app.py

import os
import re

import gradio as gr
from huggingface_hub import get_collection

os.environ["COLLECTION_SLUG_OR_URL"] = "https://huggingface.co/collections/PleIAs/finance-commons-66925e1095c7fa6e6828e26c"

def extract_collection_id(input_text):
    if url_match := re.match(r"https://huggingface\.co/collections/(.+)$", input_text):
        return url_match[1]

    # Check if input is already in the correct format
    return input_text if re.match(r"^[\w-]+/[\w-]+", input_text) else None


def load_collection():
    collection_input = os.getenv("COLLECTION_SLUG_OR_URL")
    if not collection_input:
        raise ValueError("COLLECTION_SLUG_OR_URL environment variable is not set.")

    collection_id = extract_collection_id(collection_input)
    if not collection_id:
        raise ValueError(
            "Invalid collection ID or URL in COLLECTION_SLUG_OR_URL environment variable."
        )

    collection = get_collection(collection_id)
    if dataset_ids := [
        item.item_id for item in collection.items if item.item_type == "dataset"
    ]:
        return dataset_ids, collection_id
    else:
        raise ValueError("No datasets found in this collection.")


def display_dataset(dataset_ids, index):
    dataset_id = dataset_ids[index]
    return gr.HTML(f"""<iframe
    src="https://huggingface.co/datasets/{dataset_id}/embed/viewer"
    frameborder="0"
    width="100%"
    height="560px"
></iframe>""")


def navigate_dataset(dataset_ids, index, direction):
    new_index = (index + direction) % len(dataset_ids)
    return (
        new_index,
        f"Dataset {new_index + 1} of {len(dataset_ids)}: {dataset_ids[new_index]}",
    )


def get_display_name(collection_id):
    # Pattern to match username/repo-name with an optional ID of 16 or more hexadecimal characters
    pattern = r"^(.+?)-([a-f0-9]{16,})$"
    if match := re.match(pattern, collection_id):
        return match[1]
    else:
        # If no match, return the original
        return collection_id


try:
    dataset_ids, collection_id = load_collection()
    display_name = get_display_name(collection_id)

    with gr.Blocks() as demo:
        gr.Markdown(f"<h1>Finance Commons</h1>")
        gr.Markdown(
            f"[View full collection on Hugging Face](https://huggingface.co/collections/{collection_id})"
        )

        gr.Markdown("""
        This app allows you to browse Finance Commons, the largest collection of open datasets on finance documents.""")

        index_state = gr.State(value=0)

        with gr.Row():
            left_btn = gr.Button("Previous")
            right_btn = gr.Button("Next")

        dataset_info = gr.Markdown(f"Dataset 1 of {len(dataset_ids)}: {dataset_ids[0]}")
        iframe_output = gr.HTML()
        gr.Markdown("""**Note**: This space currently mostly supports text formats.
        """)
        left_btn.click(
            navigate_dataset,
            inputs=[gr.State(dataset_ids), index_state, gr.Number(-1, visible=False)],
            outputs=[index_state, dataset_info],
        )
        right_btn.click(
            navigate_dataset,
            inputs=[gr.State(dataset_ids), index_state, gr.Number(1, visible=False)],
            outputs=[index_state, dataset_info],
        )

        index_state.change(
            display_dataset,
            inputs=[gr.State(dataset_ids), index_state],
            outputs=[iframe_output],
        )

        # Initialize the display with the first dataset
        demo.load(
            fn=lambda: display_dataset(dataset_ids, 0),
            inputs=None,
            outputs=[iframe_output],
        )

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

except Exception as e:
    print(f"Error: {str(e)}")
    print(
        "Please set the COLLECTION_SLUG_OR_URL environment variable with a valid collection ID or URL."
    )