Workflows update | Badges | Torch | New pipeline API

.github/workflows/macos-build.yml

@@ -0,0 +1,32 @@
+name: Macos Build
+
+on:
+  push:
+    branches:
+      - main
+      - spiros-dev
+  pull_request:
+    branches:
+      - main
+      - spiros-dev
+
+jobs:
+  test:
+    runs-on: macos-latest
+    strategy:
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Setup python version ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        run: |
+          pip install -r requirements.txt
+          python3 -m pip install -e .

.github/workflows/ubuntu-build.yml

@@ -0,0 +1,32 @@
+name: Ubuntu Build
+
+on:
+  push:
+    branches:
+      - main
+      - spiros-dev
+  pull_request:
+    branches:
+      - main
+      - spiros-dev
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        python-version: ["3.9", "3.10", "3.11", "3.12"]
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Setup python version ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Install dependencies
+        run: |
+          pip install -r requirements.txt
+          python3 -m pip install -e .

DLICV/__main__.py

@@ -1,20 +1,13 @@
 import argparse
-import json
-import os
-import shutil
-import sys
 import warnings
-from pathlib import Path
 
-import torch
-
-from .utils import prepare_data_folder, rename_and_copy_files
+from .dlicv_pipeline import run_pipeline
 
 warnings.simplefilter(action="ignore", category=FutureWarning)
 warnings.simplefilter(action="ignore", category=UserWarning)
 
 # VERSION = pkg_resources.require("NiChart_DLMUSE")[0].version
-VERSION = 1.0
+VERSION = "1.0.4"
 
 
 def main() -> None:
@@ -25,7 +18,6 @@ def main() -> None:
         usage="""
         DLICV v{VERSION}
         ICV calculation for structural MRI data.
-
         Required arguments:
             [-i, --in_dir]   The filepath of the input directory
             [-o, --out_dir]  The filepath of the output directory
@@ -203,142 +195,27 @@ def main() -> None:
     )
 
     args = parser.parse_args()
-    args.f = [0]
-    args.i = args.in_dir
-    args.o = args.out_dir
-
-    if args.clear_cache:
-        shutil.rmtree(os.path.join(Path(__file__).parent, "nnunet_results"))
-        shutil.rmtree(os.path.join(Path(__file__).parent, ".cache"))
-        if not args.i or not args.o:
-            print("Cache cleared and missing either -i / -o. Exiting.")
-            sys.exit(0)
-
-    if not args.i or not args.o:
-        parser.error("The following arguments are required: -i, -o")
-
-    # data conversion
-    src_folder = args.i  # input folder
-    if not os.path.exists(args.o):  # create output folder if it does not exist
-        os.makedirs(args.o)
-
-    des_folder = os.path.join(args.o, "renamed_image")
-
-    # check if -i argument is a folder, list (csv), or a single file (nii.gz)
-    if os.path.isdir(args.i):  # if args.i is a directory
-        src_folder = args.i
-        prepare_data_folder(des_folder)
-        rename_dic, rename_back_dict = rename_and_copy_files(src_folder, des_folder)
-        datalist_file = os.path.join(des_folder, "renaming.json")
-        with open(datalist_file, "w", encoding="utf-8") as f:
-            json.dump(rename_dic, f, ensure_ascii=False, indent=4)
-        print(f"Renaming dic is saved to {datalist_file}")
-
-    model_folder = os.path.join(
-        Path(__file__).parent,
-        "nnunet_results",
-        "Dataset%s_Task%s_dlicv/nnUNetTrainer__nnUNetPlans__3d_fullres/"
-        % (args.d, args.d),
-    )
-
-    if args.clear_cache:
-        shutil.rmtree(os.path.join(Path(__file__).parent, "nnunet_results"))
-        shutil.rmtree(os.path.join(Path(__file__).parent, ".cache"))
-
-    # Check if model exists. If not exist, download using HuggingFace
-    if not os.path.exists(model_folder):
-        # HF download model
-        print("DLICV model not found, downloading...")
-
-        from huggingface_hub import snapshot_download
-
-        local_src = Path(__file__).parent
-        snapshot_download(repo_id="nichart/DLICV", local_dir=local_src)
-        print("DLICV model has been successfully downloaded!")
-    else:
-        print("Loading the model...")
-
-    prepare_data_folder(args.o)
 
-    # Check for invalid arguments - advise users to see nnUNetv2 documentation
-    assert args.part_id < args.num_parts, "See nnUNetv2_predict -h."
-
-    assert args.device in [
-        "cpu",
-        "cuda",
-        "mps",
-    ], f"-device must be either cpu, mps or cuda. Other devices are not tested/supported. Got: {args.device}."
-
-    if args.device == "cpu":
-        import multiprocessing
-
-        torch.set_num_threads(
-            multiprocessing.cpu_count() // 2
-        )  # use half of the threads (better for PC)
-        device = torch.device("cpu")
-    elif args.device == "cuda":
-        # multithreading in torch doesn't help nnU-Netv2 if run on GPU
-        torch.set_num_threads(1)
-        torch.set_num_interop_threads(1)
-        device = torch.device("cuda")
-    else:
-        device = torch.device("mps")
-
-    # exports for nnunetv2 purposes
-    os.environ["nnUNet_raw"] = "/nnunet_raw/"
-    os.environ["nnUNet_preprocessed"] = "/nnunet_preprocessed"
-    os.environ["nnUNet_results"] = (
-        "/nnunet_results"  # where model will be located (fetched from HF)
-    )
-
-    from nnunetv2.inference.predict_from_raw_data import nnUNetPredictor
-
-    # Initialize nnUnetPredictor
-    predictor = nnUNetPredictor(
-        tile_step_size=args.step_size,
-        use_gaussian=True,
-        use_mirroring=not args.disable_tta,
-        perform_everything_on_device=True,
-        device=device,
-        verbose=args.verbose,
-        verbose_preprocessing=args.verbose,
-        allow_tqdm=not args.disable_progress_bar,
+    run_pipeline(
+        args.in_dir,
+        args.out_dir,
+        args.device,
+        args.clear_cache,
+        args.d,
+        args.part_id,
+        args.num_parts,
+        args.step_size,
+        args.disable_tta,
+        args.verbose,
+        args.disable_progress_bar,
+        args.chk,
+        args.save_probabilities,
+        args.continue_prediction,
+        args.npp,
+        args.nps,
+        args.prev_stage_predictions,
     )
 
-    # Retrieve the model and its weight
-    predictor.initialize_from_trained_model_folder(
-        model_folder, args.f, checkpoint_name=args.chk
-    )
-
-    # Final prediction
-    predictor.predict_from_files(
-        des_folder,
-        args.o,
-        save_probabilities=args.save_probabilities,
-        overwrite=not args.continue_prediction,
-        num_processes_preprocessing=args.npp,
-        num_processes_segmentation_export=args.nps,
-        folder_with_segs_from_prev_stage=args.prev_stage_predictions,
-        num_parts=args.num_parts,
-        part_id=args.part_id,
-    )
-
-    # After prediction, convert the image name back to original
-    files_folder = args.o
-
-    for filename in os.listdir(files_folder):
-        if filename.endswith(".nii.gz"):
-            original_name = rename_back_dict[filename]
-            os.rename(
-                os.path.join(files_folder, filename),
-                os.path.join(files_folder, original_name),
-            )
-    # Remove the (temporary) des_folder directory
-    if os.path.exists(des_folder):
-        shutil.rmtree(des_folder)
-
-    print("DLICV Process Done!")
-
 
 if __name__ == "__main__":
     main()