Merge pull request #335 from mims-harvard/evebio

Evebio

environment.yml

@@ -22,7 +22,7 @@ dependencies:
   - pip:
     - accelerate==0.33.0
     - cellxgene-census==1.15.0 
-    - datasets==2.20.0
+    - datasets<2.20.0
     - dgl==1.1.3
     - evaluate==0.4.2
     - gget==0.28.4

requirements.txt

@@ -1,7 +1,7 @@
 accelerate==0.33.0
 biopython>=1.78,<2.0
 dataclasses>=0.6,<1.0
-datasets==2.20.0
+datasets<2.20.0
 evaluate==0.4.2
 fuzzywuzzy>=0.18.0,<1.0
 huggingface_hub>=0.20.3,<1.0

tdc/metadata.py

@@ -940,6 +940,22 @@ def get_task2category():
     "geneformer_gene_median_dictionary": "pkl",
     "geneformer_gene_name_id_dict": "pkl",
     "geneformer_token_dictionary": "pkl",
+    "evebio_pharmone_v1_assay_doc": "tab",
+    "evebio_pharmone_v1_assay_table": "tab",
+    "evebio_pharmone_v1_bundle_doc": "tab",
+    "evebio_pharmone_v1_bundle_table": "tab",
+    "evebio_pharmone_v1_compound_doc": "tab",
+    "evebio_pharmone_v1_compound_table": "tab",
+    "evebio_pharmone_v1_control_doc": "tab",
+    "evebio_pharmone_v1_control_table": "tab",
+    "evebio_pharmone_v1_detailed_result_doc": "tab",
+    "evebio_pharmone_v1_detailed_result_table": "tab",
+    "evebio_pharmone_v1_observed_points_doc": "tab",
+    "evebio_pharmone_v1_observed_points_table": "tab",
+    "evebio_pharmone_v1_summary_result_doc": "tab",
+    "evebio_pharmone_v1_summary_result_table": "tab",
+    "evebio_pharmone_v1_target_doc": "tab",
+    "evebio_pharmone_v1_target_table": "tab",
 }
 
 name2id = {
@@ -1130,6 +1146,22 @@ def get_task2category():
     "geneformer_gene_median_dictionary": 10626278,
     "geneformer_gene_name_id_dict": 10626276,
     "geneformer_token_dictionary": 10626277,
+    "evebio_pharmone_v1_assay_doc": 10741530,
+    "evebio_pharmone_v1_assay_table": 10741541,
+    "evebio_pharmone_v1_bundle_doc": 10741540,
+    "evebio_pharmone_v1_bundle_table": 10741543,
+    "evebio_pharmone_v1_compound_doc": 10741535,
+    "evebio_pharmone_v1_compound_table": 10741532,
+    "evebio_pharmone_v1_control_doc": 10741539,
+    "evebio_pharmone_v1_control_table": 10741538,
+    "evebio_pharmone_v1_detailed_result_doc": 10741545,
+    "evebio_pharmone_v1_detailed_result_table": 10741534,
+    "evebio_pharmone_v1_observed_points_doc": 10741531,
+    "evebio_pharmone_v1_observed_points_table": 10741544,
+    "evebio_pharmone_v1_summary_result_doc": 10741529,
+    "evebio_pharmone_v1_summary_result_table": 10741542,
+    "evebio_pharmone_v1_target_doc": 10741536,
+    "evebio_pharmone_v1_target_table": 10741537,
 }
 
 oracle2type = {

tdc/resource/pharmone.py

@@ -0,0 +1,29 @@
+from ..utils import general_load
+"""
+Resource class for the Eve Bio (https://evebio.org/) Pharmone Map.
+"""
+
+
+class PharmoneMap(object):
+
+    def __init__(self, path="./data"):
+        self.path = path
+
+    def get_data(self):
+        return general_load('evebio_pharmone_v1_detailed_result_table',
+                            self.path, '\t')  # Load the Pharmone Map data
+
+    def get_obs_metadata(self):
+        return general_load("evebio_pharmone_v1_observed_points_table",
+                            self.path, "\t")
+
+    def get_control_data(self):
+        return general_load("evebio_pharmone_v1_control_table", self.path,
+                            "\t")  # Load the control data
+
+    def get_compound_data(self):
+        return general_load("evebio_pharmone_v1_compound_table", self.path,
+                            "\t")
+
+    def get_target_data(self):
+        return general_load("evebio_pharmone_v1_target_table", self.path, "\t")

tdc/test/test_resources.py

@@ -126,5 +126,29 @@ def tearDown(self):
             pass
 
 
+class TestPharmoneMap(unittest.TestCase):
+
+    def setUp(self):
+        print(os.getcwd())
+        pass
+
+    def test_get_data(self):
+        from tdc.resource.pharmone import PharmoneMap
+
+        resource = PharmoneMap()
+        data = resource.get_data()
+        assert isinstance(data, DataFrame), type(data)
+        assert "Compound" in data.columns
+        assert "Target_ID" in data.columns
+        assert "pXC50" in data.columns
+
+    def tearDown(self):
+        try:
+            print(os.getcwd())
+            shutil.rmtree(os.path.join(os.getcwd(), "data"))
+        except:
+            pass
+
+
 if __name__ == "__main__":
     unittest.main()