init_dataset.py

from huggingface_hub import create_repo, HfApi
from datasets import Dataset
import pandas as pd
import os

DATASET_REPO = "Sergidev/PMBMemory"

def init_dataset():
    # Check if dataset exists
    api = HfApi()
    try:
        api.repo_info(repo_id=DATASET_REPO, repo_type="dataset")
        print(f"Dataset {DATASET_REPO} already exists.")
    except Exception:
        print(f"Creating dataset {DATASET_REPO}...")
        create_repo(repo_id=DATASET_REPO, repo_type="dataset")

        # Create an empty dataframe with the required columns
        df = pd.DataFrame(columns=["timestamp", "prompt", "response", "topic"])

        # Convert to dataset and push to hub
        dataset = Dataset.from_pandas(df)
        dataset.push_to_hub(DATASET_REPO)

        print(f"Dataset {DATASET_REPO} created successfully.")

if __name__ == "__main__":
    init_dataset()