Major updates! (#44)

* added No Patching checkbox to extractor widget

* updated models adapters to support no_patching

* updated segmentation widget

* removed unused imports

* fixed some typings

* added the dataset class

* init using the new dataset for feature extraction

* init using zarr storage

* moved up no patching checkbox

* updated extractor widget and no_patching option

* updated segmentation widget using zarr storage

* clean up & formatting

* fixed extraction progress info

* updated dataset to handle images type inside

* updated prediction pipeline over a large stack

* fixed run pipeline; file dialogs open in parent dialog

* updated run_pipeline script

* removed unused import in run_pipeline

* fixed models params: image height & width as int

* ignored some typing

* updated requirements

* fixed image height & width as int

* fixed dataset & get_model_ready_image image dimention problem

* added dataset test

* fixed adapter: concat double output into one tensor

* updated mobilesam test

* updated dino adapter test

* added sam2 adapter test

* added pipeline_prediction test

* bumped version

* run_pipeline script can be used to only extract features

* updated dependencies in pyproject.toml

* fixed bug for calculating padding with no_patching

* revert back storage to hdf5; improved get_train_data

* fixed embedding_extraction test

Author: mese79

pyproject.toml

@@ -54,7 +54,9 @@ dependencies = [
     "qtpy",
     "magicgui",
     "napari[all]",
-    "h5py",
+    "tifffile",
+    "zarr>=2,<3",
+    "pims",
     "pooch",
     "tqdm>=4.66.1",
     "iopath>=0.1.10",

requirements.txt

@@ -9,7 +9,9 @@ pynrrd
 qtpy
 magicgui
 napari[all]
-h5py
+tifffile
+zarr>=2,<3
+pims
 pooch
 tqdm>=4.66.1
 iopath>=0.1.10