Add get_bytes_range() function
#196
Conversation
|
|
||
| # If we're using the /a/b/foo.zip!c/d/file.txt syntax, handle it here. | ||
| exclamation_index = url_or_filename.find("!") | ||
| if extract_archive and exclamation_index >= 0: |
There was a problem hiding this comment.
if extract_archive is true and there is no !, we should probably return some sort of error/exception
There was a problem hiding this comment.
It's possible it could be part of the filename though, no? I think this is how cached_path() behaves so I wanted to match the behavior.
There was a problem hiding this comment.
What if an archive file has an ! in its name? It's an edge case but would cause things to break. Having a separate optional parameter for the relative path of the file within the archive would be a way to get around the problem
There was a problem hiding this comment.
Having a separate optional parameter for the relative path of the file within the archive would be a way to get around the problem
I agree but I don't want to deviate from the API of cached_path(). At least for now.
We use a function like this a lot in the LLM repo so I thought we should add it here.