app/app_savta.py

from pathlib import Path

import gradio as gr
from fastai.vision.all import *

#######################
#   Data & Learner    #
#######################


class ImageImageDataLoaders(DataLoaders):
    """Create DataLoaders for image→image tasks."""

    @classmethod
    @delegates(DataLoaders.from_dblock)
    def from_label_func(
        cls,
        path: Path,
        filenames,
        label_func,
        valid_pct: float = 0.2,
        seed: int | None = None,
        item_transforms=None,
        batch_transforms=None,
        **kwargs,
    ):
        dblock = DataBlock(
            blocks=(ImageBlock(cls=PILImage), ImageBlock(cls=PILImageBW)),
            get_y=label_func,
            splitter=RandomSplitter(valid_pct, seed=seed),
            item_tfms=item_transforms,
            batch_tfms=batch_transforms,
        )
        return cls.from_dblock(dblock, filenames, path=path, **kwargs)


def get_y_fn(x: Path) -> Path:
    """Return same image as label for architecture initialization."""
    return x


def create_data(data_path: Path):
    """Create minimal data loader for model architecture initialization."""
    fnames = get_files(data_path, extensions=".jpg")
    return ImageImageDataLoaders.from_label_func(
        data_path,
        seed=42,
        bs=1,
        num_workers=0,
        valid_pct=0.0,
        filenames=fnames,
        label_func=get_y_fn,
    )


# Initialize learner with architecture
data = create_data(Path("examples"))
learner = unet_learner(
    data,
    resnet34,
    n_out=3,
    loss_func=MSELossFlat(),
    path=".",
    model_dir="models",
)
learner.load("model")

#####################
#  Inference Logic  #
#####################


def predict_depth(input_img: PILImage) -> PILImageBW:
    depth, *_ = learner.predict(input_img)
    return PILImageBW.create(depth).convert("L")


#####################
#    Gradio UI      #
#####################

title = "📷 SavtaDepth WebApp"

description_md = """
    <p style="text-align:center;font-size:1.05rem;max-width:760px;margin:auto;">
        Upload an RGB image on the left and get a grayscale depth map on the right.
    </p>
    """

footer_html = """
    <p style='text-align:center;font-size:0.9rem;'>
        <a href='https://dagshub.com/OperationSavta/SavtaDepth' target='_blank'>Project on DAGsHub</a> •
        <a href='https://colab.research.google.com/drive/1XU4DgQ217_hUMU1dllppeQNw3pTRlHy1?usp=sharing' target='_blank'>Google Colab Demo</a>
    </p>
    """

examples = [["examples/00008.jpg"], ["examples/00045.jpg"]]

input_component = gr.Image(width=640, height=480, label="Input RGB")
output_component = gr.Image(label="Predicted Depth", image_mode="L")

with gr.Blocks(title=title, theme=gr.themes.Soft()) as demo:
    gr.Markdown(f"<center><h1>{title}</h1></center>")
    gr.HTML(description_md)

    gr.Interface(
        fn=predict_depth,
        inputs=input_component,
        outputs=output_component,
        examples=examples,
        cache_examples=False,
    )

    gr.HTML(footer_html)

if __name__ == "__main__":
    demo.queue().launch()