Merge pull request #260 from mims-harvard/package_release

Change pinnacle and pentelute dataset names to more appropriate ones

tdc/benchmark_group/scdti_group.py

@@ -18,23 +18,23 @@ def __init__(self, path="./data", file_format="csv"):
         # super().__init__(name="SCDTI_Group", path=path)
         self.name = "SCDTI_Group"
         self.path = os.path.join(path, self.name)
-        # self.datasets = ["pinnacle_dti"]
-        self.dataset_names = ["pinnacle_dti"]
+        # self.datasets = ["opentargets_dti"]
+        self.dataset_names = ["opentargets_dti"]
         self.file_format = file_format
         self.split = None
 
     def get_train_valid_split(self):
         """parameters included for compatibility. this benchmark has a fixed train/test split."""
         from ..resource.dataloader import DataLoader
         if self.split is None:
-            dl = DataLoader(name="pinnacle_dti")
+            dl = DataLoader(name="opentargets_dti")
             self.split = dl.get_split()
         return self.split["train"], self.split["dev"]
 
     def get_test(self):
         from ..resource.dataloader import DataLoader
         if self.split is None:
-            dl = DataLoader(name="pinnacle_dti")
+            dl = DataLoader(name="opentargets_dti")
             self.split = dl.get_split()
         return self.split["test"]
 

tdc/dataset_configs/__init__.py

@@ -1,2 +1,2 @@
 from .config import DatasetConfig
-from .pentelute_mdm2_ace2_12ca5_config import PenteluteProteinPeptideConfig
+from .brown_mdm2_ace2_12ca5_config import BrownProteinPeptideConfig

tdc/dataset_configs/pentelute_mdm2_ace2_12ca5_config.py → tdc/dataset_configs/brown_mdm2_ace2_12ca5_config.py

@@ -2,12 +2,12 @@
 from ..feature_generators.protein_feature_generator import ProteinFeatureGenerator
 
 
-class PenteluteProteinPeptideConfig(DatasetConfig):
-    """Configuration for the pentelute-protein-peptide datasets"""
+class BrownProteinPeptideConfig(DatasetConfig):
+    """Configuration for the brown-protein-peptide datasets"""
 
     def __init__(self):
-        super(PenteluteProteinPeptideConfig, self).__init__(
-            dataset_name="pentelute_mdm2_ace2_12ca5",
+        super(BrownProteinPeptideConfig, self).__init__(
+            dataset_name="brown_mdm2_ace2_12ca5",
             data_processing_class=ProteinFeatureGenerator,
             functions_to_run=[
                 "autofill_identifier", "create_range", "insert_protein_sequence"

tdc/dataset_configs/config_map.py

@@ -1,5 +1,5 @@
-from .pentelute_mdm2_ace2_12ca5_config import PenteluteProteinPeptideConfig
-from .pinnacle_dti import PinnacleDTI
+from .brown_mdm2_ace2_12ca5_config import BrownProteinPeptideConfig
+from .opentargets_dti import OpentargetsDTI
 from .scperturb_config import SCPerturb, SCPerturb_Gene
 
 scperturb_datasets = [
@@ -23,9 +23,9 @@ class ConfigMap(dict):
     """
 
     def __init__(self):
-        self["pentelute_mdm2_ace2_12ca5"] = PenteluteProteinPeptideConfig
+        self["brown_mdm2_ace2_12ca5"] = BrownProteinPeptideConfig
         for ds in scperturb_datasets:
             self[ds] = SCPerturb
         for ds in scperturb_gene_datasets:
             self[ds] = SCPerturb_Gene
-        self["pinnacle_dti"] = PinnacleDTI
+        self["opentargets_dti"] = OpentargetsDTI

tdc/dataset_configs/opentargets_dti.py

@@ -0,0 +1,63 @@
+from .config import ResourceConfig
+from ..feature_generators.resource import ResourceFeatureGenerator
+
+
+class OpentargetsDTI(ResourceConfig):
+    """Configuration for opentargets drug-target-identification datasets"""
+
+    def __init__(self):
+        super(OpentargetsDTI, self).__init__(
+            ResourceFeatureGenerator(),
+            keys=[
+                "opentargets_ra_data_splits", "opentargets_ibd_data_splits",
+                "df"
+            ],
+            loader_functions=["split", "split", "concat"],
+            loader_args=[{
+                "dataset": ("self", "opentargets_ra_drug_evidence"),
+                "column_name":
+                    "targetId_genename",
+                "pos_train": ("self", [
+                    "opentargets_ra_data_splits", "splits", "pos_train_names"
+                ]),
+                "pos_dev":
+                    None,
+                "pos_test":
+                    ("self",
+                     ["opentargets_ra_data_splits", "splits",
+                      "pos_test_names"]),
+                "neg_train": ("self", [
+                    "opentargets_ra_data_splits", "splits", "neg_train_names"
+                ]),
+                "neg_dev":
+                    None,
+                "neg_test":
+                    ("self",
+                     ["opentargets_ra_data_splits", "splits",
+                      "neg_test_names"]),
+            }, {
+                "dataset": ("self", "opentargets_ibd_drug_evidence"),
+                "column_name":
+                    "targetId_genename",
+                "pos_train": ("self", [
+                    "opentargets_ibd_data_splits", "splits", "pos_train_names"
+                ]),
+                "pos_dev":
+                    None,
+                "pos_test": ("self", [
+                    "opentargets_ibd_data_splits", "splits", "pos_test_names"
+                ]),
+                "neg_train": ("self", [
+                    "opentargets_ibd_data_splits", "splits", "neg_train_names"
+                ]),
+                "neg_dev":
+                    None,
+                "neg_test": ("self", [
+                    "opentargets_ibd_data_splits", "splits", "neg_test_names"
+                ])
+            }, {
+                "ds_list": [
+                    "opentargets_ibd_data_splits", "opentargets_ra_data_splits"
+                ],
+                "axis": 0
+            }])

tdc/metadata.py

@@ -179,7 +179,7 @@
 
 trial_outcome_dataset_names = ['phase1', 'phase2', 'phase3']
 
-proteinpeptide_dataset_names = ['pentelute_mdm2_ace2_12ca5']
+proteinpeptide_dataset_names = ['brown_mdm2_ace2_12ca5']
 
 cellxgene_dataset_names = [
     "scperturb_drug_AissaBenevolenskaya2021",
@@ -193,29 +193,29 @@
 ]
 
 resource_dataset_names = [
-    "pinnacle_ra_data_splits",
-    "pinnacle_ibd_data_splits",
-    "pinnacle_ra_drug_evidence",
-    "pinnacle_ibd_drug_evidence",
-    "pinnacle_ra_data_splits_idx",
-    "pinnacle_ibd_data_splits_idx",
+    "opentargets_ra_data_splits",
+    "opentargets_ibd_data_splits",
+    "opentargets_ra_drug_evidence",
+    "opentargets_ibd_drug_evidence",
+    "opentargets_ra_data_splits_idx",
+    "opentargets_ibd_data_splits_idx",
 ]
 
 resources = {
-    "pinnacle_dti": {
+    "opentargets_dti": {
         "splits": [
-            "pinnacle_ra_data_splits",
-            "pinnacle_ibd_data_splits",
+            "opentargets_ra_data_splits",
+            "opentargets_ibd_data_splits",
         ],
         "datasets": [
-            "pinnacle_ra_drug_evidence",
-            "pinnacle_ibd_drug_evidence",
+            "opentargets_ra_drug_evidence",
+            "opentargets_ibd_drug_evidence",
         ],
         "all": [
-            "pinnacle_ra_data_splits",
-            "pinnacle_ibd_data_splits",
-            "pinnacle_ra_drug_evidence",
-            "pinnacle_ibd_drug_evidence",
+            "opentargets_ra_data_splits",
+            "opentargets_ibd_data_splits",
+            "opentargets_ra_drug_evidence",
+            "opentargets_ibd_drug_evidence",
         ],
     }
 }
@@ -770,7 +770,7 @@ def get_task2category():
     "phase1": "tab",
     "phase2": "tab",
     "phase3": "tab",
-    "pentelute_mdm2_ace2_12ca5": "xlsx",
+    "brown_mdm2_ace2_12ca5": "xlsx",
     "scperturb_drug_AissaBenevolenskaya2021": "h5ad",
     "scperturb_drug_SrivatsanTrapnell2020_sciplex2": "h5ad",
     "scperturb_drug_SrivatsanTrapnell2020_sciplex3": "h5ad",
@@ -779,12 +779,12 @@ def get_task2category():
     "scperturb_gene_NormanWeissman2019": "h5ad",
     "scperturb_gene_ReplogleWeissman2022_rpe1": "h5ad",
     "scperturb_gene_ReplogleWeissman2022_k562_essential": "h5ad",
-    "pinnacle_ra_data_splits": "json",
-    "pinnacle_ra_data_splits_idx": "json",
-    "pinnacle_ibd_data_splits": "json",
-    "pinnacle_ibd_data_splits_idx": "json",
-    "pinnacle_ra_drug_evidence": "tab",
-    "pinnacle_ibd_drug_evidence": "tab",
+    "opentargets_ra_data_splits": "json",
+    "opentargets_ra_data_splits_idx": "json",
+    "opentargets_ibd_data_splits": "json",
+    "opentargets_ibd_data_splits_idx": "json",
+    "opentargets_ra_drug_evidence": "tab",
+    "opentargets_ibd_drug_evidence": "tab",
 }
 
 name2id = {
@@ -899,7 +899,7 @@ def get_task2category():
     "phase1": 7331305,
     "phase2": 7331306,
     "phase3": 7331307,
-    "pentelute_mdm2_ace2_12ca5": 9649623,
+    "brown_mdm2_ace2_12ca5": 9649623,
     "scperturb_drug_AissaBenevolenskaya2021": 9845396,
     "scperturb_drug_SrivatsanTrapnell2020_sciplex2": 9845394,
     "scperturb_drug_SrivatsanTrapnell2020_sciplex3": 9845397,
@@ -908,12 +908,12 @@ def get_task2category():
     "scperturb_gene_NormanWeissman2019": 10133995,
     "scperturb_gene_ReplogleWeissman2022_rpe1": 10133996,
     "scperturb_gene_ReplogleWeissman2022_k562_essential": 10134031,
-    "pinnacle_ra_data_splits": 10141152,
-    "pinnacle_ibd_data_splits": 10141151,
-    "pinnacle_ra_data_splits_idx": 10143574,
-    "pinnacle_ibd_data_splits_idx": 10143573,
-    "pinnacle_ra_drug_evidence": 10141153,
-    "pinnacle_ibd_drug_evidence": 10141154,
+    "opentargets_ra_data_splits": 10141152,
+    "opentargets_ibd_data_splits": 10141151,
+    "opentargets_ra_data_splits_idx": 10143574,
+    "opentargets_ibd_data_splits_idx": 10143573,
+    "opentargets_ra_drug_evidence": 10141153,
+    "opentargets_ibd_drug_evidence": 10141154,
 }
 
 oracle2type = {

tdc/multi_pred/proteinpeptide.py

@@ -8,7 +8,6 @@
 import sys
 
 from ..utils import print_sys
-from ..dataset_configs.pentelute_mdm2_ace2_12ca5_config import PenteluteProteinPeptideConfig
 from . import bi_pred_dataset
 from ..metadata import dataset_names
 from ..dataset_configs.config_map import ConfigMap

tdc/test/test_benchmark.py

@@ -65,15 +65,15 @@ def test_ADME_evaluate_many(self):
     def test_SCDTI_benchmark(self):
         from tdc.resource.dataloader import DataLoader
 
-        data = DataLoader(name="pinnacle_dti")
+        data = DataLoader(name="opentargets_dti")
         group = scdti_group.SCDTIGroup()
         train, val = group.get_train_valid_split()
         assert len(val) == 0  # this benchmark has no validation set
         # test simple preds
         y_true = group.get_test()["Y"]
         results = group.evaluate(y_true)
         assert results[-1] == 1.0  # should be perfect F1 score
-        # assert it matches the PINNACLE official test scores
+        # assert it matches the opentargets official test scores
         tst = data.get_split()["test"]["Y"]
         results = group.evaluate(tst)
         assert results[-1] == 1.0

tdc/test/test_dataloaders.py

@@ -65,13 +65,13 @@ def test_cellxgene_list(self):
         assert len(df) > 0
         print(df.head())
 
-    def test_pentelute(self):
+    def test_brown(self):
         # TODO: factor out into specialized test suites for individual datasets
         # this test serves as an integration test of the data processing, data configs, and existing tdc pipeline. leave here for now.
         from tdc.multi_pred import ProteinPeptide
-        data = ProteinPeptide(name="pentelute_mdm2_ace2_12ca5")
+        data = ProteinPeptide(name="brown_mdm2_ace2_12ca5")
         assert "protein_or_rna_sequence" in data.get_data(
-        ).columns  # pentelute protein<>peptide dataset uses a data config inserting this column
+        ).columns  # brown protein<>peptide dataset uses a data config inserting this column
         data.get_split()
 
     @unittest.skip(
@@ -96,7 +96,7 @@ def test_resource_dataverse_dataloader(self):
         import pandas as pd
         from tdc.resource.dataloader import DataLoader
 
-        data = DataLoader(name="pinnacle_dti")
+        data = DataLoader(name="opentargets_dti")
         df = data.get_data()
         assert "Y" in df.columns
         split = data.get_split()