/*
* The MIT License (MIT)
*
* Copyright (c) 2007-2015 Broad Institute
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
package org.broad.igv.feature;
import org.apache.log4j.Logger;
import org.broad.igv.util.collections.LRUCache;
import htsjdk.tribble.CloseableTribbleIterator;
import htsjdk.tribble.Feature;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Collections;
import java.util.Iterator;
import java.util.List;
/**
* Class to handle caching data from any source of features
* Subclasses must override queryRaw
*
* @author jrobinso
* @date Jun 24, 2010
*/
public abstract class AbstractCacher {
private static Logger log = Logger.getLogger(AbstractCacher.class);
protected int binSize = Integer.MAX_VALUE;
protected LRUCache<String, Bin> cache;
public AbstractCacher(int binCount, int binSize) {
this.cache = new LRUCache(binCount);
setBinSize(binSize);
}
/**
* Obtain data from underlying source
*
* @param chr
* @param start
* @param end
* @return
*/
protected abstract Iterator<Feature> queryRaw(String chr, int start, int end) throws IOException;
/**
* Set the bin size. This invalidates the cache.
*
* @param newSize
*/
public void setBinSize(int newSize) {
this.binSize = newSize == 0 ? Integer.MAX_VALUE : newSize; // A binSize of zero => use a single bin for the entire chromosome
cache.clear();
}
public void close() throws IOException {
cache.clear();
}
/**
* Query the cached data, refreshing from raw data as necessary
*
* @param chr
* @param start
* @param end
* @return
* @throws IOException
*/
public Iterator<Feature> queryCached(String chr, int start, int end) throws IOException {
int startBin = start / binSize;
int endBin = end / binSize; // <= inclusive
List<Bin> tiles = getBins(chr, startBin, endBin);
if (tiles.size() == 0) {
return Collections.<Feature>emptyList().iterator();
}
// Count total # of records
int recordCount = tiles.get(0).getOverlappingRecords().size();
for (Bin t : tiles) {
recordCount += t.getContainedRecords().size();
}
List<Feature> features = new ArrayList(recordCount);
features.addAll(tiles.get(0).getOverlappingRecords());
for (Bin t : tiles) {
features.addAll(t.getContainedRecords());
}
return new BinIterator(start, end, features);
}
/**
* Return loaded tiles that span the query interval.
* <p/>
* We synchronize this method because different threads might be using
* the same source. Synchronizing here ensures that data
* is loaded as few times as possible (first caller loads it into the cache,
* the second caller accesses it from there) as well as preventing bugs stemming
* from multiple thread access
*
* @param seq
* @param startBin
* @param endBin
* @return
*/
private synchronized List<Bin> getBins(String seq, int startBin, int endBin) {
List<Bin> tiles = new ArrayList(endBin - startBin + 1);
List<Bin> tilesToLoad = new ArrayList(endBin - startBin + 1);
for (int t = startBin; t <= endBin; t++) {
String key = seq + "_" + t;
Bin tile = cache.get(key);
if (tile == null) {
if (log.isDebugEnabled()) {
log.debug("Tile cache miss: " + t);
}
int start = t * binSize;
int end = start + binSize;
tile = new Bin(t, start, end);
cache.put(key, tile);
}
tiles.add(tile);
// The current tile is loaded, load any preceding tiles we have pending
if (tile.isLoaded()) {
if (tilesToLoad.size() > 0) {
if (!loadTiles(seq, tilesToLoad)) {
return tiles;
}
}
tilesToLoad.clear();
} else {
tilesToLoad.add(tile);
}
}
if (tilesToLoad.size() > 0) {
loadTiles(seq, tilesToLoad);
}
return tiles;
}
private boolean loadTiles(String seq, List<Bin> tiles) {
assert (tiles.size() > 0);
if (log.isDebugEnabled()) {
int first = tiles.get(0).getBinNumber();
int end = tiles.get(tiles.size() - 1).getBinNumber();
log.debug("Loading tiles: " + first + "-" + end);
}
// Convert start to 1-based coordinates
int start = tiles.get(0).start + 1;
int end = tiles.get(tiles.size() - 1).end;
Iterator<Feature> iter = null;
//log.debug("Loading : " + start + " - " + end);
int featureCount = 0;
long t0 = System.currentTimeMillis();
try {
iter = queryRaw(seq, start, end);
while (iter != null && iter.hasNext()) {
Feature record = iter.next();
// Range of tile indices that this feature contributes to.
int aStart = record.getStart();
int aEnd = record.getEnd();
int idx0 = 0;
int idx1 = 0;
if (binSize > 0) {
idx0 = Math.max(0, (aStart - start) / binSize);
idx1 = Math.min(tiles.size() - 1, (aEnd - start) / binSize);
}
// Loop over tiles this read overlaps
for (int i = idx0; i <= idx1; i++) {
Bin t = tiles.get(i);
// A bin size == 0 means use a single bin for the entire chromosome. This is a confusing convention.
if (binSize == 0 || ((aStart >= t.start) && (aStart < t.end))) {
t.containedRecords.add(record);
} else if ((aEnd >= t.start) && (aStart < t.start)) {
t.overlappingRecords.add(record);
}
}
}
for (Bin t : tiles) {
t.setLoaded(true);
}
if (log.isDebugEnabled()) {
long dt = System.currentTimeMillis() - t0;
long rate = dt == 0 ? Long.MAX_VALUE : featureCount / dt;
log.debug("Loaded " + featureCount + " reads in " + dt + "ms. (" + rate + " reads/ms)");
}
return true;
} catch (IOException e) {
log.error("IOError loading feature data", e);
// TODO -- do something about this, how do we want to handle this exception?
throw new RuntimeException(e);
} finally {
if (iter != null) {
//iter.close();
}
//IGV.getInstance().resetStatusMessage();
}
}
private static class Bin {
private boolean loaded = false;
private int start;
private int end;
private int binNumber;
private List<Feature> containedRecords;
private List<Feature> overlappingRecords;
Bin(int binNumber, int start, int end) {
this.binNumber = binNumber;
this.start = start;
this.end = end;
containedRecords = new ArrayList(1000);
overlappingRecords = new ArrayList(100);
}
public int getBinNumber() {
return binNumber;
}
public int getStart() {
return start;
}
public void setStart(int start) {
this.start = start;
}
public List<Feature> getContainedRecords() {
return containedRecords;
}
public List<Feature> getOverlappingRecords() {
return overlappingRecords;
}
public boolean isLoaded() {
return loaded;
}
public void setLoaded(boolean loaded) {
this.loaded = loaded;
}
}
/**
*
*/
private class BinIterator implements CloseableTribbleIterator {
Iterator<Feature> currentFeatureIterator;
int end;
Feature nextRecord;
int start;
List<Feature> features;
BinIterator(int start, int end, List<Feature> features) {
this.features = features;
this.start = start;
this.end = end;
currentFeatureIterator = features.iterator();
advanceToFirstRecord();
}
public void close() {
// No-op
}
public boolean hasNext() {
return nextRecord != null;
}
public Feature next() {
Feature ret = nextRecord;
advanceToNextRecord();
return ret;
}
public void remove() {
// ignored
}
private void advanceToFirstRecord() {
advanceToNextRecord();
}
private void advanceToNextRecord() {
advance();
while ((nextRecord != null) && (nextRecord.getEnd() < start)) {
advance();
}
}
private void advance() {
if (currentFeatureIterator.hasNext()) {
nextRecord = currentFeatureIterator.next();
if (nextRecord.getStart() > end) {
nextRecord = null;
}
} else {
nextRecord = null;
}
}
public Iterator iterator() {
return this;
}
}
}