🐛 Bug: index_parquet_dataset() fails with TypeError when passing storage_options for S3

ld.index_parquet_dataset() fails with TypeError: filesystem() takes 1 positional argument but 4 were given when passing storage_options for S3. The bug is in litdata/utilities/parquet.py — it unpacks storage_options as positional arguments instead of keyword arguments.

To Reproduce

Steps to reproduce the behavior:

1. Install litdata==0.2.61
2. Call ld.index_parquet_dataset() with an S3 path and a storage_options dict

import litdata as ld

pq_dataset_uri = 's3://bucket/path/to/parquet/'

ld.index_parquet_dataset(
    pq_dataset_uri,
    storage_options={
        "key": "ACCESS_KEY",
        "secret": "SECRET_KEY",
        "endpoint_url": "https://s3.example.com",
    },
)

Error:

TypeError: filesystem() takes 1 positional argument but 4 were given

Stack trace excerpt:

File ~/.local/lib/python3.11/site-packages/litdata/utilities/parquet.py:150, in filesystem()
    150 self.fs = fsspec.filesystem(provider, *self.storage_options)

Root Cause

The bug is in litdata/utilities/parquet.py line ~150:

# BUG: unpacks dict keys as positional args
self.fs = fsspec.filesystem(provider, *self.storage_options)

Should be:

# FIX: pass as keyword arguments
self.fs = fsspec.filesystem(provider, **self.storage_options)

Expected behavior

The function should pass storage_options as keyword arguments to fsspec.filesystem(), allowing S3 and other remote filesystems to be initialized correctly.

Additional context

This bug affects all users trying to use index_parquet_dataset() with S3 or any filesystem requiring authentication options.