"""
Data loading utilities
"""

import json
from datasets import load_dataset


def load_dataset_from_jsonl(file):
    """Charge un dataset depuis un fichier JSONL"""
    data = []
    content = file.getvalue().decode('utf-8')
    for line in content.split('\n'):
        if line.strip():
            data.append(json.loads(line))
    return data


def load_dataset_from_hf(dataset_name, split='train'):
    """Charge un dataset depuis HuggingFace"""
    dataset = load_dataset(dataset_name, split=split)
    return list(dataset)


def filter_items_with_positive(dataset):
    """Filtre les items qui ont un champ 'positive' non vide"""
    items_with_positive = []
    for idx, item in enumerate(dataset):
        if 'positive' in item and item['positive']:
            items_with_positive.append((idx, item))
    return items_with_positive