Closed zcalhoun closed 7 months ago
We are looking into it. Thanks for raising an issue! For background, are you using the Github code or the PyPI package?
I tried using both the PyPI package and the Github code.
@zcalhoun The PyPI version needs to be updated since it has the error but we don't see an error when installing directly from the repository:
pip install "git+https://github.com/NSAPH-Projects/space@dev#egg=spacebench[all]"
We will update the PyPI version, in the meantime, can you uninstall your current version (pip uninstall spacebench
) and install from Github as above? If the problem persists, please let us know and include a code sample and information about your system so we can reproduce the error.
A new version in PyPI v0.1.4 has been released.
Hi Mauricio,
Thank you for following up. I am still seeing the issue when I try to load the datasets. I think the issue is with this line of code: https://github.com/NSAPH-Projects/space/blob/ad1630038abee4fdab3080f25fba30bd4a4df85c/spacebench/env.py#L296
Basically, you should replace the 1:
with -1
, so that you are always getting the file type. This error stems from the fact that I was using a relative reference with ./
, which doesn't work with the way you've written this line of code. I can make the change and issue a pull request if you'd like.
I can make the change and issue a pull request if you'd like.
@zcalhoun That would be fantastic! We have to be careful of the case when the extension is .tar.gz
, in which [-1]
alone wouldn't work. Maybe we can use .endswith()
and define ext
inside the if-else statement?
I see what you mean in the code -- there are two separate lines of code that use similar logic to load the .csv/.tab/.parquet files, and then a separate set of code to open up graphml/tar.gz files. I made updates to both that should address the issue I encountered (see PR #151).
First off - thank you for putting together this repository. I am looking forward to seeing this project succeed.
I had a question about loading the data -- when I tried to follow your instructions under "Getting Started," I get an error "No such file or directory [...] /healthd_dmgrcs_mortality_disc/synthetic_data.csv." I looked at the loaded data, and sure enough, the data was loaded, but it looks like the synthetic data is a .tab file. This seemed like an error, and I was wondering whether I could still trust the generated data?