Multi30k mocked testing #1554
Multi30k mocked testing #1554
Conversation
| with open(txt_file, "w") as f: | ||
| for i in range(5): | ||
| rand_string = " ".join( | ||
| random.choice(string.ascii_letters) for i in range(seed) | ||
| ) |
There was a problem hiding this comment.
One thought: since all of our datasets are utf-8 files, does it make sense to write unicode strings to make sure we don't have lingering bugs from default encodings when opening files? Maybe this is overkill, but it's been a big source of bugs when I did mostly windows development.
There was a problem hiding this comment.
Ya, agreed. I think it is a good suggestion!
There was a problem hiding this comment.
I will keep it as a follow-up item as generating random UTF-8 string is not trivial.
test/datasets/test_multi30k.py
Outdated
| rand_string = " ".join( | ||
| random.choice(string.ascii_letters) for i in range(seed) | ||
| ) | ||
| content = f'{rand_string}\n' |
There was a problem hiding this comment.
nit: can we use double quotes here instead of single quotes
Added this as a follow up item to the GH issue! |
4 participants
Reference issue #1493