Replace predicate mapping with a local JSON file

build.sbt

@@ -8,7 +8,7 @@ organization := "org.renci"
 
 name := "cam-kp-api"
 
-version := "0.3-pre1"
+version := "0.3-pre4"
 
 licenses := Seq("MIT license" -> url("https://opensource.org/licenses/MIT"))
 

src/it/resources/examples/BCL2L1-decreases-activity-abundance-chemical.json

@@ -0,0 +1,47 @@
+{
+  "description": "Example creative mode query (https://github.com/NCATSTranslator/TranslatorArchitecture/issues/80) using NCBIGene:598 (BCL2L1)",
+  "message": {
+    "query_graph": {
+      "nodes": {
+        "gene": {
+          "categories": [
+            "biolink:Gene"
+          ],
+          "ids": [
+            "NCBIGene:598"
+          ]
+        },
+        "chemical": {
+          "categories": [
+            "biolink:ChemicalEntity"
+          ]
+        }
+      },
+      "edges": {
+        "t_edge": {
+          "object": "gene",
+          "subject": "chemical",
+          "predicates": [
+            "biolink:affects"
+          ],
+          "knowledge_type": "inferred",
+          "qualifier_constraints": [
+            {
+              "qualifier_set": [
+                {
+                  "qualifier_type_id": "biolink:object_aspect_qualifier",
+                  "qualifier_value": "activity_or_abundance"
+                },
+                {
+                  "qualifier_type_id": "biolink:object_direction_qualifier",
+                  "qualifier_value": "decreased"
+                }
+              ]
+            }
+          ]
+        }
+      }
+    }
+  },
+"minExpectedResults": 6
+}

src/it/resources/examples/PRCD-increases-chemical.json

@@ -0,0 +1,47 @@
+{
+  "description": "PRCD increases chemical",
+  "message": {
+    "query_graph": {
+      "nodes": {
+        "gene": {
+          "categories": [
+            "biolink:Gene"
+          ],
+          "ids": [
+            "NCBIGene:768206"
+          ]
+        },
+        "chemical": {
+          "categories": [
+            "biolink:ChemicalEntity"
+          ]
+        }
+      },
+      "edges": {
+        "t_edge": {
+          "object": "gene",
+          "subject": "chemical",
+          "predicates": [
+            "biolink:affects"
+          ],
+          "knowledge_type": "inferred",
+          "qualifier_constraints": [
+            {
+              "qualifier_set": [
+                {
+                  "qualifier_type_id": "biolink:object_aspect_qualifier",
+                  "qualifier_value": "activity_or_abundance"
+                },
+                {
+                  "qualifier_type_id": "biolink:object_direction_qualifier",
+                  "qualifier_value": "increased"
+                }
+              ]
+            }
+          ]
+        }
+      }
+    }
+  },
+  "minExpectedResults": 1000
+}

src/it/resources/examples/STING1-increases-activity-abundance-chemical.json

@@ -0,0 +1,47 @@
+{
+  "description": "Example creative mode query (https://github.com/NCATSTranslator/TranslatorArchitecture/issues/79) using NCBIGene:340061 (STING1)",
+  "message": {
+    "query_graph": {
+      "nodes": {
+        "gene": {
+          "categories": [
+            "biolink:Gene"
+          ],
+          "ids": [
+            "NCBIGene:340061"
+          ]
+        },
+        "chemical": {
+          "categories": [
+            "biolink:ChemicalEntity"
+          ]
+        }
+      },
+      "edges": {
+        "t_edge": {
+          "object": "gene",
+          "subject": "chemical",
+          "predicates": [
+            "biolink:affects"
+          ],
+          "knowledge_type": "inferred",
+          "qualifier_constraints": [
+            {
+              "qualifier_set": [
+                {
+                  "qualifier_type_id": "biolink:object_aspect_qualifier",
+                  "qualifier_value": "activity_or_abundance"
+                },
+                {
+                  "qualifier_type_id": "biolink:object_direction_qualifier",
+                  "qualifier_value": "increased"
+                }
+              ]
+            }
+          ]
+        }
+      }
+    }
+  }
+  "minExpectedResults": 6
+}

src/it/resources/examples/genes-upstream-of-GPR35.json

@@ -19,7 +19,17 @@
         "e0": {
           "subject": "n0",
           "object": "n1",
-          "predicates": ["biolink:affects_activity_of"]
+          "predicates": ["biolink:affects"],
+          "qualifier_constraints": [
+            {
+              "qualifier_set": [
+                {
+                  "qualifier_type_id": "biolink:object_aspect_qualifier",
+                  "qualifier_value": "activity"
+                }
+              ]
+            }
+          ]
         },
         "e1": {
           "subject": "n1",

src/it/resources/examples/simple.json

@@ -21,5 +21,5 @@
     }
   },
   "minExpectedResults": 160,
-  "maxExpectedResults": 160
+  "maxExpectedResults": 1000
 }

src/it/resources/examples/swagger-example.json

@@ -17,7 +17,17 @@
       "edges": {
         "e0": {
           "predicates": [
-            "biolink:positively_regulates"
+            "biolink:regulates"
+          ],
+          "qualifier_constraints": [
+            {
+              "qualifier_set": [
+                {
+                  "qualifier_type_id": "biolink:object_direction_qualifier",
+                  "qualifier_value": "upregulated"
+                }
+              ]
+            }
           ],
           "subject": "n0",
           "object": "n1"

src/it/scala/org/renci/cam/it/ImplicitsTest.scala

@@ -4,7 +4,7 @@ import com.typesafe.scalalogging.LazyLogging
 import io.circe.generic.auto._
 import io.circe.parser._
 import io.circe.syntax._
-import io.circe.{Decoder, Encoder, KeyDecoder, KeyEncoder}
+import io.circe.{KeyDecoder, KeyEncoder}
 import org.renci.cam.Biolink.BiolinkData
 import org.renci.cam.domain._
 import org.renci.cam.{AppConfig, Biolink, HttpClient, Implicits}
@@ -39,15 +39,23 @@ object ImplicitsTest extends DefaultRunnableSpec with LazyLogging {
     testM("test Implicits.predicateOrPredicateListDecoder") {
       for {
         biolinkData <- Biolink.biolinkData
-      } yield {
-        val dataAsList = """["biolink:participates_in","biolink:related_to"]"""
-        val data = """"biolink:related_to""""
-        import biolinkData.implicits._
-        val ret = decode[List[BiolinkPredicate]](data)
-        val retWithListData = decode[List[BiolinkPredicate]](dataAsList)
-        assert(ret.toOption.get)(contains(BiolinkPredicate("related_to"))) && assert(retWithListData.toOption.get)(
-          contains(BiolinkPredicate("related_to")))
-      }
+        data = """"biolink:related_to""""
+        dataAsJson = {
+          import biolinkData.implicits._
+          decode[BiolinkPredicate](data)
+        }
+
+        dataAsList = """["biolink:participates_in","biolink:related_to"]"""
+        dataAsListAsJson = {
+          import biolinkData.implicits._
+          decode[List[BiolinkPredicate]](dataAsList)
+        }
+      } yield assert(dataAsJson)(Assertion.isRight(Assertion.equalTo(BiolinkPredicate("related_to")))) &&
+        assert(dataAsListAsJson)(
+          Assertion.isRight(
+            Assertion.contains(BiolinkPredicate("participates_in")) &&
+              Assertion.contains(BiolinkPredicate("related_to"))
+          ))
     }
   )
 

src/main/resources/application.conf

@@ -1,10 +1,11 @@
 {
-  version = "0.3-pre1"
+  version = "0.3-pre4"
   host = 0.0.0.0
   port = 8080
   port = ${?PORT}
   trapi-version = "1.3.0"
   trapi-version = ${?TRAPI_VERSION}
+  biolink-version = "v3.2.3"
   location = "http://localhost:8080"
   location = ${?LOCATION}
   sparql-endpoint = "https://cam-kp-sparql.apps.renci.org/sparql"

src/main/resources/biolink/README.md

@@ -0,0 +1,38 @@
+# Biolink mappings
+
+CAM-KP needs Biolink information in both the triplestore backend as well as the 
+frontend. This is frustrating -- if only one of them needed to know about the
+Biolink model, that would greatly simplify what we need to do here.
+
+We previously generated `predicates.csv` and `mkg-nodes.csv` from the
+[CAM Pipeline](https://github.com/ExposuresProvider/cam-pipeline) and incorporated
+them in here (via the parent directory). This was relatively easy to do when
+Biolink predicates could be mapped directly to relations in the triplestore, but
+since Biolink 3 includes predicates modified with qualifiers, this mapping is more
+complicated.
+
+This directory takes all of our predicate mapping logic and puts it into one place:
+the [`predicates.json`](./predicates.json) file in this directory is intended to provide
+the definitive set of mappings between Biolink predicate/qualifier combinations and
+triplestore relations. A single relation may be mapped to several predicate/qualifier
+combinations and vice versa.
+
+Two pieces of code interact with this file:
+1. `org.renci.cam.util.GenerateBiolinkPredicateMappings` is a standalone program that
+   can be used to regenerate this file. It currently uses the
+   [Biolink predicate_mapping.yaml](https://github.com/biolink/biolink-model/blob/master/predicate_mapping.yaml),
+   mapping information from the triplestore and a list of manual mappings added during
+   development, but in the future it may be expanded to include additional files.
+2. `org.renci.cam.domain.PredicateMappings` is a module in the code that provides a
+   programmatic interface to the contents of the `predicates.json` file, and provides
+   methods to map from Biolink predicate/qualifier combinations to relations.
+
+The `predicates.json` file can be regenerated by updating the
+Biolink version in application.conf and then running:
+
+```shell
+$ SPARQL_ENDPOINT=https://cam-kp-sparql-dev.apps.renci.org/sparql sbt "runMain org.renci.cam.util.GenerateBiolinkPredicateMappings"
+```
+
+The latest `predicates.json` file should be checked into the GitHub repository so that
+changes to it can be tracked.