feature index mismatch

[EnnyvanBeest]

Hi, I'm not sure if this is the right place to issue this as the issue arises when trying to prepare kilosort/phy output for the IBL pipeline as described here. But as the problem arises only when having already curated the data with phy2, I was hoping you have a solution/answer.

If I prepare data for the IBL pipeline directly from kilosort (v2.5) output, it works correctly. However, if I do this for already curated data (by phy2.0), I get an indexing error (model.py / memmap.py) specifically for the files []_feature[].npy; it seems like these files are incorrectly updated based on the (new) number of clusters.

I can circumvent this by first converting to IBL (for histology alignment) and only after that using phy(2) to curate, but could you look into if these feature files are updated correctly after going through phy2? It may cause additional problems when wanting to look into feature-space at a stage post-curating.

Thanks in advance!