494 reenable testing (#1074)

Signed-off-by: Wenqi Li <[email protected]>

Fixes #494


### Checks
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Notebook runs automatically `./runner [-p <regex_pattern>]`

Signed-off-by: Wenqi Li <[email protected]>

Author: wyli

.github/workflows/test-modified.yml

@@ -1,4 +1,4 @@
-name: build
+name: premerge-notebooks
 
 on:
   # quick tests of modified notebooks for every pull request
@@ -25,7 +25,12 @@ jobs:
       run: |
         which python
         nvidia-smi
+        rm -rf ../../MONAI/MONAI
         python -m pip install --upgrade pip wheel
+        pip uninstall -y monai
+        pip uninstall -y monai
+        pip uninstall -y monai-weekly
+        pip uninstall -y monai-weekly # make sure there's no existing installation
         BUILD_MONAI=0 python -m pip install git+https://github.com/Project-MONAI/MONAI#egg=MONAI
         python -m pip install -r https://raw.githubusercontent.com/Project-MONAI/MONAI/main/requirements-dev.txt
     - uses: actions/checkout@v3

deep_atlas/deep_atlas_tutorial.ipynb

@@ -70,7 +70,7 @@
     "  If we want a deep neural network to do registration, then we can use the segmentation algorithm in this manner to create an objective to train against. We refer to the resulting loss function as _anatomy loss_.\n",
     "  <br>\n",
     "  \n",
-    "- _Reg helps seg:_ Suppose we have a great registration algorithm that can deform one image to another in an anatomically realistic way. Then given just _one_ manually created segmentation, we have an equally great segmentaion algorithm:\n",
+    "- _Reg helps seg:_ Suppose we have a great registration algorithm that can deform one image to another in an anatomically realistic way. Then given just _one_ manually created segmentation, we have an equally great segmentation algorithm:\n",
     "  1. Register the image you want to segment to your _atlas_, the single image for which you already have a segmentation.\n",
     "  2. Use the deformation from the registration to warp the atlas's segmentation back onto to your image.\n",
     "  \n",
@@ -2436,7 +2436,7 @@
     "            seg2 = seg2_predicted  # Use this in anatomy loss\n",
     "\n",
     "        else:  # seg2 available, but no seg1\n",
-    "            assert('seg2' in batch.keys())\n",
+    "            assert 'seg2' in batch.keys()\n",
     "            seg2 = monai.networks.one_hot(batch['seg2'].to(device), num_segmentation_classes)\n",
     "            loss_metric = dice_loss2(seg2_predicted, seg2)\n",
     "            loss_supervised = loss_metric\n",

runner.sh

@@ -76,7 +76,6 @@ pattern="-and -name '*' -and ! -wholename '*federated_learning*'\
  -and ! -wholename '*profiling_train_base_nvtx*'\
  -and ! -wholename '*benchmark_global_mutual_information*'\
  -and ! -wholename '*spleen_segmentation_3d_visualization_basic*'\
- -and ! -wholename '*deep_atlas_tutorial*'\
  -and ! -wholename '*nuclick_infer*'\
  -and ! -wholename '*nuclick_training_notebook*'\
  -and ! -wholename '*full_gpu_inference_pipeline*'\