Chebi's atomtype fix when it is a parent

go-plus.py

@@ -19,27 +19,36 @@ def get_term(class_id):
             term = term.replace("CHEBI", "ChEBI")
     return term
 
-def get_type(term):
+def get_type(term, parent=False):
     if "ChEBI" in term: 
-        return "MoleculeNode"
+        if parent:
+            return "ConceptNode"
+        else:
+            return "MoleculeNode"
     else:
         return "ConceptNode"
 
-def evaLink(term1 , term2, predicate):
+def evaLink(term1 , term2, predicate, parent_chebis=[]):
     if not (str(term1) == "nan" or str(term2) == 'nan'):
         return ("(EvaluationLink \n" +
             "\t (PredicateNode \""+ predicate + "\")\n" +
             "\t (ListLink \n" +
-            "\t\t ({}".format(get_type(term1))  + " \"" + term1 + "\")\n" +
-            "\t\t ({}".format(get_type(term2)) + " \"" + term2 + "\")))\n" )
+            "\t\t ({}".format(get_type(term1, parent=is_parent(term1, parent_chebis)))  + " \"" + term1 + "\")\n" +
+            "\t\t ({}".format(get_type(term2, parent=is_parent(term2, parent_chebis))) + " \"" + term2 + "\")))\n" )
     else:
         return ""
 
-def inheritLink(term1 , term2):
+def is_parent(term, parent_chebis):
+    if term in parent_chebis:
+        return True
+    else:
+        return False
+
+def inheritLink(term1 , term2, parent_chebis=[]):
     if not (str(term1) == "nan" or str(term2) == 'nan'):
         return ("(InheritanceLink \n" +
-                "\t ({}".format(get_type(term1)) + " \"" + term1 + "\")\n" +
-                "\t ({}".format(get_type(term2)) + " \"" + term2 + "\"))\n" )
+                "\t ({}".format(get_type(term1, parent=is_parent(term1, parent_chebis))) + " \"" + term1 + "\")\n" +
+                "\t ({}".format(get_type(term2, parent=is_parent(term2, parent_chebis))) + " \"" + term2 + "\"))\n")
     else:
         return ""
 
@@ -49,6 +58,14 @@ def inheritLink(term1 , term2):
 if not os.path.exists("raw_data/GO-PLUS.csv.gz"):
     dataset = wget.download(source_csv, "raw_data")
 df = pd.read_csv("raw_data/GO-PLUS.csv.gz", dtype=str)
+# Parent CHEBI's should be a ConceptNode, not a MoleculeNode
+parents = df["Parents"]
+parent_chebis = []
+for i in [i.split("|") for i in parents if str(i) != "nan"]:
+    for c in i:
+        term = get_term(c)
+        if "ChEBI" in term:
+            parent_chebis.append(term)
 
 go_columns = ["negatively regulated by","negatively regulates", "positively regulated by", "positively regulates", "regulated by", "regulates", "has part", "part of"]
 uberon_columns = open("raw_data/uberon_columns.txt", "r").read().splitlines()
@@ -97,18 +114,18 @@ def inheritLink(term1 , term2):
             file_name = chebi
             file_name_with_def = chebi_with_def
 
-        file_name.write(evaLink(term, get_term(df.iloc[i]["Preferred Label"]), "has_name"))
-        file_name_with_def.write(evaLink(term, get_term(df.iloc[i]["Preferred Label"]), "has_name"))
-        file_name_with_def.write(evaLink(term, definition, "has_definition"))
+        file_name.write(evaLink(term, get_term(df.iloc[i]["Preferred Label"]), "has_name", parent_chebis=parent_chebis))
+        file_name_with_def.write(evaLink(term, get_term(df.iloc[i]["Preferred Label"]), "has_name",parent_chebis=parent_chebis))
+        file_name_with_def.write(evaLink(term, definition, "has_definition", parent_chebis=parent_chebis))
         for col in cl_columns:
             if col == "Parents":
                 parents = df.iloc[i][col]
                 if str(parents) != "nan":
                     for p in parents.split("|"): 
-                        file_name.write(inheritLink(term,get_term(p)))
-                        file_name_with_def.write(inheritLink(term, get_term(p)))
+                        file_name.write(inheritLink(term,get_term(p), parent_chebis=parent_chebis))
+                        file_name_with_def.write(inheritLink(term, get_term(p), parent_chebis=parent_chebis))
             else:
-                file_name.write(evaLink(term, get_term(df.iloc[i][col]), col.replace(" ", "_")))
-                file_name_with_def.write(evaLink(term, get_term(df.iloc[i][col]), col.replace(" ", "_")))
+                file_name.write(evaLink(term, get_term(df.iloc[i][col]), col.replace(" ", "_"), parent_chebis=parent_chebis))
+                file_name_with_def.write(evaLink(term, get_term(df.iloc[i][col]), col.replace(" ", "_"), parent_chebis=parent_chebis))
 print("Done")