Dominik Kutra

I know it's a draft, so just a quick questions: the `--skip-deglobbing` parameter would only be applicable if there were a file with the exact name as the glob expression,...

Option 3 ftw ;)

| option | votes | |:---:|:---:| |1|0| |2|1| |3|2| |4|0| so far it looks like option 3, but @Tomaz-Vieira has sneakily not voiced a strong opinion yet.

Heyo, got a quick glimpse at the code which looks great and then decided, before going into details there, to look at it from a user perspective... Some of the...

okay I could not reproduce it, and there has been a lot going on since 2017... @chaubold, it's a long shot, but would you happen to still have this data,...

> Some objections against dictionaries are: > > * the set of keys is not obvious (nor is their case sensitiveness); > > * The meaning of missing keys might...

How about we hide that from users and offer mypy compatibility: ```Python RoiDict = TypedDict('RoiDict', {"x": ..., "y", ....}, total=False) class Classifier: ... def predict(roi: RoiDict): croi = CustomRoi(**roi) ```...

Hey @pdd2110, you can achieve something fairly similar right now already, by providing this image as an atlas image in data selection. With that, the result table will include those...

Hey @pdd2110, thank you very much for that detailed comment! I hadn't thought about neighborhoods there... Exposing the neighborhood as a parameter in thresholding should not be a problem. Properly...

thx for the detailed report. I agree with all your points :) In general there is no conceivable reason for the need of access to training data at inference time...