+/*
+ * 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.ContiguousI;
import jalview.datamodel.SequenceFeature;
import java.util.ArrayList;
+import java.util.Collection;
import java.util.Collections;
-import java.util.Comparator;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
-/**
- * A data store for a set of sequence features that supports efficient lookup of
- * features overlapping a given range. Intended for (but not limited to) storage
- * of features for one sequence and feature type.
- *
- * @author gmcarstairs
- *
- */
-public class FeatureStore
+public abstract class FeatureStore implements FeatureStoreI
{
+
+ /**
+ * track last start for quick insertion of ordered features
+ */
+ protected int lastStart = -1, lastContactStart = -1;
+
/**
- * a class providing criteria for performing a binary search of a list
+ * Answers the 'length' of the feature, counting 0 for non-positional features
+ * and 1 for contact features
+ *
+ * @param feature
+ * @return
*/
- abstract static class SearchCriterion
+ protected static int getFeatureLength(SequenceFeature feature)
{
- /**
- * Answers true if the entry passes the search criterion test
- *
- * @param entry
- * @return
- */
- abstract boolean compare(SequenceFeature entry);
-
- /**
- * serves a search condition for finding the first feature whose start
- * position follows a given target location
- *
- * @param target
- * @return
- */
- static SearchCriterion byStart(final long target)
+ if (feature.isNonPositional())
+ {
+ return 0;
+ }
+ if (feature.isContactFeature())
{
- return new SearchCriterion() {
+ return 1;
+ }
+ return 1 + feature.getEnd() - feature.getBegin();
+ }
- @Override
- boolean compare(SequenceFeature entry)
- {
- return entry.getBegin() >= target;
- }
- };
+ /**
+ * Answers true if the list contains the feature, else false. This method is
+ * optimised for the condition that the list is sorted on feature start
+ * position ascending, and will give unreliable results if this does not hold.
+ *
+ * @param list
+ * @param feature
+ * @return
+ */
+ @Override
+ public boolean listContains(List<SequenceFeature> list,
+ SequenceFeature feature)
+ {
+ if (list == null || feature == null)
+ {
+ return false;
}
- /**
- * serves a search condition for finding the first feature whose end
- * position is at or follows a given target location
- *
- * @param target
- * @return
+ return (getEquivalentFeatureIndex(list, feature) >= 0);
+ }
+
+ /**
+ * Binary search for the index (>= 0) of a feature in a list.
+ *
+ * @param list
+ * @param feature
+ * @return index if found; -1 if not
+ */
+ protected int getEquivalentFeatureIndex(List<SequenceFeature> list,
+ SequenceFeature feature)
+ {
+
+ /*
+ * locate the first entry in the list which does not precede the feature
*/
- static SearchCriterion byEnd(final long target)
+ int begin = feature.begin;
+ int pos = findFirstBegin(list, begin);
+ int len = list.size();
+ while (pos < len)
{
- return new SearchCriterion()
+ SequenceFeature sf = list.get(pos);
+ if (sf.begin > begin)
{
-
- @Override
- boolean compare(SequenceFeature entry)
- {
- return entry.getEnd() >= target;
- }
- };
+ return -1; // no match found
+ }
+ if (sf.equals(feature))
+ {
+ return pos;
+ }
+ pos++;
}
+ return -1;
+ }
- /**
- * serves a search condition for finding the first feature which follows the
- * given range as determined by a supplied comparator
- *
- * @param target
- * @return
- */
- static SearchCriterion byFeature(final ContiguousI to,
- final Comparator<ContiguousI> rc)
+ /**
+ * A helper method to return the maximum of two floats, where a non-NaN value
+ * is treated as 'greater than' a NaN value (unlike Math.max which does the
+ * opposite)
+ *
+ * @param f1
+ * @param f2
+ */
+ protected static float max(float f1, float f2)
+ {
+ if (Float.isNaN(f1))
{
- return new SearchCriterion()
- {
+ return Float.isNaN(f2) ? f1 : f2;
+ }
+ else
+ {
+ return Float.isNaN(f2) ? f1 : Math.max(f1, f2);
+ }
+ }
- @Override
- boolean compare(SequenceFeature entry)
- {
- return rc.compare(entry, to) >= 0;
- }
- };
+ /**
+ * A helper method to return the minimum of two floats, where a non-NaN value
+ * is treated as 'less than' a NaN value (unlike Math.min which does the
+ * opposite)
+ *
+ * @param f1
+ * @param f2
+ */
+ protected static float min(float f1, float f2)
+ {
+ if (Float.isNaN(f1))
+ {
+ return Float.isNaN(f2) ? f1 : f2;
+ }
+ else
+ {
+ return Float.isNaN(f2) ? f1 : Math.min(f1, f2);
}
}
List<SequenceFeature> nonPositionalFeatures;
/*
- * An ordered list of features, with the promise that no feature in the list
- * properly contains any other. This constraint allows bounded linear search
- * of the list for features overlapping a region.
- * Contact features are not included in this list.
- */
- List<SequenceFeature> nonNestedFeatures;
-
- /*
* contact features ordered by first contact position
*/
List<SequenceFeature> contactFeatureStarts;
List<SequenceFeature> contactFeatureEnds;
/*
- * Nested Containment List is used to hold any features that are nested
- * within (properly contained by) any other feature. This is a recursive tree
- * which supports depth-first scan for features overlapping a range.
- * It is used here as a 'catch-all' fallback for features that cannot be put
- * into a simple ordered list without invalidating the search methods.
+ * IntervalStore holds remaining features and provides efficient
+ * query for features overlapping any given interval
*/
- NCList<SequenceFeature> nestedFeatures;
+ Collection<SequenceFeature> features;
/*
* Feature groups represented in stored positional features
*/
public FeatureStore()
{
- nonNestedFeatures = new ArrayList<SequenceFeature>();
- positionalFeatureGroups = new HashSet<String>();
- nonPositionalFeatureGroups = new HashSet<String>();
+ positionalFeatureGroups = new HashSet<>();
+ nonPositionalFeatureGroups = new HashSet<>();
positionalMinScore = Float.NaN;
positionalMaxScore = Float.NaN;
nonPositionalMinScore = Float.NaN;
nonPositionalMaxScore = Float.NaN;
- // we only construct nonPositionalFeatures, contactFeatures
- // or the NCList if we need to
+ // we only construct nonPositionalFeatures, contactFeatures if we need to
}
/**
- * 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.
+ * Add a contact feature to the lists that hold them ordered by start (first
+ * contact) and by end (second contact) position, ensuring the lists remain
+ * ordered, and returns true. This method allows duplicate features to be
+ * added, so test before calling to avoid this.
*
* @param feature
+ * @return
*/
- public boolean addFeature(SequenceFeature feature)
+ protected synchronized boolean addContactFeature(SequenceFeature feature)
{
- if (contains(feature))
+ if (contactFeatureStarts == null)
{
- return false;
+ contactFeatureStarts = new ArrayList<>();
+ contactFeatureEnds = new ArrayList<>();
}
/*
- * keep a record of feature groups
+ * insert into list sorted by start (first contact position):
+ * binary search the sorted list to find the insertion point
*/
- if (!feature.isNonPositional())
- {
- positionalFeatureGroups.add(feature.getFeatureGroup());
- }
+ contactFeatureStarts.add(
+ findFirstBegin(contactFeatureStarts, feature.begin), feature);
+ /*
+ * insert into list sorted by end (second contact position):
+ * binary search the sorted list to find the insertion point
+ */
+ contactFeatureEnds.add(findFirstEnd(contactFeatureEnds, feature.end),
+ feature);
+
+ return true;
+ }
+
+ /**
+ * 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 feature
+ */
- boolean added = false;
+ @Override
+ public boolean addFeature(SequenceFeature feature)
+ {
+ // if (contains(feature))
+ // {
+ // return false;
+ // }
+
+ // /*
+ // * keep a record of feature groups
+ // */
+ // if (!feature.isNonPositional())
+ // {
+ // positionalFeatureGroups.add(feature.getFeatureGroup());
+ // }
if (feature.isContactFeature())
{
- added = addContactFeature(feature);
+ if (containsContactFeature(feature))
+ {
+ return false;
+ }
+ positionalFeatureGroups.add(feature.getFeatureGroup());
+ if (feature.begin > lastContactStart)
+ {
+ lastContactStart = feature.begin;
+ }
+ addContactFeature(feature);
}
else if (feature.isNonPositional())
{
- added = addNonPositionalFeature(feature);
+ if (containsNonPositional(feature))
+ {
+ return false;
+ }
+
+ addNonPositionalFeature(feature);
}
else
{
- added = addNonNestedFeature(feature);
- if (!added)
+ // allow for check with
+ if (checkContainsPositionalFeatureForAdd(feature)
+ || !addPositionalFeature(feature))
{
- /*
- * detected a nested feature - put it in the NCList structure
- */
- added = addNestedFeature(feature);
+ return false;
+ }
+ positionalFeatureGroups.add(feature.getFeatureGroup());
+ // addPositionalFeature(feature);
+ if (feature.begin > lastStart)
+ {
+ lastStart = feature.begin;
}
}
- if (added)
+ /*
+ * record the total extent of positional features, to make
+ * getTotalFeatureLength possible; we count the length of a
+ * contact feature as 1
+ */
+ totalExtent += getFeatureLength(feature);
+
+ /*
+ * record the minimum and maximum score for positional
+ * and non-positional features
+ */
+ float score = feature.getScore();
+ if (!Float.isNaN(score))
{
- /*
- * record the total extent of positional features, to make
- * getTotalFeatureLength possible; we count the length of a
- * contact feature as 1
- */
- totalExtent += getFeatureLength(feature);
-
- /*
- * record the minimum and maximum score for positional
- * and non-positional features
- */
- float score = feature.getScore();
- if (!Float.isNaN(score))
+ if (feature.isNonPositional())
{
- if (feature.isNonPositional())
- {
- nonPositionalMinScore = min(nonPositionalMinScore, score);
- nonPositionalMaxScore = max(nonPositionalMaxScore, score);
- }
- else
- {
- positionalMinScore = min(positionalMinScore, score);
- positionalMaxScore = max(positionalMaxScore, score);
- }
+ nonPositionalMinScore = min(nonPositionalMinScore, score);
+ nonPositionalMaxScore = max(nonPositionalMaxScore, score);
+ }
+ else
+ {
+ positionalMinScore = min(positionalMinScore, score);
+ positionalMaxScore = max(positionalMaxScore, score);
}
}
- return added;
+ return true;
}
- /**
- * Answers true if this store contains the given feature (testing by
- * SequenceFeature.equals), else false
- *
- * @param feature
- * @return
- */
- public boolean contains(SequenceFeature feature)
+ private void addFeaturesForGroup(String group,
+ Collection<SequenceFeature> sfs, List<SequenceFeature> result)
{
- if (feature.isNonPositional())
- {
- return nonPositionalFeatures == null ? false : nonPositionalFeatures
- .contains(feature);
- }
-
- if (feature.isContactFeature())
+ if (sfs == null)
{
- return contactFeatureStarts == null ? false : listContains(
- contactFeatureStarts, feature);
+ return;
}
-
- if (listContains(nonNestedFeatures, feature))
+ for (SequenceFeature sf : sfs)
{
- return true;
+ String featureGroup = sf.getFeatureGroup();
+ if (group == null && featureGroup == null
+ || group != null && group.equals(featureGroup))
+ {
+ result.add(sf);
+ }
}
-
- return nestedFeatures == null ? false : nestedFeatures
- .contains(feature);
}
/**
- * Answers the 'length' of the feature, counting 0 for non-positional features
- * and 1 for contact features
+ * Adds one feature to the IntervalStore that can manage nested features
+ * (creating the IntervalStore if necessary)
*
- * @param feature
- * @return
+ * @return true if added -- allowing for late checking during addition
*/
- protected static int getFeatureLength(SequenceFeature feature)
- {
- if (feature.isNonPositional())
- {
- return 0;
- }
- if (feature.isContactFeature())
- {
- return 1;
- }
- return 1 + feature.getEnd() - feature.getBegin();
- }
+ abstract protected boolean addPositionalFeature(SequenceFeature feature);
/**
* Adds the feature to the list of non-positional features (with lazy
{
if (nonPositionalFeatures == null)
{
- nonPositionalFeatures = new ArrayList<SequenceFeature>();
+ nonPositionalFeatures = new ArrayList<>();
}
nonPositionalFeatures.add(feature);
}
/**
- * Adds one feature to the NCList that can manage nested features (creating
- * the NCList if necessary), and returns true. If the feature is already
- * stored in the NCList (by equality test), then it is not added, and this
- * method returns false.
- */
- protected synchronized boolean addNestedFeature(SequenceFeature feature)
- {
- if (nestedFeatures == null)
- {
- nestedFeatures = new NCList<>(feature);
- return true;
- }
- return nestedFeatures.add(feature, false);
- }
-
- /**
- * Add a feature to the list of non-nested features, maintaining the ordering
- * of the list. A check is made for whether the feature is nested in (properly
- * contained by) an existing feature. If there is no nesting, the feature is
- * added to the list and the method returns true. If nesting is found, the
- * feature is not added and the method returns false.
+ * Answers true if this store contains the given feature (testing by
+ * SequenceFeature.equals), else false
*
* @param feature
* @return
*/
- protected boolean addNonNestedFeature(SequenceFeature feature)
+ @Override
+ public boolean contains(SequenceFeature feature)
{
- synchronized (nonNestedFeatures)
+ if (feature.isNonPositional())
{
- /*
- * find the first stored feature which doesn't precede the new one
- */
- int insertPosition = binarySearch(nonNestedFeatures,
- SearchCriterion.byFeature(feature, RangeComparator.BY_START_POSITION));
-
- /*
- * fail if we detect feature enclosure - of the new feature by
- * the one preceding it, or of the next feature by the new one
- */
- if (insertPosition > 0)
- {
- if (encloses(nonNestedFeatures.get(insertPosition - 1), feature))
- {
- return false;
- }
- }
- if (insertPosition < nonNestedFeatures.size())
- {
- if (encloses(feature, nonNestedFeatures.get(insertPosition)))
- {
- return false;
- }
- }
+ return containsNonPositional(feature);
- /*
- * checks passed - add the feature
- */
- nonNestedFeatures.add(insertPosition, feature);
-
- return true;
}
- }
- /**
- * Answers true if range1 properly encloses range2, else false
- *
- * @param range1
- * @param range2
- * @return
- */
- protected boolean encloses(ContiguousI range1, ContiguousI range2)
- {
- int begin1 = range1.getBegin();
- int begin2 = range2.getBegin();
- int end1 = range1.getEnd();
- int end2 = range2.getEnd();
- if (begin1 == begin2 && end1 > end2)
- {
- return true;
- }
- if (begin1 < begin2 && end1 >= end2)
+ if (feature.isContactFeature())
{
- return true;
+ return containsContactFeature(feature);
+
}
- return false;
+
+ return containsPositionalFeature(feature);
+
}
/**
- * Add a contact feature to the lists that hold them ordered by start (first
- * contact) and by end (second contact) position, ensuring the lists remain
- * ordered, and returns true. This method allows duplicate features to be
- * added, so test before calling to avoid this.
+ * A check that can be overridden if the check is being done during the add
+ * operation itself.
*
* @param feature
* @return
*/
- protected synchronized boolean addContactFeature(SequenceFeature feature)
+ protected boolean checkContainsPositionalFeatureForAdd(
+ SequenceFeature feature)
{
- if (contactFeatureStarts == null)
- {
- contactFeatureStarts = new ArrayList<SequenceFeature>();
- }
- if (contactFeatureEnds == null)
- {
- contactFeatureEnds = new ArrayList<SequenceFeature>();
- }
+ return containsPositionalFeature(feature);
+ }
- /*
- * binary search the sorted list to find the insertion point
- */
- int insertPosition = binarySearch(contactFeatureStarts,
- SearchCriterion.byFeature(feature,
- RangeComparator.BY_START_POSITION));
- contactFeatureStarts.add(insertPosition, feature);
- // and resort to mak siccar...just in case insertion point not quite right
- Collections.sort(contactFeatureStarts, RangeComparator.BY_START_POSITION);
-
- insertPosition = binarySearch(contactFeatureStarts,
- SearchCriterion.byFeature(feature,
- RangeComparator.BY_END_POSITION));
- contactFeatureEnds.add(feature);
- Collections.sort(contactFeatureEnds, RangeComparator.BY_END_POSITION);
+ private boolean containsPositionalFeature(SequenceFeature feature)
+ {
+ return features == null || feature.begin > lastStart ? false
+ : containsFeature(feature);
+ }
- return true;
+ private boolean containsContactFeature(SequenceFeature feature)
+ {
+ return contactFeatureStarts != null && feature.begin <= lastContactStart
+ && listContains(contactFeatureStarts, feature);
}
+ private boolean containsNonPositional(SequenceFeature feature)
+ {
+ return nonPositionalFeatures == null ? false
+ : nonPositionalFeatures.contains(feature);
+ }
+
+ abstract protected boolean containsFeature(SequenceFeature feature);
+
/**
- * Answers true if the list contains the feature, else false. This method is
- * optimised for the condition that the list is sorted on feature start
- * position ascending, and will give unreliable results if this does not hold.
+ * 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 features
- * @param feature
- * @return
+ * @param sf
*/
- protected static boolean listContains(List<SequenceFeature> features,
- SequenceFeature feature)
+
+ @Override
+ public synchronized boolean delete(SequenceFeature sf)
{
- if (features == null || feature == null)
- {
- return false;
- }
+ boolean removed = false;
/*
- * locate the first entry in the list which does not precede the feature
+ * try contact positions (and if found, delete
+ * from both lists of contact positions)
*/
- int pos = binarySearch(features,
- SearchCriterion.byFeature(feature, RangeComparator.BY_START_POSITION));
- int len = features.size();
- while (pos < len)
+ if (!removed && contactFeatureStarts != null)
{
- SequenceFeature sf = features.get(pos);
- if (sf.getBegin() > feature.getBegin())
- {
- return false; // no match found
- }
- if (sf.equals(feature))
+ removed = contactFeatureStarts.remove(sf);
+ if (removed)
{
- return true;
+ contactFeatureEnds.remove(sf);
}
- pos++;
}
- return false;
- }
- /**
- * Returns a (possibly empty) list of features whose extent overlaps the given
- * range. The returned list is not ordered. Contact features are included if
- * either of the contact points lies within the range.
- *
- * @param start
- * start position of overlap range (inclusive)
- * @param end
- * end position of overlap range (inclusive)
- * @return
- */
- public List<SequenceFeature> findOverlappingFeatures(long start, long end)
- {
- List<SequenceFeature> result = new ArrayList<>();
-
- findNonNestedFeatures(start, end, result);
-
- findContactFeatures(start, end, result);
-
- if (nestedFeatures != null)
+ /*
+ * if not found, try non-positional features
+ */
+ if (!removed && nonPositionalFeatures != null)
{
- result.addAll(nestedFeatures.findOverlaps(start, end));
+ removed = nonPositionalFeatures.remove(sf);
}
- return result;
- }
-
- /**
- * Adds contact features to the result list where either the second or the
- * first contact position lies within the target range
- *
- * @param from
- * @param to
- * @param result
- */
- protected void findContactFeatures(long from, long to,
- List<SequenceFeature> result)
- {
- if (contactFeatureStarts != null)
+ /*
+ * if not found, try nested features
+ */
+ if (!removed && features != null)
{
- findContactStartFeatures(from, to, result);
+ removed = findAndRemoveNonContactFeature(sf);
}
- if (contactFeatureEnds != null)
+
+ if (removed)
{
- findContactEndFeatures(from, to, result);
+ rescanAfterDelete();
}
+
+ return removed;
}
- /**
- * Adds to the result list any contact features whose end (second contact
- * point), but not start (first contact point), lies in the query from-to
- * range
- *
- * @param from
- * @param to
- * @param result
- */
- protected void findContactEndFeatures(long from, long to,
- List<SequenceFeature> result)
- {
- /*
- * find the first contact feature (if any) that does not lie
- * entirely before the target range
- */
- int startPosition = binarySearch(contactFeatureEnds,
- SearchCriterion.byEnd(from));
- for (; startPosition < contactFeatureEnds.size(); startPosition++)
- {
- SequenceFeature sf = contactFeatureEnds.get(startPosition);
- if (!sf.isContactFeature())
- {
- System.err.println("Error! non-contact feature type "
- + sf.getType() + " in contact features list");
- continue;
- }
+ abstract protected boolean findAndRemoveNonContactFeature(SequenceFeature sf);
- int begin = sf.getBegin();
- if (begin >= from && begin <= to)
- {
- /*
- * this feature's first contact position lies in the search range
- * so we don't include it in results a second time
- */
- continue;
- }
+ abstract protected void findContactFeatures(long from, long to,
+ List<SequenceFeature> result);
- int end = sf.getEnd();
- if (end >= from && end <= to)
- {
- result.add(sf);
- }
- if (end > to)
- {
- break;
- }
- }
- }
+ abstract protected int findFirstBegin(List<SequenceFeature> list,
+ long pos);
- /**
- * Adds non-nested features to the result list that lie within the target
- * range. Non-positional features (start=end=0), contact features and nested
- * features are excluded.
- *
- * @param from
- * @param to
- * @param result
- */
- protected void findNonNestedFeatures(long from, long to,
- List<SequenceFeature> result)
- {
- /*
- * find the first feature whose end position is
- * after the target range start
- */
- int startIndex = binarySearch(nonNestedFeatures,
- SearchCriterion.byEnd(from));
+ abstract protected int findFirstEnd(List<SequenceFeature> list, long pos);
- final int startIndex1 = startIndex;
- int i = startIndex1;
- while (i < nonNestedFeatures.size())
- {
- SequenceFeature sf = nonNestedFeatures.get(i);
- if (sf.getBegin() > to)
- {
- break;
- }
- if (sf.getBegin() <= to && sf.getEnd() >= from)
- {
- result.add(sf);
- }
- i++;
- }
+ @Override
+ public List<SequenceFeature> findOverlappingFeatures(long start, long end)
+ {
+ return findOverlappingFeatures(start, end, null);
}
- /**
- * Adds contact features whose start position lies in the from-to range to the
- * result list
- *
- * @param from
- * @param to
- * @param result
- */
- protected void findContactStartFeatures(long from, long to,
- List<SequenceFeature> result)
+ @Override
+ public List<SequenceFeature> getContactFeatures()
{
- int startPosition = binarySearch(contactFeatureStarts,
- SearchCriterion.byStart(from));
-
- for (; startPosition < contactFeatureStarts.size(); startPosition++)
- {
- SequenceFeature sf = contactFeatureStarts.get(startPosition);
- if (!sf.isContactFeature())
- {
- System.err.println("Error! non-contact feature type "
- + sf.getType() + " in contact features list");
- continue;
- }
- int begin = sf.getBegin();
- if (begin >= from && begin <= to)
- {
- result.add(sf);
- }
- }
+ return getContactFeatures(new ArrayList<>());
}
/**
- * Answers a list of all positional features stored, in no guaranteed order
+ * Answers a list of all contact features. If there are none, returns an
+ * immutable empty list.
*
* @return
*/
- public List<SequenceFeature> getPositionalFeatures()
- {
- /*
- * add non-nested features (may be all features for many cases)
- */
- List<SequenceFeature> result = new ArrayList<>();
- result.addAll(nonNestedFeatures);
- /*
- * add any contact features - from the list by start position
- */
+ @Override
+ public List<SequenceFeature> getContactFeatures(
+ List<SequenceFeature> result)
+ {
if (contactFeatureStarts != null)
{
result.addAll(contactFeatureStarts);
}
-
- /*
- * add any nested features
- */
- if (nestedFeatures != null)
- {
- result.addAll(nestedFeatures.getEntries());
- }
-
return result;
}
/**
- * Answers a list of all contact features. If there are none, returns an
- * immutable empty list.
+ * Answers the number of positional (or non-positional) features stored.
+ * Contact features count as 1.
*
+ * @param positional
* @return
*/
- public List<SequenceFeature> getContactFeatures()
+
+ @Override
+ public int getFeatureCount(boolean positional)
{
- if (contactFeatureStarts == null)
+ if (!positional)
{
- return Collections.emptyList();
+ return nonPositionalFeatures == null ? 0
+ : nonPositionalFeatures.size();
}
- return new ArrayList<>(contactFeatureStarts);
+
+ return (contactFeatureStarts == null ? 0 : contactFeatureStarts.size())
+ + features.size();
+
}
/**
- * Answers a list of all non-positional features. If there are none, returns
- * an immutable empty list.
+ * Answers the set of distinct feature groups stored, possibly including null,
+ * as an unmodifiable view of the set. The parameter determines whether the
+ * groups for positional or for non-positional features are returned.
*
+ * @param positionalFeatures
* @return
*/
- public List<SequenceFeature> getNonPositionalFeatures()
- {
- if (nonPositionalFeatures == null)
- {
- return Collections.emptyList();
- }
- return new ArrayList<>(nonPositionalFeatures);
- }
- /**
- * 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
- */
- public synchronized boolean delete(SequenceFeature sf)
+ @Override
+ public Set<String> getFeatureGroups(boolean positionalFeatures)
{
- /*
- * try the non-nested positional features first
- */
- boolean removed = nonNestedFeatures.remove(sf);
-
- /*
- * if not found, try contact positions (and if found, delete
- * from both lists of contact positions)
- */
- if (!removed && contactFeatureStarts != null)
- {
- removed = contactFeatureStarts.remove(sf);
- if (removed)
- {
- contactFeatureEnds.remove(sf);
- }
- }
-
- boolean removedNonPositional = false;
-
- /*
- * if not found, try non-positional features
- */
- if (!removed && nonPositionalFeatures != null)
- {
- removedNonPositional = nonPositionalFeatures.remove(sf);
- removed = removedNonPositional;
- }
-
- /*
- * if not found, try nested features
- */
- if (!removed && nestedFeatures != null)
+ if (positionalFeatures)
{
- removed = nestedFeatures.delete(sf);
+ return Collections.unmodifiableSet(positionalFeatureGroups);
}
-
- if (removed)
+ else
{
- rescanAfterDelete();
+ return nonPositionalFeatureGroups == null
+ ? Collections.<String> emptySet()
+ : Collections.unmodifiableSet(nonPositionalFeatureGroups);
}
+ }
- return removed;
+ @Override
+ public Collection<SequenceFeature> getFeatures()
+ {
+ return features;
}
/**
- * Rescan all features to recompute any cached values after an entry has been
- * deleted. This is expected to be an infrequent event, so performance here is
- * not critical.
+ * Answers a list of all either positional or non-positional features whose
+ * feature group matches the given group (which may be null)
+ *
+ * @param positional
+ * @param group
+ * @return
*/
- protected synchronized void rescanAfterDelete()
- {
- positionalFeatureGroups.clear();
- nonPositionalFeatureGroups.clear();
- totalExtent = 0;
- positionalMinScore = Float.NaN;
- positionalMaxScore = Float.NaN;
- nonPositionalMinScore = Float.NaN;
- nonPositionalMaxScore = Float.NaN;
- /*
- * scan non-positional features for groups and scores
- */
- for (SequenceFeature sf : getNonPositionalFeatures())
- {
- nonPositionalFeatureGroups.add(sf.getFeatureGroup());
- float score = sf.getScore();
- nonPositionalMinScore = min(nonPositionalMinScore, score);
- nonPositionalMaxScore = max(nonPositionalMaxScore, score);
- }
+ @Override
+ public List<SequenceFeature> getFeaturesForGroup(boolean positional,
+ String group)
+ {
+ List<SequenceFeature> result = new ArrayList<>();
/*
- * scan positional features for groups, scores and extents
+ * if we know features don't include the target group, no need
+ * to inspect them for matches
*/
- for (SequenceFeature sf : getPositionalFeatures())
+ if (positional && !positionalFeatureGroups.contains(group)
+ || !positional && !nonPositionalFeatureGroups.contains(group))
{
- positionalFeatureGroups.add(sf.getFeatureGroup());
- float score = sf.getScore();
- positionalMinScore = min(positionalMinScore, score);
- positionalMaxScore = max(positionalMaxScore, score);
- totalExtent += getFeatureLength(sf);
+ return result;
}
- }
- /**
- * A helper method to return the minimum of two floats, where a non-NaN value
- * is treated as 'less than' a NaN value (unlike Math.min which does the
- * opposite)
- *
- * @param f1
- * @param f2
- */
- protected static float min(float f1, float f2)
- {
- if (Float.isNaN(f1))
+ if (positional)
{
- return Float.isNaN(f2) ? f1 : f2;
+ addFeaturesForGroup(group, contactFeatureStarts, result);
+ addFeaturesForGroup(group, features, result);
}
else
{
- return Float.isNaN(f2) ? f1 : Math.min(f1, f2);
+ addFeaturesForGroup(group, nonPositionalFeatures, result);
}
+ return result;
}
/**
- * A helper method to return the maximum of two floats, where a non-NaN value
- * is treated as 'greater than' a NaN value (unlike Math.max which does the
- * opposite)
+ * Answers the maximum score held for positional or non-positional features.
+ * This may be Float.NaN if there are no features, are none has a non-NaN
+ * score.
*
- * @param f1
- * @param f2
+ * @param positional
+ * @return
*/
- protected static float max(float f1, float f2)
+
+ @Override
+ public float getMaximumScore(boolean positional)
{
- if (Float.isNaN(f1))
- {
- return Float.isNaN(f2) ? f1 : f2;
- }
- else
- {
- return Float.isNaN(f2) ? f1 : Math.max(f1, f2);
- }
+ return positional ? positionalMaxScore : nonPositionalMaxScore;
}
/**
- * Answers true if this store has no features, else false
+ * Answers the minimum score held for positional or non-positional features.
+ * This may be Float.NaN if there are no features, are none has a non-NaN
+ * score.
*
+ * @param positional
* @return
*/
- public boolean isEmpty()
- {
- boolean hasFeatures = !nonNestedFeatures.isEmpty()
- || (contactFeatureStarts != null && !contactFeatureStarts
- .isEmpty())
- || (nonPositionalFeatures != null && !nonPositionalFeatures
- .isEmpty())
- || (nestedFeatures != null && nestedFeatures.size() > 0);
- return !hasFeatures;
+ @Override
+ public float getMinimumScore(boolean positional)
+ {
+ return positional ? positionalMinScore : nonPositionalMinScore;
}
- /**
- * Answers the set of distinct feature groups stored, possibly including null,
- * as an unmodifiable view of the set. The parameter determines whether the
- * groups for positional or for non-positional features are returned.
- *
- * @param positionalFeatures
- * @return
- */
- public Set<String> getFeatureGroups(boolean positionalFeatures)
+ @Override
+ public List<SequenceFeature> getNonPositionalFeatures()
{
- if (positionalFeatures)
- {
- return Collections.unmodifiableSet(positionalFeatureGroups);
- }
- else
- {
- return nonPositionalFeatureGroups == null ? Collections
- .<String> emptySet() : Collections
- .unmodifiableSet(nonPositionalFeatureGroups);
- }
+ return getNonPositionalFeatures(new ArrayList<>());
}
/**
- * Performs a binary search of the (sorted) list to find the index of the
- * first entry which returns true for the given comparator function. Returns
- * the length of the list if there is no such entry.
+ * Answers a list of all non-positional features. If there are none, returns
+ * an immutable empty list.
*
- * @param features
- * @param sc
* @return
*/
- protected static int binarySearch(List<SequenceFeature> features,
- SearchCriterion sc)
- {
- int start = 0;
- int end = features.size() - 1;
- int matched = features.size();
- while (start <= end)
+ @Override
+ public List<SequenceFeature> getNonPositionalFeatures(
+ List<SequenceFeature> result)
+ {
+ if (nonPositionalFeatures != null)
{
- int mid = (start + end) / 2;
- SequenceFeature entry = features.get(mid);
- boolean compare = sc.compare(entry);
- if (compare)
- {
- matched = mid;
- end = mid - 1;
- }
- else
- {
- start = mid + 1;
- }
+ result.addAll(nonPositionalFeatures);
}
+ return result;
+ }
- return matched;
+ @Override
+ public List<SequenceFeature> getPositionalFeatures()
+ {
+ return getPositionalFeatures(new ArrayList<>());
}
/**
- * Answers the number of positional (or non-positional) features stored.
- * Contact features count as 1.
+ * Answers a list of all positional features stored, in no guaranteed order
*
- * @param positional
* @return
*/
- public int getFeatureCount(boolean positional)
- {
- if (!positional)
- {
- return nonPositionalFeatures == null ? 0 : nonPositionalFeatures
- .size();
- }
- int size = nonNestedFeatures.size();
+ @Override
+ public List<SequenceFeature> getPositionalFeatures(
+ List<SequenceFeature> result)
+ {
+ /*
+ * add any contact features - from the list by start position
+ */
if (contactFeatureStarts != null)
{
- // note a contact feature (start/end) counts as one
- size += contactFeatureStarts.size();
+ result.addAll(contactFeatureStarts);
}
- if (nestedFeatures != null)
+ /*
+ * add any nested features
+ */
+ if (features != null)
{
- size += nestedFeatures.size();
+ result.addAll(features);
}
- return size;
+ return result;
}
/**
*
* @return
*/
+
+ @Override
public int getTotalFeatureLength()
{
return totalExtent;
}
/**
- * Answers the minimum score held for positional or non-positional features.
- * This may be Float.NaN if there are no features, are none has a non-NaN
- * score.
+ * Answers true if this store has no features, else false
*
- * @param positional
* @return
*/
- public float getMinimumScore(boolean positional)
- {
- return positional ? positionalMinScore : nonPositionalMinScore;
- }
- /**
- * Answers the maximum score held for positional or non-positional features.
- * This may be Float.NaN if there are no features, are none has a non-NaN
- * score.
- *
- * @param positional
- * @return
- */
- public float getMaximumScore(boolean positional)
+ @Override
+ public boolean isEmpty()
{
- return positional ? positionalMaxScore : nonPositionalMaxScore;
+ boolean hasFeatures = (contactFeatureStarts != null
+ && !contactFeatureStarts.isEmpty())
+ || (nonPositionalFeatures != null
+ && !nonPositionalFeatures.isEmpty())
+ || features.size() > 0;
+
+ return !hasFeatures;
}
/**
- * Answers a list of all either positional or non-positional features whose
- * feature group matches the given group (which may be null)
- *
- * @param positional
- * @param group
- * @return
+ * Rescan all features to recompute any cached values after an entry has been
+ * deleted. This is expected to be an infrequent event, so performance here is
+ * not critical.
*/
- public List<SequenceFeature> getFeaturesForGroup(boolean positional,
- String group)
+ protected synchronized void rescanAfterDelete()
{
- List<SequenceFeature> result = new ArrayList<>();
-
+ positionalFeatureGroups.clear();
+ nonPositionalFeatureGroups.clear();
+ totalExtent = 0;
+ positionalMinScore = Float.NaN;
+ positionalMaxScore = Float.NaN;
+ nonPositionalMinScore = Float.NaN;
+ nonPositionalMaxScore = Float.NaN;
/*
- * if we know features don't include the target group, no need
- * to inspect them for matches
+ * scan non-positional features for groups and scores
*/
- if (positional && !positionalFeatureGroups.contains(group)
- || !positional && !nonPositionalFeatureGroups.contains(group))
+ if (nonPositionalFeatures != null)
{
- return result;
+ List<SequenceFeature> list = nonPositionalFeatures;
+ for (int i = 0, n = list.size(); i < n; i++)
+ {
+ SequenceFeature sf = list.get(i);
+ nonPositionalFeatureGroups.add(sf.getFeatureGroup());
+ float score = sf.getScore();
+ nonPositionalMinScore = min(nonPositionalMinScore, score);
+ nonPositionalMaxScore = max(nonPositionalMaxScore, score);
+ }
}
- List<SequenceFeature> sfs = positional ? getPositionalFeatures()
- : getNonPositionalFeatures();
+ /*
+ * scan positional features for groups, scores and extents
+ */
+
+ rescanPositional(contactFeatureStarts);
+ rescanPositional(features);
+ }
+
+ private void rescanPositional(Collection<SequenceFeature> sfs)
+ {
+ if (sfs == null)
+ {
+ return;
+ }
for (SequenceFeature sf : sfs)
{
- String featureGroup = sf.getFeatureGroup();
- if (group == null && featureGroup == null || group != null
- && group.equals(featureGroup))
- {
- result.add(sf);
- }
+ positionalFeatureGroups.add(sf.getFeatureGroup());
+ float score = sf.getScore();
+ positionalMinScore = min(positionalMinScore, score);
+ positionalMaxScore = max(positionalMaxScore, score);
+ totalExtent += getFeatureLength(sf);
}
- return result;
}
/**
- * Adds the shift value to the start and end of all positional features.
- * Returns true if at least one feature was updated, else false.
+ * Adds the shift amount to the start and end of all positional features whose
+ * start position is at or after fromPosition. Returns true if at least one
+ * feature was shifted, else false.
*
- * @param shift
+ * @param fromPosition
+ * @param shiftBy
* @return
*/
- public synchronized boolean shiftFeatures(int shift)
+
+ @Override
+ public synchronized boolean shiftFeatures(int fromPosition, int shiftBy)
{
/*
* Because begin and end are final fields (to ensure the data store's
* (Although a simple shift of all values would preserve data integrity!)
*/
boolean modified = false;
- for (SequenceFeature sf : getPositionalFeatures())
+ List<SequenceFeature> list = getPositionalFeatures();
+ for (int i = 0, n = list.size(); i < n; i++)
{
- modified = true;
- int newBegin = sf.getBegin() + shift;
- int newEnd = sf.getEnd() + shift;
-
- /*
- * sanity check: don't shift left of the first residue
- */
- if (newEnd > 0)
+ SequenceFeature sf = list.get(i);
+ if (sf.getBegin() >= fromPosition)
{
- newBegin = Math.max(1, newBegin);
- SequenceFeature sf2 = new SequenceFeature(sf, newBegin, newEnd,
- sf.getFeatureGroup(), sf.getScore());
- addFeature(sf2);
+ modified = true;
+ int newBegin = sf.getBegin() + shiftBy;
+ int newEnd = sf.getEnd() + shiftBy;
+
+ /*
+ * sanity check: don't shift left of the first residue
+ */
+ if (newEnd > 0)
+ {
+ newBegin = Math.max(1, newBegin);
+ SequenceFeature sf2 = new SequenceFeature(sf, newBegin, newEnd,
+ sf.getFeatureGroup(), sf.getScore());
+ addFeature(sf2);
+ }
+ delete(sf);
}
- delete(sf);
}
return modified;
}
+
}