Upgrade gocamgen dependency; some typing

geneontology · Apr 29, 2020 · efb589a · efb589a · dustine32 · Apr 29, 2020
1 parent 67e915f
commit efb589a
Show file tree

Hide file tree

Showing 4 changed files with 31 additions and 51 deletions.
diff --git a/pathway_connections.py b/pathway_connections.py
@@ -26,7 +26,7 @@ def __init__(self, mapping_file=None):
         self.mappings = []
         if mapping_file:
             with open(mapping_file) as mf:
-                mappings = yaml.load(mf)
+                mappings = yaml.safe_load(mf)
             for m in mappings:
                 self.mappings.append(MechanismToGoMapping(
                     mechanism=m["MECHANISM"],
@@ -42,10 +42,16 @@ def go_id_by_mechanism(self, mechanism):
 
 
 class AnnotatorOrcidMapping:
-    def __init__(self, annotator_name, orcid):
+    def __init__(self, annotator_name, orcid: str):
         self.annotator_name = annotator_name
         self.orcid = orcid
 
+    def full_orcid_uri(self):
+        orcid_prefix = "http://orcid.org/"
+        if self.orcid.startswith(orcid_prefix):
+            return self.orcid
+        return orcid_prefix + self.orcid
+
 class AnnotatorOrcidMappingSet:
     def __init__(self, file=None):
         self.mappings = []
@@ -61,7 +67,7 @@ def __init__(self, file=None):
     def orcid_by_name(self, annotator_name):
         for m in self.mappings:
             if m.annotator_name == annotator_name:
-                return m.orcid
+                return m.full_orcid_uri()
 
 # TODO: Refactor `PathwayConnection`
 # * Connect causal statements together in networkx graph
@@ -94,6 +100,7 @@ def __init__(self, entity_a: SignorEntity, entity_b: SignorEntity, mechanism, ef
             "term": None
         }
 
+        self.annotator = None
         if annotator:
             self.annotator = self.ANNOTATOR_ORCID_MAPPING.orcid_by_name(annotator)
 
@@ -247,9 +254,9 @@ def upper_first(iterator):
     return itertools.chain([next(iterator).upper()], iterator)
 
 
-class PathwayConnectionSet():
+class PathwayConnectionSet:
     def __init__(self):
-        self.connections = []
+        self.connections: List[PathwayConnection] = []
 
     @staticmethod
     def parse_file(filename):
@@ -273,22 +280,21 @@ def parse_file(filename):
 
         return pc_set
 
-
-    def add(self, pathway_connection):
+    def add(self, pathway_connection: PathwayConnection):
         existing_connection = self.find(pathway_connection)
         if existing_connection:
             # Causal statement already exists so just add reference
             existing_connection.references = set(existing_connection.references) | set(pathway_connection.references)
         else:
             self.connections.append(pathway_connection)
 
-    def contains(self, pathway_connection, check_ref=False):
+    def contains(self, pathway_connection: PathwayConnection, check_ref=False):
         for connection in self.connections:
             if connection.equals(pathway_connection, check_ref=check_ref):
                 return True
         return False
 
-    def find(self, pathway_connection, check_ref=False):
+    def find(self, pathway_connection: PathwayConnection, check_ref=False):
         for connection in self.connections:
             if connection.equals(pathway_connection, check_ref=check_ref):
                 return connection
@@ -325,4 +331,4 @@ def remove_list(self, pc_list):
         new_connection_list = self.connections
         for dead_pc in pc_list:
             new_connection_list = [pc for pc in new_connection_list if not pc.equals(dead_pc)]
-        self.connections = new_connection_list
+        self.connections = new_connection_list
diff --git a/pathway_importer.py b/pathway_importer.py
@@ -9,6 +9,7 @@
 ro = OboRO()
 ENABLED_BY = URIRef(expand_uri(ro.enabled_by))
 HAS_INPUT = URIRef(expand_uri("RO:0002233"))
+EXP_ECO_CODE = "ECO:0000269"
 
 parser = argparse.ArgumentParser()
 parser.add_argument('-f', "--filename", type=str, required=True,
@@ -80,9 +81,13 @@ def generate_model(filename, title):
         pc.mechanism["uri"] = model.declare_individual(pc.mechanism["term"])
         # Emit mechanism -enabled_by -> entity_a
         pc.enabled_by_stmt_a = model.writer.emit(pc.mechanism["uri"], ENABLED_BY, pc.entity_a.uri)
-        axiom_a = model.add_axiom(pc.enabled_by_stmt_a,
-                                  GoCamEvidence("EXP", ["PMID:" + pmid for pmid in pc.references]))
-        model.add_evidence(axiom_a, "EXP", ["PMID:" + pmid for pmid in pc.references])
+        contributors = []
+        if pc.annotator:
+            contributors = [pc.annotator]
+        evidence = GoCamEvidence(EXP_ECO_CODE, ["PMID:" + pmid for pmid in pc.references],
+                                 contributors=contributors)
+        axiom_a = model.add_axiom(pc.enabled_by_stmt_a, evidence=evidence)
+        # model.add_evidence(axiom_a, "EXP", ["PMID:" + pmid for pmid in pc.references])
 
     # Now that the a's are declared, go check on the b's.
     for pc in p_connections.connections:
@@ -105,7 +110,11 @@ def generate_model(filename, title):
             regulation_triple = (mechanism_uri, URIRef(expand_uri(pc.relation)), regulated_activity_uri)
             model.writer.emit(*regulation_triple)
             regulation_axiom = model.writer.emit_axiom(*regulation_triple)
-            model.add_evidence(regulation_axiom, "EXP", ["PMID:" + pmid for pmid in pc.references])
+            if pc.annotator:
+                contributors = [pc.annotator]
+            evidence = GoCamEvidence(EXP_ECO_CODE, ["PMID:" + pmid for pmid in pc.references],
+                                     contributors=contributors)
+            model.add_evidence(regulation_axiom, evidence=evidence)
 
     print(skipped_count, "causal statements skipped")
     print(len(p_connections.connections), "pathway_connections at finish")

diff --git a/requirements.txt b/requirements.txt
@@ -1,36 +1,2 @@
-argh==0.26.2
-astroid==1.6.1
-cachier==1.2.2
-certifi==2018.1.18
-chardet==3.0.4
-click==6.7
-decorator==4.2.1
-gocamgen==0.0.4
-idna==2.6
-isodate==0.6.0
-isort==4.3.4
-jsobject==0.10.2
-lazy-object-proxy==1.3.1
-marshmallow==2.15.0
-mccabe==0.6.1
-networkx==1.11
-numpy==1.14.2
-ontobio==0.2.36
-pandas==0.22.0
-pathtools==0.1.2
-portalocker==1.2.0
-prefixcommons==0.1.4
-pylint==1.8.2
-pyparsing==2.2.0
-pysolr==3.7.0
-python-dateutil==2.7.2
-pytz==2018.3
-PyYAML==3.12
-rdflib==4.2.2
+gocamgen==0.0.5
 requests==2.18.4
-scipy==1.0.1
-six==1.11.0
-SPARQLWrapper==1.8.1
-urllib3==1.22
-watchdog==0.8.3
-wrapt==1.10.11
diff --git a/test.py b/test.py
@@ -14,7 +14,7 @@ class TestSignor2Gocam(unittest.TestCase):
 
     def test_mechanism_map_loading(self):
         with open(M_FILE) as mf:
-            mappings = yaml.load(mf)
+            mappings = yaml.safe_load(mf)
         self.assertGreater(len(mappings), 1)
 
     def test_mechanism_mapping_set(self):
@@ -127,7 +127,6 @@ def test_distinct_entity_instance_per_stmt(self):
 
         # TODO: Now check how many instances of most_prominent_entity are in model. Should this == input count?
         print(most_prominent_entity, entity_counts[most_prominent_entity])
-        pass
 
 
 if __name__ == '__main__':