Illegal instruction in test code

[KANG-BIOINFO]

It seems there's some weird bug in test.py code. Could you take a look and see why it happens? I ran it on my Mac:

P14-5847:sctriangulate_test kang$ /Users/kang/miniconda3/envs/aipy/bin/python3.7 -W ignore mini_test.py 
===================
import modules test: ✓ 
====================
Preprocessing...
Simulating doublets...
Embedding transcriptomes using PCA...
Calculating doublet scores...
Illegal instruction: 4

[frankligy]

It seems to be a conda version issue, mine is conda 4.10.1 and my mac version is 11.4 See (https://github.com/conda/conda/issues/9678)

I carefully compare our dependencies difference, and they are the same, so the dependencies are not the culprit here. dependency_mac.txt

I will keep looking at this issue, and in the meantime, I would suggest a version update on your conda environment.

[nsalomonis]

Has it been tested in a virtual environment with PyPI without conda (local and cluster)?

On Sun, Oct 10, 2021 at 3:21 PM Guangyuan(Frank) Li < @.***> wrote:

It seems to be a conda version issue, mine is conda 4.10.1 and my mac version is 11.4 See (conda/conda#9678 https://github.com/conda/conda/issues/9678)

I carefully compare our dependencies difference dependency_mac.txt https://github.com/frankligy/scTriangulate/files/7318076/dependency_mac.txt

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/frankligy/scTriangulate/issues/3#issuecomment-939538084, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABD5TSOSSNRZI6CU57AWRRLUGHRUDANCNFSM5FWVDNMQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

[frankligy]

Hello @KANG-BIOINFO just want to give you an update here, as I realize that the another potential cause is due to scrublet package, look at this issue (https://github.com/swolock/scrublet/issues/28). I experienced that same issue when using annoy==1.7.0, downgrading to 1.6.3 indeed help me to pass through this hurdle.

With that being said, I used annoy==1.7.0 in my python=3.6 environment and it works well, so it is still mysterious to me at this time point that I am going to keep looking into it.

Another reason could be the RAM is exceeded.

[KANG-BIOINFO]

I see. Thanks for the updates. Let me know if you want to test on other versions of python (3.7 or higher).

[Bon-jour]

I updated annoy-1.15.1 to solve this problem, my python version is Python 3.9.16, scanpy 1.9.3