merge main

.bumpversion.cfg

@@ -0,0 +1,8 @@
+[bumpversion]
+current_version = 0.1.8
+tag = True
+commit = True
+
+[bumpversion:file:pyproject.toml]
+
+[bumpversion:file:cyto_dl/__init__.py]

.github/workflows/bump.yml

@@ -0,0 +1,53 @@
+---
+name: bump
+on:
+  pull_request:
+    types: [closed]
+    branches: [main]
+
+jobs:
+  bump:
+    # make sure this job runs only when the PR is not a version bump PR and the pr is merged to main (not just closed)
+    if: github.event.pull_request.title != 'admin/version-bump' && github.event.pull_request.merged == true
+    runs-on: ubuntu-latest
+    environment: release
+    permissions:
+      id-token: write
+      contents: write # Add this permission to allow pushing the version bump
+      pull-requests: write
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v2
+        with:
+          fetch-depth: 0 # Fetch all history for all branches and tags
+
+      - name: Setup PDM
+        uses: pdm-project/setup-pdm@v2
+        with:
+          python-version: "3.10"
+
+      - name: Install bumpversion
+        run: pip install bumpversion
+
+      - name: Configure Git
+        run: |
+          git config user.name github-actions
+          git config user.email [email protected]
+
+      - name: Delete remote branch # delete the branch if it exists
+        run: |
+          git push origin --delete admin/version-bump || true
+
+      - name: Apply bumpversion minor
+        run: |
+          git checkout -b admin/version-bump
+          bumpversion minor
+          git push origin admin/version-bump
+
+      - name: Create Pull Request
+        uses: peter-evans/create-pull-request@v5
+        with:
+          branch: admin/version-bump
+          title: admin/version-bump
+          body: "Automated version bump"
+          base: main

.github/workflows/publish.yml

@@ -1,11 +1,13 @@
 ---
 name: publish
 on:
-  push:
+  pull_request:
+    types: [closed]
     branches: [main]
-
 jobs:
   publish:
+    # only publish when the PR is a version bump PR and the pr is merged to main
+    if: github.event.pull_request.title == 'admin/version-bump' && github.event.pull_request.merged == true
     runs-on: ubuntu-latest
     environment: release
     permissions:

configs/data/im2im/segmentation_plugin.yaml

@@ -33,14 +33,16 @@ transforms:
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [TRAIN] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col1_channel}
       - _target_: monai.transforms.LoadImaged
         keys: ${target_col2}
         allow_missing_keys: True
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [TRAIN] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col2_channel}
       - _target_: monai.transforms.ThresholdIntensityd
         allow_missing_keys: True
         keys:
@@ -108,14 +110,16 @@ transforms:
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [TEST] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col1_channel}
       - _target_: monai.transforms.LoadImaged
         keys: ${target_col2}
         allow_missing_keys: True
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [TEST] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col2_channel}
       # load merging mask - assumed not to exist by default
       - _target_: cyto_dl.image.io.PolygonLoaderd
         keys:
@@ -184,14 +188,16 @@ transforms:
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [VAL] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col1_channel}
       - _target_: monai.transforms.LoadImaged
         keys: ${target_col2}
         allow_missing_keys: True
         reader:
           - _target_: cyto_dl.image.io.MonaiBioReader
             dimension_order_out: ${eval:'"CZYX" if ${spatial_dims}==3 else "CYX"'}
-            C: 0
+            # [VAL] Exposing C as a configurable param for multi-channel segs
+            C: ${target_col2_channel}
 
       - _target_: monai.transforms.ThresholdIntensityd
         allow_missing_keys: True

configs/experiment/im2im/segmentation_plugin.yaml

@@ -29,6 +29,8 @@ test: False
 source_col: raw
 target_col1: seg1
 target_col2: seg2
+target_col1_channel: 0
+target_col2_channel: 0
 merge_mask_col: merge_mask
 exclude_mask_col: exclude_mask
 base_image_col: base_image

cyto_dl/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.1.7"
+__version__ = "0.1.8"
 
 
 # silence bio packages warnings

cyto_dl/train.py

@@ -8,6 +8,7 @@
 import hydra
 import lightning
 import pyrootutils
+import torch
 from lightning import Callback, LightningDataModule, LightningModule, Trainer
 from lightning.pytorch.loggers.logger import Logger
 from omegaconf import DictConfig, OmegaConf
@@ -94,6 +95,16 @@ def train(cfg: DictConfig) -> Tuple[dict, dict]:
 
     if cfg.get("train"):
         log.info("Starting training!")
+
+        if cfg.get("weights_only"):
+            assert cfg.get(
+                "ckpt_path"
+            ), "ckpt_path must be provided to with argument weights_only=True"
+            # load model from state dict to get around trainer.max_epochs limit, useful for resuming model training from existing weights
+            state_dict = torch.load(cfg["ckpt_path"])["state_dict"]
+            model.load_state_dict(state_dict)
+            cfg["ckpt_path"] = None
+
         if isinstance(data, LightningDataModule):
             trainer.fit(model=model, datamodule=data, ckpt_path=cfg.get("ckpt_path"))
         else:

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "pdm.pep517.api"
 
 [project]
 name = "cyto-dl"
-version = "0.1.7"
+version = "0.1.8"
 description = """\
   Collection of representation learning models, techniques, callbacks, utils, \
   used to create latent variable models of cell shape, morphology and \

tests/test_polygon_loader.py

@@ -7,8 +7,9 @@
 
 def test_load_polygon(tmp_path):
     # create random image and find contours
-    image = random_shapes((500, 500), max_shapes=1, num_channels=1)[0].squeeze()
-    image[image == 255] = 0
+    image = random_shapes((500, 500), max_shapes=1, num_channels=1, random_seed=3)[0].squeeze()
+    # create polygon from background
+    image = image == 255
     contours = find_contours(label(image))
 
     # Convert contours to polygons and save
@@ -24,8 +25,8 @@ def test_load_polygon(tmp_path):
     reconstructed = transform(data)["poly"][0]
 
     # Check that the reconstructed mask is the close to as the original
-    iou = np.sum(np.logical_and(reconstructed[0], image > 0)) / np.sum(
-        np.logical_or(reconstructed[0], image > 0)
+    iou = np.sum(np.logical_and(reconstructed[0] > 0, image > 0)) / np.sum(
+        np.logical_or(reconstructed[0] > 0, image > 0)
     )
 
     assert iou > 0.8