SmartCache for `PatchWSIDataset`

[bhashemian]

Is your feature request related to a problem? Please describe.
Add SmartCache functionality for the new PatchWSIDataset. #4152 (comment)

[Nic-Ma]

Hi @drbeh ,

Thanks for raising this up, I think we discussed in #4152 (comment) that it's not necessary to implement this SmartCachePatchWSIDataset anymore.
You can easily use 2 datasets:

dataset = PatchWSIDataset(...)
cache = SmartCacheDataset(data=dataset, ...)

What do you think?

Thanks.

[bhashemian]

Oh @Nic-Ma! I think there were some misunderstandings. I thought you were saying that we must implement SmartCachePatchWSIDataset! and you liked my next comments that I said I will add SamrtCache in the following PR: #4152 (comment)

So you meant that we had to implement it before but it is not necessary anymore, right?!

P.S. The confusion comes from using "must" instead of "had to"! ;)

[Nic-Ma]

Sorry man, maybe I didn't make myself clear in that comment.
Yes, we had to implement in Clara Train 4.0 before, but I think it's not necessary now.
I liked your next comments because I thought you will add the use case of 2 connected datasets in the tutorial or unit test.

Thanks.

[bhashemian]

That's fine! so we can close this.

[Nic-Ma]

Yeah, let's close it for now.
Maybe we will reopen this PR for some unknown cases in the future.

Thanks.