MEDS-DEV 0.0.2__tar.gz → 0.0.3__tar.gz

{meds_dev-0.0.2 → meds_dev-0.0.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: MEDS_DEV
-Version: 0.0.2
+Version: 0.0.3
 Summary: Task configuration and helper files for the MEDS-DEV Benchmark
 Author: Nassim Oufattole
 Author-email: "Matthew B. A. McDermott" <matthew_mcdermott@hms.harvard.edu>, Ethan Steinberg <ethanid@stanford.edu>, Shalmali Joshi <shalmali.joshi@columbia.edu>, "Teya S. Bergamaschi" <teya@mit.edu>, Aleksia Kolo <aleksiak@mit.edu>, Young Sang Choi <young.sang.choi@columbia.edu>, Chao Pang <cp3016@cumc.columbia.edu>, Kamilė Stankevičiūtė <kamile.stankeviciute@cst.cam.ac.uk>, Vincent Jeanselme <vincent.jeanselme@mrc-bsu.cam.ac.uk>, Justin Xu <justin.xu@ndm.ox.ac.uk>

{meds_dev-0.0.2 → meds_dev-0.0.3}/src/MEDS_DEV/configs/_extract_task.yaml

@@ -3,6 +3,7 @@ defaults:
 
 dataset: ???
 dataset_dir: ???
+dataset_predicates_path: null
 task: ???
 output_dir: ???
 do_overwrite: False
@@ -22,4 +23,6 @@ hydra:
 
       You can specify "dataset" to give the name of the dataset you want to build and "dataset_dir" to point
       to the extracted dataset on disk, "output_dir" to say where the final, labeled task cohort
-      should be stored on disk, and "task" to dictate which task should be extracted.
+      should be stored on disk, and "task" to dictate which task should be extracted. If you overwrite
+      "dataset_predicates_path", then it will look at that location for the predicates file, rather than in
+      the MEDS-DEV repository location. This is useful for local datasets.

{meds_dev-0.0.2 → meds_dev-0.0.3}/src/MEDS_DEV/tasks/__main__.py

@@ -16,11 +16,15 @@ logger = logging.getLogger(__name__)
 def main(cfg: DictConfig):
     if cfg.task not in TASKS:
         raise ValueError(f"Task {cfg.task} not currently configured")
-    if cfg.dataset not in DATASETS:
-        raise ValueError(f"Dataset {cfg.dataset} not currently configured")
 
     task_config_path = TASKS[cfg.task]["criteria_fp"]
-    dataset_predicates_path = DATASETS[cfg.dataset]["predicates"]
+    if cfg.get("dataset_predicates_path", None):
+        logger.info(f"Using provided (local) predicates path: {cfg.dataset_predicates_path}")
+        dataset_predicates_path = Path(cfg.dataset_predicates_path)
+    else:
+        if cfg.dataset not in DATASETS:
+            raise ValueError(f"Dataset {cfg.dataset} not currently configured")
+        dataset_predicates_path = DATASETS[cfg.dataset]["predicates"]
 
     if dataset_predicates_path is None:
         raise ValueError(f"Predicates not found for dataset {cfg.dataset}")

{meds_dev-0.0.2 → meds_dev-0.0.3}/src/MEDS_DEV.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: MEDS_DEV
-Version: 0.0.2
+Version: 0.0.3
 Summary: Task configuration and helper files for the MEDS-DEV Benchmark
 Author: Nassim Oufattole
 Author-email: "Matthew B. A. McDermott" <matthew_mcdermott@hms.harvard.edu>, Ethan Steinberg <ethanid@stanford.edu>, Shalmali Joshi <shalmali.joshi@columbia.edu>, "Teya S. Bergamaschi" <teya@mit.edu>, Aleksia Kolo <aleksiak@mit.edu>, Young Sang Choi <young.sang.choi@columbia.edu>, Chao Pang <cp3016@cumc.columbia.edu>, Kamilė Stankevičiūtė <kamile.stankeviciute@cst.cam.ac.uk>, Vincent Jeanselme <vincent.jeanselme@mrc-bsu.cam.ac.uk>, Justin Xu <justin.xu@ndm.ox.ac.uk>