Resolve Xrefs equivalence

src/OBO.NET/DBXref.fs

@@ -1,16 +1,40 @@
 namespace OBO.NET
 
 
+open FSharpAux
+open ControlledVocabulary
+
 open System
 
 
-/// Representation of dbxrefs.
-type DBXref = {
+/// Representation of DBXrefs.
+type DBXref = 
+    {
         Name        : string
         Description : string
         Modifiers   : string
     }
 
+    /// Parses a given string to a DBXref
+    static member ofString (v : string) =
+        let xrefRegex = Text.RegularExpressions.Regex("""(?<xrefName>^([^"{])*)(\s?)(?<xrefDescription>\"(.*?)\")?(\s?)(?<xrefModifiers>\{(.*?)}$)?""")
+        let matches = xrefRegex.Match(v.Trim()).Groups
+        {
+            Name        = matches.Item("xrefName")          .Value |> String.trim
+            Description = matches.Item("xrefDescription")   .Value |> String.trim
+            Modifiers   = matches.Item("xrefModifiers")     .Value |> String.trim
+        }
+
+    /// Returns the corresponding CvTerm of the DBXref with empty name.
+    member this.ToCvTerm() = {
+        Name        = ""
+        Accession   = this.Name
+        RefUri      = 
+            String.split ':' this.Name
+            |> Array.head
+            |> String.trim
+    }
+
 
 /// Functions for working with DBXrefs.
 module DBXref =
@@ -37,16 +61,15 @@ module DBXref =
 
     //Note that the trailing modifiers (like all trailing modifiers) do not need to be decoded or round-tripped by parsers; trailing modifiers can always be optionally ignored. However, all parsers must be able to gracefully ignore trailing modifiers. It is important to recognize that lines which accept a dbxref list may have a trailing modifier for each dbxref in the list, and another trailing modifier for the line itself.
 
+    // EXAMPLE (taken from GO_Slim Agr): "xref: RO:0002093"
+
     let trimComment (line : string) = 
         line.Split('!').[0].Trim()
 
-    let private xrefRegex = 
-        Text.RegularExpressions.Regex("""(?<xrefName>^([^"{])*)(\s?)(?<xrefDescription>\"(.*?)\")?(\s?)(?<xrefModifiers>\{(.*?)}$)?""")
-
+    [<Obsolete "Use `DBXref.ofString` instead">]
     let parseDBXref (v : string) =
-        let matches = xrefRegex.Match(v.Trim()).Groups
-        {
-            Name = matches.Item("xrefName").Value
-            Description = matches.Item("xrefDescription").Value
-            Modifiers = matches.Item("xrefModifiers").Value
-        }
+        DBXref.ofString v
+
+    /// Creates a CvTerm (with an empty name) of a given DBXref.
+    let toCvTerm (dbxref : DBXref) =
+        dbxref.ToCvTerm()

src/OBO.NET/OBO.NET.fsproj

@@ -38,6 +38,7 @@
 
   <ItemGroup>
     <PackageReference Include="ARCtrl.ISA" Version="1.0.0-beta.7" />
+    <PackageReference Include="ControlledVocabulary" Version="1.0.0" />
     <PackageReference Include="Microsoft.SourceLink.GitHub" Version="[1.1.1]" PrivateAssets="All" />
     <PackageReference Include="FSharpAux" Version="[2.0.0]" />
   </ItemGroup>

src/OBO.NET/OboOntology.fs

@@ -519,6 +519,47 @@ type OboOntology =
     static member getSynonyms term (onto : OboOntology) =
         onto.GetSynonyms term
 
+    /// <summary>Checks if the given terms are treated as equivalent in the OboOntology.</summary>
+    /// <remarks>Note that term1 must be part of the OboOntology while term2 must be part of another ontology.</remarks>
+    member this.AreTermsEquivalent(term1 : OboTerm, term2 : OboTerm) =
+        term1.Xrefs 
+        |> List.exists (
+            fun x1 -> 
+                term2.Id = x1.Name
+                &&
+                this.TreatXrefsAsEquivalents
+                |> List.exists (fun t -> t = (term2.ToCvTerm()).RefUri)
+        )
+
+    /// <summary>Checks if the given terms are treated as equivalent in the given OboOntology.</summary>
+    /// <remarks>Note that term1 must be part of the given OboOntology while term2 must be part of another ontology.</remarks>
+    static member areTermsEquivalent term1 term2 (onto : OboOntology) =
+        onto.AreTermsEquivalent(term1, term2)
+
+    /// Returns all terms of the OboOntology that have equivalent terms in a given second OboOntology.
+    member this.ReturnAllEquivalentTerms(onto : OboOntology) =
+        if List.exists (fun x -> x = onto.Terms.Head.ToCvTerm().RefUri) this.TreatXrefsAsEquivalents then
+            this.Terms
+            |> Seq.filter (
+                fun t ->
+                    List.isEmpty t.Xrefs |> not
+                    &&
+                    t.Xrefs
+                    |> List.exists (
+                        fun x -> 
+                            onto.Terms
+                            |> List.exists (
+                                fun t2 -> 
+                                    (DBXref.toCvTerm x).Accession = t2.Id
+                            ) 
+                    ) 
+            )
+        else Seq.empty
+
+    /// Returns all terms of the first given OboOntology that have equivalent terms in the second given OboOntology.
+    static member returnAllEquivalentTerms (onto1 : OboOntology) onto2 =
+        onto1.ReturnAllEquivalentTerms(onto2)
+
 
 type OboTermDef = 
     {

src/OBO.NET/OboTerm.fs

@@ -3,10 +3,12 @@
 open DBXref
 open TermSynonym
 
-open ARCtrl.ISA
-
 open System
 
+open ARCtrl.ISA
+open ControlledVocabulary
+open FSharpAux
+
 
 /// Models the entities in an OBO Ontology.
 type OboTerm = 
@@ -335,7 +337,7 @@
                     propertyValues builtIn createdBy creationDate
 
             | "xref" | "xref_analog" | "xref_unk" -> 
                 let v = (split.[1..] |> String.concat ": ") |> parseDBXref
                 OboTerm.fromLines verbose en (lineNumber + 1)
                     id name isAnonymous altIds definition comment subsets synonyms (v::xrefs) isA 
                     intersectionOf unionOf disjointFrom relationships isObsolete replacedby consider 
@@ -540,6 +542,18 @@
     static member getRelatedTermIds (term : OboTerm) =
         term.GetRelatedTermIds()
 
+    /// Returns the corresponding CvTerm of the OboTerm.
+    member this.ToCvTerm() =
+        let tsr = 
+            String.split ':' this.Id
+            |> Array.head
+            |> String.trim
+        CvTerm.create(this.Id, this.Name, tsr)
+
+    /// Returns the corresponding CvTerm of the given OboTerm.
+    static member toCvTerm (term : OboTerm) =
+        term.ToCvTerm()
+
 
 /// Representation of a the relation an OboTerm can have with other OboTerms.
 type TermRelation<'a> =

tests/OBO.NET.Tests/DBXref.Tests.fs

@@ -0,0 +1,35 @@
+namespace OBO.NET.Tests
+
+
+open OBO.NET
+
+open Expecto
+open ControlledVocabulary
+
+
+module DBXref =
+
+    [<Tests>]
+    let dbxrefTests =
+        testList "DBXref" [
+
+            let testDBXref = DBXref.ofString """test:1 "testDesc" {testMod}"""
+
+            testList "ofString" [
+                testCase "returns correct DBXref" <| fun _ ->
+                    let expected = {Name = "test:1"; Description = "\"testDesc\""; Modifiers = "{testMod}"}
+                    Expect.equal testDBXref.Name expected.Name "Name does not match"
+                    Expect.equal testDBXref.Description expected.Description "Description does not match"
+                    Expect.equal testDBXref.Modifiers expected.Modifiers "Modifiers do not match"
+            ]
+
+            testList "ToCvTerm" [
+                testCase "returns correct CvTerm" <| fun _ ->
+                    let actual = testDBXref.ToCvTerm()
+                    let expected = CvTerm.create("test:1", "", "test")
+                    Expect.equal actual.Accession expected.Accession "TANs are not equal"
+                    Expect.equal actual.RefUri expected.RefUri "TSRs are not equal"
+                    Expect.equal actual.Name expected.Name "Names are not equal"
+            ]
+
+        ]

tests/OBO.NET.Tests/OBO.NET.Tests.fsproj

@@ -12,6 +12,7 @@
     <None Include="References\IncorrectHeaderTags.obo" />
     <None Include="References\DuplicateHeaderTags.obo" />
     <None Include="references\CorrectHeaderTags.obo" />
+    <Compile Include="DBXref.Tests.fs" />
     <Compile Include="TermSynonym.Tests.fs" />
     <Compile Include="OboTerm.Tests.fs" />
     <Compile Include="OboOntology.Tests.fs" />
@@ -32,6 +33,4 @@
     <PackageReference Update="FSharp.Core" Version="7.0.401" />
   </ItemGroup>
 
-  <ItemGroup />
-
 </Project>

tests/OBO.NET.Tests/OboOntology.Tests.fs

@@ -36,14 +36,20 @@ module OboOntologyTests =
             let testTerm4 =
                 OboTerm.Create(
                     "id:5",
-                    Name = "testTerm4"
+                    Name = "testTerm4",
+                    Xrefs = [DBXref.ofString "check:1"]
                 )
             let testTerm5 =
                 OboTerm.Create(
                     "id:6",
                     Name = "testTerm5",
                     Synonyms = [TermSynonym.parseSynonym None 0 "\"testTerm1\" EXACT []"; TermSynonym.parseSynonym None 1 "\"testTerm2\" BROAD []"; TermSynonym.parseSynonym None 2 "\"testTerm0\" NARROW []"]
                 )
+            let testTerm6 =
+                OboTerm.Create(
+                    "check:1",
+                    Name = "checkTerm1"
+                )
 
             let testFile1Path = Path.Combine(__SOURCE_DIRECTORY__, "References", "CorrectHeaderTags.obo")
             let testFile2Path = Path.Combine(__SOURCE_DIRECTORY__, "References", "IncorrectHeaderTags.obo")
@@ -125,7 +131,8 @@ module OboOntologyTests =
                     Expect.equal (Option.map (fun o -> o.TypeDefs) testFile1) typedefsExpected "Terms did not match"
             ]
 
-            let testOntology = OboOntology.Create([testTerm1; testTerm2; testTerm3; testTerm4; testTerm5], [], "")
+            let testOntology = OboOntology.Create([testTerm1; testTerm2; testTerm3; testTerm4; testTerm5], [], "", TreatXrefsAsEquivalents = ["check"])
+            let testOntology2 = OboOntology.Create([testTerm6], [], "")
 
             testList "GetRelatedTerms" [
                 testCase "returns correct related terms" <| fun _ ->
@@ -171,4 +178,15 @@ module OboOntologyTests =
                     let expected = seq {Exact, testTerm5, Some testTerm1; Broad, testTerm5, Some testTerm2; Narrow, testTerm5, None}
                     Expect.sequenceEqual actual expected "is not equal"
             ]
+
+            testList "AreTermsEquivalent" [
+                testCase "checks equivalence correctly" <| fun _ ->
+                    Expect.isTrue (testOntology.AreTermsEquivalent(testTerm4, testTerm6)) "is not equal"
+            ]
+
+            testList "ReturnAllEquivalentTerms" [
+                testCase "returns correct terms" <| fun _ ->
+                    let actual = testOntology.ReturnAllEquivalentTerms(testOntology2)
+                    Expect.sequenceEqual actual (seq {testTerm4}) "is not equal"
+            ]
         ]

tests/OBO.NET.Tests/OboTerm.Tests.fs

@@ -2,10 +2,13 @@ namespace OBO.NET.Tests
 
 open Expecto
 open OBO.NET
+open ControlledVocabulary
 
 
 module OboTermTests =
 
+    let testTerm1 = OboTerm.Create("test:001", Name = "TestTerm1")
+
     [<Tests>]
     let oboTermTest =
         testList "OboTerm" [
@@ -35,4 +38,12 @@ module OboTermTests =
                     ]
                     Expect.sequenceEqual actual expected ""
             ]
+            testList "ToCvTerm" [
+                testCase "returns correct CvTerm" <| fun _ ->
+                    let actual = OboTerm.toCvTerm testTerm1
+                    let expected = CvTerm.create("test:001", "TestTerm1", "test")
+                    Expect.equal actual.RefUri expected.RefUri "TSRs are different"
+                    Expect.equal actual.Name expected.Name "Names are different"
+                    Expect.equal actual.Accession expected.Accession "TANs are different"
+            ]
         ]