From 63cead20b09743e899a22cb89f1c09e4d41cc8c0 Mon Sep 17 00:00:00 2001
From: gmungoc <g.m.carstairs@dundee.ac.uk>
Date: Tue, 12 Feb 2019 08:44:08 +0000
Subject: [PATCH] JAL-3195 revised SequenceIdMatcher (tests todo)

---
 src/jalview/analysis/SequenceIdMatcher.java |   76 ++++++++++++++++++++-------
 src/jalview/io/FeaturesFile.java            |    2 +-
 2 files changed, 58 insertions(+), 20 deletions(-)

diff --git a/src/jalview/analysis/SequenceIdMatcher.java b/src/jalview/analysis/SequenceIdMatcher.java
index 3d4cbe7..9888647 100755
--- a/src/jalview/analysis/SequenceIdMatcher.java
+++ b/src/jalview/analysis/SequenceIdMatcher.java
@@ -27,7 +27,7 @@ import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.HashMap;
 import java.util.List;
-import java.util.Vector;
+import java.util.Map;
 
 /**
  * Routines for approximate Sequence Id resolution by name using string
@@ -37,11 +37,14 @@ import java.util.Vector;
  */
 public class SequenceIdMatcher
 {
-  private HashMap<SeqIdName, SequenceI> names;
+  private HashMap<SeqIdName, List<SequenceI>> names;
+
+  private Map<SeqIdName, List<SequenceI>> excludes;
 
   public SequenceIdMatcher(List<SequenceI> seqs)
   {
-    names = new HashMap<SeqIdName, SequenceI>();
+    names = new HashMap<>();
+    excludes = new HashMap<>();
     addAll(seqs);
   }
 
@@ -65,9 +68,8 @@ public class SequenceIdMatcher
    */
   public void add(SequenceI seq)
   {
-    // TODO: deal with ID collisions - SequenceI should be appended to list
-    // associated with this key.
-    names.put(new SeqIdName(seq.getDisplayId(true)), seq);
+    SeqIdName key = new SeqIdName(seq.getDisplayId(true));
+    addMatchCandidate(key, seq);
     SequenceI dbseq = seq;
     while (dbseq.getDatasetSequence() != null)
     {
@@ -77,18 +79,45 @@ public class SequenceIdMatcher
     if (dbseq.getDBRefs() != null)
     {
       DBRefEntry dbr[] = dbseq.getDBRefs();
-      SeqIdName sid = null;
       for (int r = 0; r < dbr.length; r++)
       {
-        sid = new SeqIdName(dbr[r].getAccessionId());
-        if (!names.containsKey(sid))
+        DBRefEntry dbref = dbr[r];
+        SeqIdName sid = new SeqIdName(dbref.getAccessionId());
+        if (dbref.getMap() != null
+                && dbref.getMap().getMap().isTripletMap())
         {
-          names.put(sid, seq);
+          /*
+           * dbref with 3:1 or 1:3 mapping (e.g. CDS/protein);
+           * mark as not a valid match for this id
+           */
+          List<SequenceI> excluded = excludes.get(sid);
+          if (excluded == null)
+          {
+            excludes.put(sid, excluded = new ArrayList<>());
+          }
+          excluded.add(seq);
+          System.out.println("Excluding " + sid + "->" + seq);
+          continue;
         }
+        addMatchCandidate(sid, seq);
       }
     }
   }
 
+  void addMatchCandidate(SeqIdName key, SequenceI seq)
+  {
+    List<SequenceI> namesList = names.get(key);
+    if (namesList == null)
+    {
+      names.put(key, namesList = new ArrayList<>());
+    }
+    if (!namesList.contains(seq))
+    {
+      namesList.add(seq);
+      System.out.println("Adding " + key + "->" + seq);
+    }
+  }
+
   /**
    * convenience method to make a matcher from concrete array
    * 
@@ -130,21 +159,21 @@ public class SequenceIdMatcher
   private List<SequenceI> pickbestMatches(SeqIdName candName,
           List<SequenceI> matches)
   {
-    ArrayList<SequenceI> best = new ArrayList<SequenceI>();
+    List<SequenceI> best = new ArrayList<>();
     if (candName == null || matches == null || matches.size() == 0)
     {
       return null;
     }
     SequenceI match = matches.remove(0);
     best.add(match);
-    names.put(new SeqIdName(match.getName()), match);
+    addMatchCandidate(new SeqIdName(match.getName()), match);
     int matchlen = match.getName().length();
     int namlen = candName.id.length();
     while (matches.size() > 0)
     {
       // look through for a better one.
       SequenceI cand = matches.remove(0);
-      names.put(new SeqIdName(cand.getName()), cand);
+      addMatchCandidate(new SeqIdName(cand.getName()), cand);
       int q, w, candlen = cand.getName().length();
       // keep the one with an id 'closer' to the given seqnam string
       if ((q = Math.abs(matchlen - namlen)) > (w = Math
@@ -251,13 +280,22 @@ public class SequenceIdMatcher
    *          SeqIdName
    * @return SequenceI
    */
-  private SequenceI findIdMatch(
-          jalview.analysis.SequenceIdMatcher.SeqIdName nam)
+  private SequenceI findIdMatch(SeqIdName nam)
   {
-    Vector matches = new Vector();
+    List<SequenceI> matches = new ArrayList<>();
     while (names.containsKey(nam))
     {
-      matches.addElement(names.remove(nam));
+      List<SequenceI> candidates = names.remove(nam);
+      List<SequenceI> except = excludes.get(nam);
+      int j = candidates.size();
+      for (int i = 0; i < j; i++)
+      {
+        SequenceI candidate = candidates.get(i);
+        if (!except.contains(candidate))
+        {
+          matches.add(candidate);
+        }
+      }
     }
     return pickbestMatch(nam, matches);
   }
@@ -272,10 +310,10 @@ public class SequenceIdMatcher
   private List<SequenceI> findAllIdMatches(
           jalview.analysis.SequenceIdMatcher.SeqIdName nam)
   {
-    ArrayList<SequenceI> matches = new ArrayList<SequenceI>();
+    List<SequenceI> matches = new ArrayList<>();
     while (names.containsKey(nam))
     {
-      matches.add(names.remove(nam));
+      matches.addAll(names.remove(nam));
     }
     List<SequenceI> r = pickbestMatches(nam, matches);
     return r;
diff --git a/src/jalview/io/FeaturesFile.java b/src/jalview/io/FeaturesFile.java
index 169da5a..559ca79 100755
--- a/src/jalview/io/FeaturesFile.java
+++ b/src/jalview/io/FeaturesFile.java
@@ -393,7 +393,7 @@ public class FeaturesFile extends AlignFile implements FeaturesSourceI
     }
     String desc = gffColumns[0];
     String seqId = gffColumns[1];
-    SequenceI seq = findSequence(seqId, alignment, null, relaxedIdMatching);
+    SequenceI seq;
 
     if (!ID_NOT_SPECIFIED.equals(seqId))
     {
-- 
1.7.10.2