Is Metrics.evaluate's `use_cache=True` default setting intentional?

[djstreat]

Question

I'm wondering if the use_cache setting on Metrics.evaluate is intented to on be default.

Further Information

Hi all. Just ran into an issue when using synthcity within a jupyter notebook kernel. Didn't realize that the function is set to use previously cached results instead of rerunning the program when called by default. It makes some sense, but was somewhat confusing to find out changes I was making earlier in my notebook were not propagating to the results I was being shown when running cells later down the line.

It might just be that my particular use case isn't common (using synthcity within a notebook), but I wanted to make sure it was intentional behavior.

Screenshots

N/a

System Information

• OS: Windows 11
• OS Version: 10.0.26100 Build 26100
• Language Version: Python 3.11.9
• Package Manager Version: uv 0.8.22

Additional Context

Add any other context about the problem here.