TypeError: unsupported operand type(s) for +

[FloWuenne]

Hi Cylinter team,

installed Cylinter in a fresh conda environment on my Macbook today and wanted to test with the example dataset.
I downloaded the example dataset using the following demo command:

prep cylinter_demo ~/Desktop/cylinter_demo  # Enter Synapse ID and password when prompted.

When I try to run the example data using the following command, I get the error below. Trying with a specific module gives the same error.

Thanks for the help!

cylinter ~/Desktop/cylinter_demo/config.yml

Traceback (most recent call last):
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/bin/cylinter", line 6, in <module>
    from cylinter.cylinter import main
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/lib/python3.10/site-packages/cylinter/cylinter.py", line 7, in <module>
    from . import pipeline, components
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/lib/python3.10/site-packages/cylinter/pipeline.py", line 4, in <module>
    from . import components
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/lib/python3.10/site-packages/cylinter/components.py", line 15, in <module>
    import hdbscan
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/lib/python3.10/site-packages/hdbscan/__init__.py", line 1, in <module>
    from .hdbscan_ import HDBSCAN, hdbscan
  File "/Users/florian_wuennemann/miniconda3/envs/cylinter/lib/python3.10/site-packages/hdbscan/hdbscan_.py", line 40, in <module>
    FAST_METRICS = KDTree.valid_metrics + BallTree.valid_metrics + ["cosine", "arccos"]
TypeError: unsupported operand type(s) for +: 'builtin_function_or_method' and 'builtin_function_or_method'

[gjbaker]

Hello and thank you for your comment!

We have recently observed this error and believe it is caused by an update to the scikit-learn API that has made it incompatible with the density-based clustering package used by CyLinter (HDBSCAN). Running (pip install 'scikit-learn<1.3.0') should fix the problem. We will soon be pushing an updated version of the conda-build recipe that takes care of the issue at the point of initial installation.

[FloWuenne]

That fixed it! Thank you very much for the rapid response @gjbaker ! 😊