Difference in numbers from condastats and anaconda.org

[nicrie]

• condastats version: 0.2.1
• Python version: 3.11.4
• Operating System: Ubuntu 22.04

Description

Getting the total/monthly download number from condastats for a package (xeofs) differs from the number you find on the corresponding anaconda website

What I Did

condastats overall xeofs

The above command results in 1191, whereas the anaconda website states 4091. Where does the difference come from?

btw: Thanks for this super cool tool! :)

[sophiamyang]

Sorry for the late reply. @seibert could you help with this question? Thanks!

[phwuil]

Hi @nicrie and @sophiamyang , it is quite frustrating to see that @nicrie and I have quite the same configuration, and he can get stats from condastats and I cannot :-) (see #17) ... Can we try to find the difference between our configurations ?

• condastats==0.2.1

• pandas==2.1.1

• pyarrow==13.0.0

• dask==2023.9.3

• python 3.11.5

• linux (manjaro 23.0)

(what else ?)

[jakirkham]

@jezdez would it make sense to transfer this to https://github.com/ContinuumIO/anaconda-package-data ?

[jakirkham]

@jezdez think running this gh command should work

gh issue transfer https://github.com/conda-incubator/condastats/issues/18 https://github.com/ContinuumIO/anaconda-package-data

Unfortunately lack the permissions to complete it

[jezdez]

@jakirkham IIUC we can't move this ticket to the other repo, since it's not stored in the same organization sadly, per https://docs.github.com/en/issues/tracking-your-work-with-issues/transferring-an-issue-to-another-repository

[nicrie]

In that case, would you advice me to open another issue to anaconda-package-data?

[jakirkham]

Thanks Jannis! 🙏

Yes @nicrie could you please raise a new issue on ContinuumIO/anaconda-package-data? Once open please link it here and close this one out

Edit: Admittedly there is some overlap with issue ( ContinuumIO/anaconda-package-data#52 ). So it may make sense to consolidate there. Please read it over and comment over there if you agree

[nicrie]

Thanks, @jakirkham! I hadn't seen that thread before. It looks like the underlying issue (linked in the thread you shared) is discussed here:

ContinuumIO/anaconda-package-data#45 (comment)

They made a change in June March 2024, and now the download stats seem more accurate for me. I'll keep an eye on that discussion. Thanks for pointing me in the right direction and for your patience!

[jakirkham]

Yes there is some work ongoing to improve the stats, which was part of my interest in rolling this into that work (it also looked related to other issues we have seen). Glad we are already tracking it in the existing discussion.

Thanks for checking! 🙏