JAL-3621 corrected sequence feature sort and test
[jalview.git] / src / jalview / datamodel / features / SequenceFeatures.java
index 6165d0a..8ac4991 100644 (file)
@@ -1,15 +1,38 @@
+/*
+ * 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.datamodel.features;
 
-import jalview.datamodel.SequenceFeature;
-
 import java.util.ArrayList;
-import java.util.Arrays;
-import java.util.HashMap;
+import java.util.Collections;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Map;
 import java.util.Map.Entry;
 import java.util.Set;
+import java.util.TreeMap;
+
+import intervalstore.api.IntervalI;
+import jalview.datamodel.SequenceFeature;
+import jalview.io.gff.SequenceOntologyFactory;
+import jalview.io.gff.SequenceOntologyI;
 
 /**
  * A class that stores sequence features in a way that supports efficient
@@ -19,9 +42,8 @@ import java.util.Set;
  * @author gmcarstairs
  *
  */
-public class SequenceFeatures
+public class SequenceFeatures implements SequenceFeaturesI
 {
-
   /*
    * map from feature type to structured store of features for that type
    * null types are permitted (but not a good idea!)
@@ -33,20 +55,42 @@ public class SequenceFeatures
    */
   public SequenceFeatures()
   {
-    featureStore = new HashMap<String, FeatureStore>();
+    /*
+     * use a TreeMap so that features are returned in alphabetical order of type
+     * ? wrap as a synchronized map for add and delete operations
+     */
+    // featureStore = Collections
+    // .synchronizedSortedMap(new TreeMap<String, FeatureStore>());
+    featureStore = new TreeMap<>();
   }
 
   /**
-   * Adds one sequence feature to the store, and returns true, unless the
-   * feature is already contained in the store, in which case this method
-   * returns false. Containment is determined by SequenceFeature.equals()
-   * comparison.
-   * 
-   * @param sf
+   * Constructor given a list of features
    */
+  public SequenceFeatures(List<SequenceFeature> features)
+  {
+    this();
+    if (features != null)
+    {
+      for (SequenceFeature feature : features)
+      {
+        add(feature);
+      }
+    }
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
   public boolean add(SequenceFeature sf)
   {
     String type = sf.getType();
+    if (type == null)
+    {
+      System.err.println("Feature type may not be null: " + sf.toString());
+      return false;
+    }
 
     if (featureStore.get(type) == null)
     {
@@ -56,138 +100,170 @@ public class SequenceFeatures
   }
 
   /**
-   * Returns a (possibly empty) list of features, optionally restricted to
-   * specified types, which overlap the given (inclusive) sequence position
-   * range
-   * 
-   * @param from
-   * @param to
-   * @param type
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public List<SequenceFeature> findFeatures(int from, int to,
           String... type)
   {
-    List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+    List<SequenceFeature> result = new ArrayList<>();
 
-    for (String featureType : varargToTypes(type))
+    for (FeatureStore featureSet : varargToTypes(type))
     {
-      FeatureStore features = featureStore.get(featureType);
-      if (features != null)
-      {
-        result.addAll(features.findOverlappingFeatures(from, to));
-      }
+      result.addAll(featureSet.findOverlappingFeatures(from, to));
     }
 
     return result;
   }
 
   /**
-   * Answers a list of all features stored, optionally restricted to specified
-   * types, in no particular guaranteed order
-   * 
-   * @param type
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public List<SequenceFeature> getAllFeatures(String... type)
   {
-    List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+    List<SequenceFeature> result = new ArrayList<>();
 
     result.addAll(getPositionalFeatures(type));
 
-    result.addAll(getNonPositionalFeatures(type));
+    result.addAll(getNonPositionalFeatures());
 
     return result;
   }
 
   /**
-   * Answers a list of all positional features, optionally restricted to
-   * specified types, in no particular guaranteed order
-   * 
-   * @param type
-   * @return
+   * {@inheritDoc}
    */
+  @Override
+  public List<SequenceFeature> getFeaturesByOntology(String... ontologyTerm)
+  {
+    if (ontologyTerm == null || ontologyTerm.length == 0)
+    {
+      return new ArrayList<>();
+    }
+
+    Set<String> featureTypes = getFeatureTypes(ontologyTerm);
+    if (featureTypes.isEmpty())
+    {
+      /*
+       * no features of the specified type or any sub-type
+       */
+      return new ArrayList<>();
+    }
+
+    return getAllFeatures(featureTypes.toArray(new String[featureTypes
+            .size()]));
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public int getFeatureCount(boolean positional, String... type)
+  {
+    int result = 0;
+
+    for (FeatureStore featureSet : varargToTypes(type))
+    {
+      result += featureSet.getFeatureCount(positional);
+    }
+    return result;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public int getTotalFeatureLength(String... type)
+  {
+    int result = 0;
+
+    for (FeatureStore featureSet : varargToTypes(type))
+    {
+      result += featureSet.getTotalFeatureLength();
+    }
+    return result;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
   public List<SequenceFeature> getPositionalFeatures(String... type)
   {
-    List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+    List<SequenceFeature> result = new ArrayList<>();
 
-    for (String featureType : varargToTypes(type))
+    for (FeatureStore featureSet : varargToTypes(type))
     {
-      FeatureStore featureSet = featureStore.get(featureType);
-      if (featureSet != null)
-      {
-        result.addAll(featureSet.getPositionalFeatures());
-      }
+      result.addAll(featureSet.getPositionalFeatures());
     }
     return result;
   }
 
   /**
    * A convenience method that converts a vararg for feature types to an
-   * Iterable, replacing the value with the stored feature types if it is null
-   * or empty
+   * Iterable over matched feature sets. If no types are specified, all feature
+   * sets are returned. If one or more types are specified, feature sets for
+   * those types are returned, preserving the order of the types.
    * 
    * @param type
    * @return
    */
-  protected Iterable<String> varargToTypes(String... type)
+  protected Iterable<FeatureStore> varargToTypes(String... type)
   {
-    return type == null || type.length == 0 ? featureStore
-            .keySet() : Arrays.asList(type);
+    if (type == null || type.length == 0)
+    {
+      /*
+       * no vararg parameter supplied - return all
+       */
+      return featureStore.values();
+    }
+
+    List<FeatureStore> types = new ArrayList<>();
+    for (String theType : type)
+    {
+      if (theType != null && featureStore.containsKey(theType))
+      {
+        types.add(featureStore.get(theType));
+      }
+    }
+    return types;
   }
 
   /**
-   * Answers a list of all contact features, optionally restricted to specified
-   * types, in no particular guaranteed order
-   * 
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public List<SequenceFeature> getContactFeatures(String... type)
   {
-    List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+    List<SequenceFeature> result = new ArrayList<>();
 
-    for (String featureType : varargToTypes(type))
+    for (FeatureStore featureSet : varargToTypes(type))
     {
-      FeatureStore featureSet = featureStore.get(featureType);
-      if (featureSet != null)
-      {
-        result.addAll(featureSet.getContactFeatures());
-      }
+      result.addAll(featureSet.getContactFeatures());
     }
     return result;
   }
 
   /**
-   * Answers a list of all non-positional features, optionally restricted to
-   * specified types, in no particular guaranteed order
-   * 
-   * @param type
-   *          if no type is specified, all are returned
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public List<SequenceFeature> getNonPositionalFeatures(String... type)
   {
-    List<SequenceFeature> result = new ArrayList<SequenceFeature>();
+    List<SequenceFeature> result = new ArrayList<>();
 
-    for (String featureType : varargToTypes(type))
+    for (FeatureStore featureSet : varargToTypes(type))
     {
-      FeatureStore featureSet = featureStore.get(featureType);
-      if (featureSet != null)
-      {
-        result.addAll(featureSet.getNonPositionalFeatures());
-      }
+      result.addAll(featureSet.getNonPositionalFeatures());
     }
     return result;
   }
 
   /**
-   * Deletes the given feature from the store, returning true if it was found
-   * (and deleted), else false. This method makes no assumption that the feature
-   * is in the 'expected' place in the store, in case it has been modified since
-   * it was added.
-   * 
-   * @param sf
+   * {@inheritDoc}
    */
+  @Override
   public boolean delete(SequenceFeature sf)
   {
     for (FeatureStore featureSet : featureStore.values())
@@ -201,10 +277,9 @@ public class SequenceFeatures
   }
 
   /**
-   * Answers true if this store contains at least one feature, else false
-   * 
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public boolean hasFeatures()
   {
     for (FeatureStore featureSet : featureStore.values())
@@ -218,49 +293,30 @@ public class SequenceFeatures
   }
 
   /**
-   * Returns a set of the distinct feature groups present in the collection. The
-   * set may include null. The boolean parameter determines whether the groups
-   * for positional or for non-positional features are returned. The optional
-   * type parameter may be used to restrict to groups for specified feature
-   * types.
-   * 
-   * @param positionalFeatures
-   * @param type
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public Set<String> getFeatureGroups(boolean positionalFeatures,
           String... type)
   {
-    Set<String> groups = new HashSet<String>();
+    Set<String> groups = new HashSet<>();
 
-    Iterable<String> types = varargToTypes(type);
-
-    for (String featureType : types)
+    for (FeatureStore featureSet : varargToTypes(type))
     {
-      FeatureStore featureSet = featureStore.get(featureType);
-      if (featureSet != null)
-      {
-        groups.addAll(featureSet.getFeatureGroups(positionalFeatures));
-      }
+      groups.addAll(featureSet.getFeatureGroups(positionalFeatures));
     }
 
     return groups;
   }
 
   /**
-   * Answers the set of distinct feature types for which there is at least one
-   * feature with one of the given feature group(s). The parameter determines
-   * whether the groups for positional or for non-positional features are
-   * returned.
-   * 
-   * @param positionalFeatures
-   * @param groups
-   * @return
+   * {@inheritDoc}
    */
+  @Override
   public Set<String> getFeatureTypesForGroups(boolean positionalFeatures,
           String... groups)
   {
-    Set<String> result = new HashSet<String>();
+    Set<String> result = new HashSet<>();
 
     for (Entry<String, FeatureStore> featureType : featureStore.entrySet())
     {
@@ -281,4 +337,134 @@ public class SequenceFeatures
 
     return result;
   }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public Set<String> getFeatureTypes(String... soTerm)
+  {
+    Set<String> types = new HashSet<>();
+    for (Entry<String, FeatureStore> entry : featureStore.entrySet())
+    {
+      String type = entry.getKey();
+      if (!entry.getValue().isEmpty() && isOntologyTerm(type, soTerm))
+      {
+        types.add(type);
+      }
+    }
+    return types;
+  }
+
+  /**
+   * Answers true if the given type matches one of the specified terms (or is a
+   * sub-type of one in the Sequence Ontology), or if no terms are supplied.
+   * Answers false if filter terms are specified and the given term does not
+   * match any of them.
+   * 
+   * @param type
+   * @param soTerm
+   * @return
+   */
+  protected boolean isOntologyTerm(String type, String... soTerm)
+  {
+    if (soTerm == null || soTerm.length == 0)
+    {
+      return true;
+    }
+    SequenceOntologyI so = SequenceOntologyFactory.getInstance();
+    for (String term : soTerm)
+    {
+      if (type.equals(term) || so.isA(type, term))
+      {
+        return true;
+      }
+    }
+    return false;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public float getMinimumScore(String type, boolean positional)
+  {
+    return featureStore.containsKey(type) ? featureStore.get(type)
+            .getMinimumScore(positional) : Float.NaN;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public float getMaximumScore(String type, boolean positional)
+  {
+    return featureStore.containsKey(type) ? featureStore.get(type)
+            .getMaximumScore(positional) : Float.NaN;
+  }
+
+  /**
+   * A convenience method to sort features by start position ascending (if on
+   * forward strand), or end position descending (if on reverse strand)
+   * 
+   * @param features
+   * @param forwardStrand
+   */
+  public static void sortFeatures(List<? extends IntervalI> features,
+          final boolean forwardStrand)
+  {
+    Collections.sort(features,
+            forwardStrand
+                    ? IntervalI.COMPARE_BEGIN_ASC_END_DESC
+                    : IntervalI.COMPARE_END_DESC);
+  }
+
+  /**
+   * {@inheritDoc} This method is 'semi-optimised': it only inspects features
+   * for types that include the specified group, but has to inspect every
+   * feature of those types for matching feature group. This is efficient unless
+   * a sequence has features that share the same type but are in different
+   * groups - an unlikely case.
+   * <p>
+   * For example, if RESNUM feature is created with group = PDBID, then features
+   * would only be retrieved for those sequences associated with the target
+   * PDBID (group).
+   */
+  @Override
+  public List<SequenceFeature> getFeaturesForGroup(boolean positional,
+          String group, String... type)
+  {
+    List<SequenceFeature> result = new ArrayList<>();
+    for (FeatureStore featureSet : varargToTypes(type))
+    {
+      if (featureSet.getFeatureGroups(positional).contains(group))
+      {
+        result.addAll(featureSet.getFeaturesForGroup(positional, group));
+      }
+    }
+    return result;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public boolean shiftFeatures(int fromPosition, int shiftBy)
+  {
+    boolean modified = false;
+    for (FeatureStore fs : featureStore.values())
+    {
+      modified |= fs.shiftFeatures(fromPosition, shiftBy);
+    }
+    return modified;
+  }
+
+  /**
+   * {@inheritDoc}
+   */
+  @Override
+  public void deleteAll()
+  {
+    featureStore.clear();
+  }
 }