I changed pheno_jaccard_ig to be a csr matrix so it is easier to save and manipulate - before it was a COO matrix which threw some errors.

[dshefter]

[codecov[bot]]

Codecov Report

Merging #2295 (1f7fc46) into master (ab2177c) will decrease coverage by 0.06%. The diff coverage is 50.00%.

:exclamation: Current head 1f7fc46 differs from pull request most recent head 63142f7. Consider uploading reports for the commit 63142f7 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2295      +/-   ##
==========================================
- Coverage   71.88%   71.82%   -0.06%     
==========================================
  Files          98       98              
  Lines       11509    11540      +31     
==========================================
+ Hits         8273     8289      +16     
- Misses       3236     3251      +15     

Impacted Files	Coverage Δ
scanpy/external/tl/_phenograph.py	28.57% <50.00%> (+4.32%)	:arrow_up:

... and 41 files with indirect coverage changes

[Zethson]

Hi,

could you please provide more details? What issues did you run into?

[dshefter]

Hi,

The issues I was mostly running into were that when saving the anndata variable as a h5ad file, 'pheno_jaccard_ig' was not compatible with this action. So, I had to either remove pheno_jaccard_ig from the anndata object and then save it as h5ad or convert it to a sparse matrix. This also happened with a few other functions I tried on the anndata object, and I kept getting the error "this function is not compatible with COO matrix format", always talking about pheno_jaccard_ig. Therefore, since a sparse matrix object does not have any problems with the functions I was running on adata, changing pheno_jaccard_ig to a sparse matrix from the start makes sense to circumvent any of those issues I was getting before.

I hope this makes sense. Thank you, Deena Shefter

On Wed, Jul 27, 2022 at 6:10 PM Lukas Heumos @.***> wrote:

Hi,

could you please provide more details? What issues did you run into?

— Reply to this email directly, view it on GitHub https://github.com/scverse/scanpy/pull/2295#issuecomment-1197424392, or unsubscribe https://github.com/notifications/unsubscribe-auth/AMILUOEK7J64GU3YOR7DB53VWGXULANCNFSM534YT5ZA . You are receiving this because you authored the thread.Message ID: @.***>