Patch cath (#25)

* add missing attribute * update version strings * add overwrite attr to FoldClassification * add test to check for overwrite * fix operator typo * fix config interpolation error * fix formatting * fix paths in test * add overwrite to EC * add missing overwrites * ignore atom3d * update changelog --------- Co-authored-by: Arian Jamasb <[email protected]>
a-r-j · Aug 31, 2023 · 932841b · 932841b
1 parent 3092377
commit 932841b
Show file tree

Hide file tree

Showing 9 changed files with 38 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,3 +1,7 @@
+### 0.2.3 (31/08/2023)
+
+* Minor patch; adds missing `overwrite` attribute to `CATHDataModule`, `FoldClassificationDataModule` and `GeneOntologyDataModule`. ([#25](https://github.com/a-r-j/ProteinWorkshop/pull/25))
+
 ### 0.2.2 (30/08/2023)
 
 * Fixes raw data download triggered by absence of PDB when using pre-processed datasets ([#24](https://github.com/a-r-j/ProteinWorkshop/pull/24))

diff --git a/docs/source/conf.py b/docs/source/conf.py
@@ -9,7 +9,7 @@
 # https://www.sphinx-doc.org/en/master/usage/configuration.html#project-information
 project = "Protein Workshop"
 author = "Arian R. Jamasb"
-release = "0.2.2"
+release = "0.2.3"
 copyright = f"{datetime.datetime.now().year}, {author}"
 
 # -- General configuration ---------------------------------------------------

diff --git a/proteinworkshop/datasets/cath.py b/proteinworkshop/datasets/cath.py
@@ -67,6 +67,7 @@ def __init__(
             self.transform = None
 
         self.in_memory = in_memory
+        self.overwrite = overwrite
 
         self.batch_size = batch_size
         self.pin_memory = pin_memory

diff --git a/proteinworkshop/datasets/ec_reaction.py b/proteinworkshop/datasets/ec_reaction.py
@@ -50,6 +50,7 @@ def __init__(
         self.num_workers = num_workers
         self.shuffle_labels = shuffle_labels
         self.format = format
+        self.overwrite = overwrite
 
         self.prepare_data_per_node = True
         logger.info(

diff --git a/proteinworkshop/datasets/fold_classification.py b/proteinworkshop/datasets/fold_classification.py
@@ -55,6 +55,7 @@ def __init__(
         self.structure_dir = self.data_dir / "pdbstyle-1.75"
 
         self.in_memory = in_memory
+        self.overwrite = overwrite
 
         self.dataset_fraction = dataset_fraction
         self.batch_size = batch_size

diff --git a/proteinworkshop/datasets/go.py b/proteinworkshop/datasets/go.py
@@ -61,6 +61,7 @@ def __init__(
         self.format = format
 
         self.in_memory = in_memory
+        self.overwrite = overwrite
 
         self.batch_size = batch_size
         self.pin_memory = pin_memory

diff --git a/proteinworkshop/datasets/masif_site.py b/proteinworkshop/datasets/masif_site.py
@@ -42,6 +42,7 @@ def __init__(
         else:
             self.transform = None
 
+        self.overwrite = overwrite
         self.in_memory = in_memory
         self.dataset_fraction = dataset_fraction
         self.obsolete = obsolete

diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "proteinworkshop"
-version = "0.2.2"
+version = "0.2.3"
 description = ""
 authors = ["Arian Jamasb <[email protected]>"]
 readme = "README.md"

diff --git a/tests/dataset/test_datasets.py b/tests/dataset/test_datasets.py
@@ -46,3 +46,30 @@ def test_instantiate_datasets(tmp_path):
 
         assert dataset, f"Dataset {t} not instantiated!"
         assert isinstance(dataset, LightningDataModule)
+
+
+def test_datasets_have_overwrite_attr(tmp_path):
+        for t in os.listdir(DATASET_CONFIG_DIR):
+            config_path = DATASET_CONFIG_DIR / t
+            cfg = omegaconf.OmegaConf.load(config_path)
+
+            if "data_dir" in cfg.datamodule:
+                cfg.datamodule.data_dir = tmp_path
+
+            if "path" in cfg.datamodule:
+                cfg.datamodule.path = tmp_path
+
+            if "pdb_dir" in cfg.datamodule:
+                cfg.datamodule.pdb_dir = tmp_path
+
+            if "transforms" in cfg.datamodule:
+                cfg.datamodule.transforms = None
+
+            if "transform" in cfg.datamodule:
+                cfg.datamodule.transform = None
+
+            if cfg.datamodule._target_ in {"graphein.ml.datasets.foldcomp_dataset.FoldCompLightningDataModule", "proteinworkshop.datasets.atom3d_datamodule.ATOM3DDataModule"}:
+                continue 
+            else:
+                dm = instantiate(cfg.datamodule)
+                assert hasattr(dm, "overwrite"), f"Datamodules {dm} has no overwrite attribute"