Closed ethanbb closed 3 months ago
By the way, did the lockfile stuff in your other branch work?
By the way, did the lockfile stuff in your other branch work?
Probably, still doing some testing though.
is this good to go to merge?
is this good to go to merge?
Yes it should be!
This extends the function
get_params_diffs
to work when the params are nested under different groups (as caiman now expects) and/or include types that are comparable but not hashable, such as lists. (The hashability issue comes from creating aset
, which requires elements to be hashable; after playing a bit with converting non-hashable types to hashable ones, I decided to bypass that entirely by just using a different method based on equality testing.)I had to make these changes to get it to work with my existing code, where I am exporting the entire params dict from a CNMFParams object and modifying it as needed before passing to
df.caiman.add_item
. I figured it could help others to make this function more robust.