Intro example not working

[XenonLamb]

I tried to run the introduction notebook, but encountered the following error:

`---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
Cell In[3], line 1
----> 1 import mlcroissant as mlc
3 # FileObjects and FileSets define the resources of the dataset.
4 distribution = [
5 # gpt-3 is hosted on a GitHub repository:
6 mlc.FileObject(
(...)
22 ),
23 ]

File ~/.local/lib/python3.9/site-packages/mlcroissant/init.py:3
1 """Defines the public interface to the mlcroissant package."""
----> 3 from mlcroissant._src import torch
4 from mlcroissant._src.core import constants
5 from mlcroissant._src.core.constants import DataType

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/torch/init.py:3
1 """PyTorch utilities public API."""
----> 3 from mlcroissant._src.torch.torch_adapter import LoaderFactory
4 from mlcroissant._src.torch.torch_adapter import LoaderSpecificationDataType
6 all = ["LoaderFactory", "LoaderSpecificationDataType"]

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/torch/torch_adapter/init.py:3
1 """PyTorch dataloader-based public API."""
----> 3 from mlcroissant._src.torch.torch_adapter.dataloader import LoaderFactory
4 from mlcroissant._src.torch.torch_adapter.dataloader import LoaderSpecificationDataType
6 all = ["LoaderFactory", "LoaderSpecificationDataType"]

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/torch/torch_adapter/dataloader.py:11
8 from typing import Any, Dict, Optional
10 from mlcroissant._src.core.optional import deps
---> 11 from mlcroissant._src.datasets import Dataset
13 try:
14 dp = deps.torchdata_datapipes

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/datasets.py:11
8 from absl import logging
9 from etils import epath
---> 11 from mlcroissant._src.core.context import Context
12 from mlcroissant._src.core.graphs import utils as graphs_utils
13 from mlcroissant._src.core.issues import ValidationError

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/core/context.py:12
9 from etils import epath
10 import networkx as nx
---> 12 from mlcroissant._src.core.issues import Issues
13 from mlcroissant._src.core.rdf import Rdf
16 class CroissantVersion(enum.Enum):

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/core/issues.py:16
11 class GenerationError(Exception):
12 """Error during the generation of the dataset."""
15 @dataclasses.dataclass(frozen=True)
---> 16 class Issues:
17 """Issues during the validation of the format.
18
19 Issues can either be errors (blocking) or warnings (informative).
20
21 We use sets to represent errors and warnings to avoid repeated strings.
22 """
24 _errors: set[tuple[str, Any]] = dataclasses.field(default_factory=set, hash=False)

File ~/.local/lib/python3.9/site-packages/mlcroissant/_src/core/issues.py:27, in Issues()
24 _errors: set[tuple[str, Any]] = dataclasses.field(default_factory=set, hash=False)
25 _warnings: set[tuple[str, Any]] = dataclasses.field(default_factory=set, hash=False)
---> 27 def _wrap_in_context(self, context: str | None, issue: str) -> str:
28 if context is None:
29 return issue

TypeError: unsupported operand type(s) for |: 'type' and 'NoneType'`

[luisoala]

hi @XenonLamb thx for posting

i just reran all cells of the notebook on colab and it worked fine on my end

can you reproduce your result on colab https://githubtocolab.com/mlcommons/croissant/blob/main/python/mlcroissant/recipes/introduction.ipynb ?

cc #694

[brendon-boldt]

Python version issue, most likely. X | Y typing syntax is supported from 3.10 onward (https://peps.python.org/pep-0604/), and the stack trace is from Python 3.9.

[XenonLamb]

Python version issue, most likely. X | Y typing syntax is supported from 3.10 onward (https://peps.python.org/pep-0604/), and the stack trace is from Python 3.9.

Yes after running on a python 3.10 kernel, the example seems to work. Thank you for the clarification!