app.py

import json
import os
import re
from datetime import datetime
from pathlib import Path

import gradio as gr
from huggingface_hub import CommitScheduler, HfApi
from huggingface_hub.utils import HfHubHTTPError


HF_TOKEN = os.getenv("HF_TOKEN")

JSON_DATASET_DIR = Path("dataset")
JSON_DATASET_DIR.mkdir(parents=True, exist_ok=True)
JSON_DATASET_PATH = JSON_DATASET_DIR / "dataset.jsonl"

scheduler = CommitScheduler(
    repo_id="librarian-bots/collection_cloner-usage-stats",
    repo_type="dataset",
    folder_path=JSON_DATASET_DIR,
    path_in_repo=str(JSON_DATASET_PATH),
    token=HF_TOKEN,
)


def save_json(source_slug: str, destination_slug: str) -> None:
    with scheduler.lock:
        with JSON_DATASET_PATH.open("a") as f:
            if source_slug.startswith("hf_"):  # catch people accidentally adding tokens
                return None
            if destination_slug.startswith("hf_"):
                return None
            json.dump(
                {
                    "source_collection": source_slug,
                    "destination_collection": destination_slug,
                    "datetime": datetime.now().isoformat(),
                },
                f,
            )
            f.write("\n")


def extract_slug(url):
    pattern = r"https://huggingface\.co/collections/(.*)"
    return match.group(1) if (match := re.search(pattern, url)) else None


def clone_collection(
    source_slug, dest_title, token, dest_namespace=None, private=False, exist_ok=False
):
    api = HfApi(token=token)
    source_slug = source_slug.strip()
    # check if formatted as url
    if source_slug.startswith("https://huggingface.co./collections/"):
        source_slug = extract_slug(source_slug)
    collection = api.get_collection(source_slug)
    if not collection:
        raise gr.Error(
            f"Collection {source_slug} does not exist or you do not have access to it."
        )
    description = f"Copied from {collection.title} using https://huggingface.co./spaces/librarian-bots/collection_cloner."
    if dest_namespace == "username":
        dest_namespace = None
    new_collection = api.create_collection(
        dest_title,
        namespace=dest_namespace,
        exists_ok=exist_ok,
        private=private,
        description=description,
        token=token,
    )
    for item in collection.items:
        try:
            api.add_collection_item(
                new_collection.slug, item.item_id, item_type=item.item_type
            )
        except HfHubHTTPError as e:
            gr.Info(
                f"Failed to add item {item.item_id} to collection {new_collection.slug} because it already exists in this collection."
            )
    if not private:
        save_json(collection.slug, new_collection.slug)
    return f"[Collection]({collection.url}) has been cloned into [{new_collection.slug}]({new_collection.url})"


title = (
    """<h1 style='text-align: center;'> &#129516; Collection Cloner &#129516;</h1>"""
)


with gr.Blocks(css="style.css") as demo:
    gr.HTML(title)
    gr.HTML(
        """<p style='text-align: center;'>
This space allows you to clone a <a href="https://huggingface.co./docs/hub/collections">Collection</a> from the Hugging Face Hub into your own namespace.<p>
<p style='text-align: center;'> You can edit this cloned Collection to your liking!</p>"""
    )
    gr.Markdown(
        """
                **Note**: To track interest in this feature this Space keeps a record of clones which are cloned into public collection. Clones into private Collections are not tracked."""
    )
    gr.Markdown("## Authentication")
    gr.Markdown(
        "Token is required to create a new collection and clone private collections. You can get your token from your [profile page](https://huggingface.co./settings/token)."
    )
    with gr.Row():
        token = gr.Textbox(
            label="Token",
            type="password",
        )
    with gr.Column():
        gr.Markdown("## Source Collection")
        source_slug = gr.Textbox(
            label="Source Collection slug or URL",
            placeholder="e.g. username/collection-slug",
        )
        gr.Markdown("## Destination Collection info")

        dest_title = gr.Textbox(
            label="Destination Title",
        )
        dest_namespace = gr.Textbox(
            value="username",
            label="Destination Namespace (optional - defaults to your username))",
            interactive=True,
        )
        with gr.Row():
            private = gr.Checkbox(
                False,
                label="Make new collection private?",
            )
            overwrite = gr.Checkbox(
                False,
                label="Overwrite any collection with same slug as the destination?",
            )
    submit_btn = gr.Button("Clone Collection")
    response = gr.Markdown()
    submit_btn.click(
        clone_collection,
        [
            source_slug,
            dest_title,
            token,
            dest_namespace,
            private,
            overwrite,
        ],
        response,
    )

demo.launch()