learning_rate),
            weight_decay=float(weight_decay),
            batch_size=int(batch_size),
            epochs=int(epochs),
            fine_tune_mode=str(fine_tune_mode),
            model_tag=model_tag,
        )

        models = list_saved_models()
        selected = result["model_name"] if result["model_name"] in models else None

        return (
            result["logs"],
            result["history"],
            result["summary"],
            result["classification_report"],
            result["confusion_matrix"],
            result["confusion_matrix_path"],
            gr.update(choices=models, value=selected),
        )

    except Exception as e:
        return (
            f"Échec de l’entraînement :\n{str(e)}",
            None,
            None,
            None,
            None,
            None,
            gr.update(),
        )


@spaces.GPU(duration=120)
def evaluate_saved_model_callback(model_name):
    try:
        summary, report_df, cm_df, cm_path = evaluate_saved_model(model_name)
        return summary, report_df, cm_df, cm_path
    except Exception as e:
        return {"Erreur": str(e)}, None, None, None


@spaces.GPU(duration=60)
def predict_uploaded_image_callback(model_name, image):
    try:
        return predict_uploaded_image(model_name, image)
    except Exception as e:
        return f"Échec de la prédiction :\n{str(e)}", None


@spaces.GPU(duration=60)
def test_random_sample_callback(model_name):
    try:
        return test_random_sample(model_name)
    except Exception as e:
        return None, f"Échec du test aléatoire :\n{str(e)}", None


def refresh_models_dropdown():
    models = list_saved_models()
    return gr.update(choices=models, value=models[0] if models else None)


def get_model_info(model_name: str):
    if not model_name:
        return {"message": "Aucun modèle sélectionné."}

    meta_file = model_meta_path(model_name)

    try:
        with open(meta_file, "r", encoding="utf-8") as f:
            return json.load(f)
    except FileNotFoundError:
        return {"message": "Métadonnées introuvables."}


initial_models = list_saved_models()


with gr.Blocks(title="Classification d’images microscopiques") as demo:
    gr.Markdown("# Classification d’images microscopiques de charbons de bois")
    gr.Markdown(
        "Application pédagogique pour explorer un jeu de données d’images microscopiques, "
        "entraîner un modèle de classification et analyser ses performances."
    )

    with gr.Tabs():

        with gr.Tab("1. Explorer le jeu de données"):
            gr.Markdown("## Comprendre le jeu de données avant l’entraînement")

            load_dataset_btn = gr.Button(
                "Charger les informations du dataset",
                variant="primary",
            )

            dataset_summary = gr.JSON(label="Résumé général du dataset")

            class_distribution = gr.Dataframe(
                label="Distribution des images par split et par classe",
                interactive=False,
            )

            gr.Markdown("## Visualisation des images")

            with gr.Row():
                split_selector = gr.Dropdown(
                    choices=["train", "validation", "test"],
                    value="train",
                    label="Split",
                )
                class_selector = gr.Dropdown(
                    choices=["Toutes les classes"],
                    value="Toutes les classes",
                    label="Classe",
                )
                max_images = gr.Slider(
                    minimum=4,
                    maximum=48,
                    value=24,
                    step=4,
                    label="Nombre d’images à afficher",
                )

            refresh_gallery_btn = gr.Button("Afficher des exemples")

            image_gallery = gr.Gallery(
                label="Exemples d’images",
                columns=4,
                height=600,
            )

        with gr.Tab("2. Entraîner un modèle"):
            gr.Markdown("## Entraînement avec ResNet18 pré-entraîné")
            gr.Markdown(
                "Paramètres par défaut recommandés : fine-tuning de la dernière couche convolutionnelle "
                "du ResNet18, faible taux d’apprentissage, augmentation légère des données."
            )

            with gr.Row():
                with gr.Column():
                    dropout = gr.Slider(
                        minimum=0.0,
                        maximum=0.8,
                        value=0.4,
                        step=0.05,
                        label="Dropout",
                    )

                    fc_dim = gr.Dropdown(
                        choices=[64, 128, 256, 512],
                        value=256,
                        label="Dimension de la couche cachée",
                    )

                    learning_rate = gr.Number(
                        value=0.00001,
                        label="Taux d’apprentissage",
                    )

                    weight_decay = gr.Number(
                        value=0.0001,
                        label="Weight decay",
                    )

                    batch_size = gr.Dropdown(
                        choices=[8, 16, 32, 64],
                        value=16,
                        label="Taille du batch",
                    )

                    epochs = gr.Slider(
                        minimum=1,
                        maximum=80,
                        value=30,
                        step=1,
                        label="Nombre d’époques",
                    )

                    fine_tune_mode = gr.Dropdown(
                        choices=["frozen", "layer4", "full"],
                        value="layer4",
                        label="Mode de fine-tuning",
                        info=(
                            "frozen = seul le classifieur est entraîné ; "
                            "layer4 = dernière partie du ResNet18 + classifieur ; "
                            "full = tout le réseau est ajusté."
                        ),
                    )

                    model_tag = gr.Textbox(
                        label="Nom court du modèle",
                        placeholder="ex. charbon_resnet18_layer4",
                    )

                    train_btn = gr.Button("Lancer l’entraînement", variant="primary")

                with gr.Column():
                    train_status = gr.Textbox(
                        label="Journal d’entraînement",
                        lines=18,
                    )
                    train_history = gr.JSON(label="Historique d’entraînement")
                    train_summary = gr.JSON(label="Résumé final")

            gr.Markdown("## Résultats sur le test set")

            train_report = gr.Dataframe(
                label="Rapport de classification",
                interactive=False,
            )

            train_confusion_matrix = gr.Dataframe(
                label="Matrice de confusion",
                interactive=False,
            )

            train_confusion_matrix_image = gr.Image(
                label="Matrice de confusion - figure",
                type="filepath",
            )

        with gr.Tab("3. Tester et analyser un modèle"):
            gr.Markdown("## Sélectionner un modèle sauvegardé")

            with gr.Row():
                with gr.Column():
                    model_selector = gr.Dropdown(
                        choices=initial_models,
                        value=initial_models[0] if initial_models else None,
                        label="Modèle sauvegardé",
                    )

                    refresh_btn = gr.Button("Actualiser la liste des modèles")
                    load_info_btn = gr.Button("Afficher les informations du modèle")
                    model_info = gr.JSON(label="Métadonnées du modèle")

                with gr.Column():
                    evaluate_btn = gr.Button(
                        "Évaluer le modèle sur le test set",
                        variant="primary",
                    )
                    eval_summary = gr.JSON(label="Résumé des métriques")

            eval_report = gr.Dataframe(
                label="Rapport de classification",
                interactive=False,
            )

            eval_confusion_matrix = gr.Dataframe(
                label="Matrice de confusion",
                interactive=False,
            )

            eval_confusion_matrix_image = gr.Image(
                label="Matrice de confusion - figure",
                type="filepath",
            )

            gr.Markdown("## Prédiction sur une image importée")

            with gr.Row():
                with gr.Column():
                    upload_image = gr.Image(type="pil", label="Importer une image")
                    predict_btn = gr.Button("Prédire la classe", variant="primary")

                with gr.Column():
                    predict_text = gr.Textbox(label="Résultat de la prédiction", lines=7)
                    predict_probs = gr.Label(label="Probabilités par classe")

            gr.Markdown("## Test sur un échantillon aléatoire du test set")

            random_test_btn = gr.Button("Tester un échantillon aléatoire")

            with gr.Row():
                random_sample_image = gr.Image(type="pil", label="Image test aléatoire")
                random_sample_text = gr.Textbox(label="Résultat sur l’échantillon", lines=7)
                random_sample_probs = gr.Label(label="Probabilités par classe")

    load_dataset_btn.click(
        fn=load_dataset_overview_callback,
        inputs=None,
        outputs=[dataset_summary, class_distribution, class_selector],
    )

    refresh_gallery_btn.click(
        fn=refresh_gallery_callback,
        inputs=[split_selector, class_selector, max_images],
        outputs=image_gallery,
    )

    train_btn.click(
        fn=train_callback,
        inputs=[
            dropout,
            fc_dim,
            learning_rate,
            weight_decay,
            batch_size,
            epochs,
            fine_tune_mode,
            model_tag,
        ],
        outputs=[
            train_status,
            train_history,
            train_summary,
            train_report,
            train_confusion_matrix,
            train_confusion_matrix_image,
            model_selector,
        ],
    )

    refresh_btn.click(
        fn=refresh_models_dropdown,
        inputs=None,
        outputs=model_selector,
    )

    load_info_btn.click(
        fn=get_model_info,
        inputs=model_selector,
        outputs=model_info,
    )

    evaluate_btn.click(
        fn=evaluate_saved_model_callback,
        inputs=model_selector,
        outputs=[
            eval_summary,
            eval_report,
            eval_confusion_matrix,
            eval_confusion_matrix_image,
        ],
    )

    predict_btn.click(
        fn=predict_uploaded_image_callback,
        inputs=[model_selector, upload_image],
        outputs=[predict_text, predict_probs],
    )

    random_test_btn.click(
        fn=test_random_sample_callback,
        inputs=model_selector,
        outputs=[random_sample_image, random_sample_text, random_sample_probs],
    )


if __name__ == "__main__":
    demo.launch(ssr_mode=False)