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
*/
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);
}
/**
- * add more sequences to this matcher - also used by the constructor
+ * Adds sequences to this matcher
*
* @param seqs
*/
{
for (SequenceI seq : seqs)
{
- // TODO: deal with ID collisions - SequenceI should be appended to list
- // associated with this key.
- names.put(new SeqIdName(seq.getDisplayId(true)), seq);
- SequenceI dbseq = seq;
- while (dbseq.getDatasetSequence() != null)
- {
- dbseq = dbseq.getDatasetSequence();
- }
- // add in any interesting identifiers
- if (dbseq.getDBRefs() != null)
+ add(seq);
+ }
+ }
+
+ /**
+ * 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)
+ {
+ dbseq = dbseq.getDatasetSequence();
+ }
+ // add in any interesting identifiers
+ if (dbseq.getDBRefs() != null)
+ {
+ DBRefEntry dbr[] = dbseq.getDBRefs();
+ for (int r = 0; r < dbr.length; r++)
{
- DBRefEntry dbr[] = dbseq.getDBRefs();
- SeqIdName sid = null;
- for (int r = 0; r < dbr.length; r++)
+ DBRefEntry dbref = dbr[r];
+ SeqIdName sid = new SeqIdName(dbref.getAccessionId());
+ if (dbref.getMap() != null
+ && dbref.getMap().getMap().isTripletMap())
{
- sid = new SeqIdName(dbr[r].getAccessionId());
- if (!names.containsKey(sid))
+ /*
+ * 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)
{
- names.put(sid, seq);
+ 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
*
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.abs(candlen
- - namlen))
- && candlen > matchlen)
+ if ((q = Math.abs(matchlen - namlen)) > (w = Math
+ .abs(candlen - namlen)) && candlen > matchlen)
{
best.clear();
match = cand;
* 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);
}
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;
}
- private class SeqIdName
+ class SeqIdName
{
String id;
{
if (s != null)
{
- id = new String(s);
+ id = s.toLowerCase();
}
else
{
@Override
public int hashCode()
{
- return ((id.length() >= 4) ? id.substring(0, 4).hashCode() : id
- .hashCode());
+ return ((id.length() >= 4) ? id.substring(0, 4).hashCode()
+ : id.hashCode());
}
@Override
}
if (s instanceof SeqIdName)
{
- return this.equals(((SeqIdName) s).id);
+ return this.stringequals(((SeqIdName) s).id);
}
else
{
if (s instanceof String)
{
- return this.equals((String) s);
+ return this.stringequals(((String) s).toLowerCase());
}
}
* @param s
* @return boolean
*/
- public boolean equals(String s)
+ private boolean stringequals(String s)
{
if (id.length() > s.length())
{
- return id.startsWith(s) ? (WORD_SEP.indexOf(id.charAt(s.length())) > -1)
+ 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;
+ 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;
+ }
}
}
git://source.jalview.org / jalview.git / blobdiff