X-Git-Url: http://source.jalview.org/gitweb/?a=blobdiff_plain;f=src%2Fjalview%2Fanalysis%2FSequenceIdMatcher.java;h=98886472f1de8e750b46438ddb1680a81fa206c8;hb=63cead20b09743e899a22cb89f1c09e4d41cc8c0;hp=03528dba46ef6085b8d185d510bfbaa0ada99447;hpb=5a6113ce9fef6f59b895a40a69b55b120735c533;p=jalview.git

diff --git a/src/jalview/analysis/SequenceIdMatcher.java b/src/jalview/analysis/SequenceIdMatcher.java
index 03528db..9888647 100755
--- a/src/jalview/analysis/SequenceIdMatcher.java
+++ b/src/jalview/analysis/SequenceIdMatcher.java
@@ -1,137 +1,411 @@
+/*
+ * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
+ * Copyright (C) $$Year-Rel$$ The Jalview Authors
+ * 
+ * This file is part of Jalview.
+ * 
+ * Jalview is free software: you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License 
+ * as published by the Free Software Foundation, either version 3
+ * of the License, or (at your option) any later version.
+ *  
+ * Jalview is distributed in the hope that it will be useful, but 
+ * WITHOUT ANY WARRANTY; without even the implied warranty 
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR 
+ * PURPOSE.  See the GNU General Public License for more details.
+ * 
+ * You should have received a copy of the GNU General Public License
+ * along with Jalview.  If not, see <http://www.gnu.org/licenses/>.
+ * The Jalview Authors are detailed in the 'AUTHORS' file.
+ */
 package jalview.analysis;
 
-import java.util.Vector;
-import java.util.Hashtable;
+import jalview.datamodel.DBRefEntry;
 import jalview.datamodel.SequenceI;
 
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
 /**
- * <p>Title: </p>
- * SequenceIdMatcher
- * <p>Description: </p>
- * Routine which does approximate Sequence Id resolution by name using string containment rather than equivalence
- * <p>Copyright: Copyright (c) 2004</p>
- *
- * <p>Company: Dundee University</p>
- *
- * @author not attributable
- * @version 1.0
+ * Routines for approximate Sequence Id resolution by name using string
+ * containment (on word boundaries) rather than equivalence. It also attempts to
+ * resolve ties where no exact match is available by picking the the id closest
+ * to the query.
  */
 public class SequenceIdMatcher
 {
+  private HashMap<SeqIdName, List<SequenceI>> names;
+
+  private Map<SeqIdName, List<SequenceI>> excludes;
 
-  private class SeqIdName
+  public SequenceIdMatcher(List<SequenceI> seqs)
   {
-    String id;
+    names = new HashMap<>();
+    excludes = new HashMap<>();
+    addAll(seqs);
+  }
 
-    SeqIdName(String s)
+  /**
+   * Adds sequences to this matcher
+   * 
+   * @param seqs
+   */
+  public void addAll(List<SequenceI> seqs)
+  {
+    for (SequenceI seq : seqs)
     {
-      id = new String(s);
+      add(seq);
     }
+  }
 
-    public int hashCode()
+  /**
+   * Adds one sequence to this matcher
+   * 
+   * @param seq
+   */
+  public void add(SequenceI seq)
+  {
+    SeqIdName key = new SeqIdName(seq.getDisplayId(true));
+    addMatchCandidate(key, seq);
+    SequenceI dbseq = seq;
+    while (dbseq.getDatasetSequence() != null)
     {
-      return (id.substring(0, 4).hashCode());
+      dbseq = dbseq.getDatasetSequence();
     }
-
-    public boolean equals(Object s)
+    // add in any interesting identifiers
+    if (dbseq.getDBRefs() != null)
     {
-      if (s instanceof SeqIdName)
+      DBRefEntry dbr[] = dbseq.getDBRefs();
+      for (int r = 0; r < dbr.length; r++)
       {
-        return this.equals( (SeqIdName) s);
-      }
-      else
-      {
-        if (s instanceof String)
+        DBRefEntry dbref = dbr[r];
+        SeqIdName sid = new SeqIdName(dbref.getAccessionId());
+        if (dbref.getMap() != null
+                && dbref.getMap().getMap().isTripletMap())
         {
-          return this.equals( (String) s);
+          /*
+           * 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);
       }
-      return false;
     }
+  }
 
-    public boolean equals(SeqIdName s)
+  void addMatchCandidate(SeqIdName key, SequenceI seq)
+  {
+    List<SequenceI> namesList = names.get(key);
+    if (namesList == null)
     {
-      if (id.startsWith(s.id) || s.id.startsWith(id))
-      {
-        return true;
-      }
-      return false;
+      names.put(key, namesList = new ArrayList<>());
     }
-
-    public boolean equals(String s)
+    if (!namesList.contains(seq))
     {
-      if (id.startsWith(s) || s.startsWith(id))
-      {
-        return true;
-      }
-      return false;
+      namesList.add(seq);
+      System.out.println("Adding " + key + "->" + seq);
     }
   }
 
-  private Hashtable names;
+  /**
+   * convenience method to make a matcher from concrete array
+   * 
+   * @param sequences
+   */
+  public SequenceIdMatcher(SequenceI[] sequences)
+  {
+    this(Arrays.asList(sequences));
+  }
+
+  /**
+   * returns the closest SequenceI in matches to SeqIdName and returns all the
+   * matches to the names hash.
+   * 
+   * @param candName
+   *          SeqIdName
+   * @param matches
+   *          List of SequenceI objects
+   * @return SequenceI closest SequenceI to SeqIdName
+   */
+  private SequenceI pickbestMatch(SeqIdName candName,
+          List<SequenceI> matches)
+  {
+    List<SequenceI> st = pickbestMatches(candName, matches);
+    return st == null || st.size() == 0 ? null : st.get(0);
+  }
 
-  public SequenceIdMatcher(SequenceI[] seqs)
+  /**
+   * returns the closest SequenceI in matches to SeqIdName and returns all the
+   * matches to the names hash.
+   * 
+   * @param candName
+   *          SeqIdName
+   * @param matches
+   *          Vector of SequenceI objects
+   * @return Object[] { SequenceI closest SequenceI to SeqIdName, SequenceI[]
+   *         ties }
+   */
+  private List<SequenceI> pickbestMatches(SeqIdName candName,
+          List<SequenceI> matches)
   {
-    names = new Hashtable();
-    for (int i = 0; i < seqs.length; i++)
+    List<SequenceI> best = new ArrayList<>();
+    if (candName == null || matches == null || matches.size() == 0)
+    {
+      return null;
+    }
+    SequenceI match = matches.remove(0);
+    best.add(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);
+      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
+              .abs(candlen - namlen)) && candlen > matchlen)
+      {
+        best.clear();
+        match = cand;
+        matchlen = candlen;
+        best.add(match);
+      }
+      if (q == w && candlen == matchlen)
+      {
+        // record any ties
+        best.add(cand);
+      }
+    }
+    if (best.size() == 0)
     {
-      names.put(new SeqIdName(seqs[i].getName()), seqs[i]);
+      return null;
     }
+    ;
+    return best;
   }
 
-  SequenceI findIdMatch(SequenceI seq)
+  /**
+   * get SequenceI with closest SequenceI.getName() to seq.getName()
+   * 
+   * @param seq
+   *          SequenceI
+   * @return SequenceI
+   */
+  public SequenceI findIdMatch(SequenceI seq)
   {
     SeqIdName nam = new SeqIdName(seq.getName());
-    if (names.containsKey(nam))
-    {
-      return (SequenceI) names.get(nam);
-    }
-    return null;
+    return findIdMatch(nam);
+  }
+
+  public SequenceI findIdMatch(String seqnam)
+  {
+    SeqIdName nam = new SeqIdName(seqnam);
+    return findIdMatch(nam);
   }
 
-  SequenceI findIdMatch(String seqnam)
+  /**
+   * Find all matches for a given sequence name.
+   * 
+   * @param seqnam
+   *          string to query Matcher with.
+   * @return a new array or (possibly) null
+   */
+  public SequenceI[] findAllIdMatches(String seqnam)
   {
+
     SeqIdName nam = new SeqIdName(seqnam);
-    if (names.containsKey(nam))
+    List<SequenceI> m = findAllIdMatches(nam);
+    if (m != null)
     {
-      return (SequenceI) names.get(nam);
+      return m.toArray(new SequenceI[m.size()]);
     }
     return null;
   }
 
   /**
-   * @method findIdMatch
-   *
-   * Return pointers to sequences (or sequence object containers)
-   * which have same Id as a given set of different sequence objects
-   *
-   * @param seqs SequenceI[]
+   * findIdMatch
+   * 
+   * Return pointers to sequences (or sequence object containers) which have
+   * same Id as a given set of different sequence objects
+   * 
+   * @param seqs
+   *          SequenceI[]
    * @return SequenceI[]
    */
-
-  SequenceI[] findIdMatch(SequenceI[] seqs)
+  public SequenceI[] findIdMatch(SequenceI[] seqs)
   {
-    SequenceI[] namedseqs = new SequenceI[seqs.length];
-
+    SequenceI[] namedseqs = null;
     int i = 0;
     SeqIdName nam;
+
     if (seqs.length > 0)
     {
+      namedseqs = new SequenceI[seqs.length];
       do
       {
         nam = new SeqIdName(seqs[i].getName());
+
         if (names.containsKey(nam))
         {
-          namedseqs[i] = (SequenceI) names.get(nam);
+          namedseqs[i] = findIdMatch(nam);
         }
         else
         {
           namedseqs[i] = null;
         }
-      }
-      while (i++ < seqs.length);
+      } while (++i < seqs.length);
     }
+
     return namedseqs;
   }
 
+  /**
+   * core findIdMatch search method
+   * 
+   * @param nam
+   *          SeqIdName
+   * @return SequenceI
+   */
+  private SequenceI findIdMatch(SeqIdName nam)
+  {
+    List<SequenceI> matches = new ArrayList<>();
+    while (names.containsKey(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);
+  }
+
+  /**
+   * core findIdMatch search method for finding all equivalent matches
+   * 
+   * @param nam
+   *          SeqIdName
+   * @return SequenceI[]
+   */
+  private List<SequenceI> findAllIdMatches(
+          jalview.analysis.SequenceIdMatcher.SeqIdName nam)
+  {
+    List<SequenceI> matches = new ArrayList<>();
+    while (names.containsKey(nam))
+    {
+      matches.addAll(names.remove(nam));
+    }
+    List<SequenceI> r = pickbestMatches(nam, matches);
+    return r;
+  }
+
+  class SeqIdName
+  {
+    String id;
+
+    SeqIdName(String s)
+    {
+      if (s != null)
+      {
+        id = s.toLowerCase();
+      }
+      else
+      {
+        id = "";
+      }
+    }
+
+    @Override
+    public int hashCode()
+    {
+      return ((id.length() >= 4) ? id.substring(0, 4).hashCode()
+              : id.hashCode());
+    }
+
+    @Override
+    public boolean equals(Object s)
+    {
+      if (s == null)
+      {
+        return false;
+      }
+      if (s instanceof SeqIdName)
+      {
+        return this.stringequals(((SeqIdName) s).id);
+      }
+      else
+      {
+        if (s instanceof String)
+        {
+          return this.stringequals(((String) s).toLowerCase());
+        }
+      }
+
+      return false;
+    }
+
+    /**
+     * Characters that define the end of a unique sequence ID at the beginning
+     * of an arbitrary ID string JBPNote: This is a heuristic that will fail for
+     * arbritrarily extended sequence id's (like portions of an aligned set of
+     * repeats from one sequence)
+     */
+    private String WORD_SEP = "~. |#\\/<>!\"" + ((char) 0x00A4)
+            + "$%^*)}[@',?_";
+
+    /**
+     * matches if one ID properly contains another at a whitespace boundary.
+     * TODO: (JBPNote) These are not efficient. should use char[] for speed
+     * todo: (JBPNote) Set separator characters appropriately
+     * 
+     * @param s
+     * @return boolean
+     */
+    private boolean stringequals(String s)
+    {
+      if (id.length() > s.length())
+      {
+        return id.startsWith(s)
+                ? (WORD_SEP.indexOf(id.charAt(s.length())) > -1)
+                : false;
+      }
+      else
+      {
+        return s.startsWith(id)
+                ? (s.equals(id) ? true
+                        : (WORD_SEP.indexOf(s.charAt(id.length())) > -1))
+                : false;
+      }
+    }
+
+    /**
+     * toString method returns the wrapped sequence id. For debugging purposes
+     * only, behaviour not guaranteed not to change.
+     */
+    @Override
+    public String toString()
+    {
+      return id;
+    }
+  }
 }