csv_provider.py

import asyncio
import pandas as pd
from huggingface_hub import HfApi, hf_hub_download
from mcp.server import Server
from mcp.types import Resource, ResourceResponse
import os

server = Server("csv-provider")

HF_TOKEN = os.environ.get("HF_TOKEN")
api = HfApi()

# Resource: csv://repo_id/path/to/file.csv
@server.resource("csv://{repo_id}/{path}")
async def get_csv(repo_id: str, path: str) -> ResourceResponse:
    try:
        # Télécharger le fichier CSV depuis le dataset privé
        local_path = hf_hub_download(
            repo_id=repo_id,
            filename=path,
            repo_type="dataset",
            token=HF_TOKEN
        )

        # Charger et prévisualiser le CSV
        df = pd.read_csv(local_path)
        preview = df.head(5).to_string()

        return ResourceResponse(
            resource=Resource(
                uri=f"csv://{repo_id}/{path}",
                name=f"CSV: {path}",
                description=f"Aperçu des données du dataset {repo_id}",
                mimeType="text/plain",
                text=preview,
            )
        )
    except Exception as e:
        return ResourceResponse(
            resource=Resource(
                uri=f"csv://{repo_id}/{path}",
                name=f"Erreur CSV",
                description=str(e),
                mimeType="text/plain",
                text=f"Impossible de lire {repo_id}/{path} : {e}",
            )
        )

if __name__ == "__main__":
    asyncio.run(server.run())