/*
* Copyright (c) 2007-2012 The Broad Institute, Inc.
* SOFTWARE COPYRIGHT NOTICE
* This software and its documentation are the copyright of the Broad Institute, Inc. All rights are reserved.
*
* This software is supplied without any warranty or guaranteed support whatsoever. The Broad Institute is not responsible for its use, misuse, or functionality.
*
* This software is licensed under the terms of the GNU Lesser General Public License (LGPL),
* Version 2.1 which is available at http://www.opensource.org/licenses/lgpl-2.1.php.
*/
package org.broad.igv.cbio;
import biz.source_code.base64Coder.Base64Coder;
import com.google.common.base.Predicate;
import org.broad.igv.AbstractHeadlessTest;
import org.broad.igv.track.Track;
import org.broad.igv.track.TrackLoader;
import org.broad.igv.util.*;
import org.broad.igv.util.collections.CollUtils;
import org.jgrapht.EdgeFactory;
import org.jgrapht.VertexFactory;
import org.jgrapht.generate.WheelGraphGenerator;
import org.junit.After;
import org.junit.Before;
import org.junit.Test;
import org.w3c.dom.Document;
import org.w3c.dom.Element;
import org.w3c.dom.NamedNodeMap;
import org.w3c.dom.Node;
import javax.xml.parsers.DocumentBuilderFactory;
import java.io.*;
import java.util.Collection;
import java.util.HashSet;
import java.util.List;
import java.util.Set;
import java.util.zip.GZIPInputStream;
import static org.junit.Assert.*;
/**
* Test our GeneNetwork class, which implements jgraphT graph interface.
* See http://www.jgrapht.org.
* <p/>
* Notes:
* Modification methods tend to return true for any modification,
* and false for none. So graph.removeAll(nodeSet) would return
* true if only half of the nodes were actually removed, although
* odds are an exception would be thrown for any which could
* not be removed.
* <p/>
* User: jacob
* Date: 2012/02/02
*/
public class GeneNetworkTest extends AbstractHeadlessTest {
private static String testpath = TestUtils.DATA_DIR + "xml/tp53network.xml";
private static String drugTestPath = TestUtils.DATA_DIR + "xml/EGFR_withdrugs.xml.gz";
private GeneNetwork network;
@Before
public void setUp() throws Exception {
network = new GeneNetwork();
}
@After
public void tearDown() throws Exception {
network = null;
GeneNetwork.BASE_URL = GeneNetwork.REAL_URL;
}
public static void main(String[] args) throws IOException {
String netpath = testpath;
if (args != null && args.length > 0) {
netpath = args[0];
}
GeneNetwork network = new GeneNetwork();
network.loadNetwork(netpath);
String viewPath = network.outputForcBioView();
BrowserLauncher.openURL("file://" + viewPath);
}
@Test
public void testLoadLocal() throws Exception {
assertTrue("Failed to load network", network.loadNetwork(testpath) > 0);
}
/**
* Test that we don't filter out drug nodes
*
* @throws Exception
*/
@Test
public void testFilterWithDrugs() throws Exception {
assertTrue("Failed to load network", network.loadNetwork(drugTestPath) > 0);
Collection<Node> nonGenesBeforeFilter = CollUtils.filter(network.vertexSet(), GeneNetwork.isNotGene);
assertTrue("Bad test setup, no non-gene nodes", nonGenesBeforeFilter.size() > 0);
doTestAnnotation(network);
int genesRemoved = network.filterGenesRange(GeneNetwork.PERCENT_MUTATED, 0, 10.0f);
assertTrue("Bad test setup, Filter didn't remove any genes", genesRemoved > 0);
Set<Node> nonGenesAfterFilter =
new HashSet<Node>(CollUtils.filter(network.vertexSet(), GeneNetwork.isNotGene));
for (Node nonGene : nonGenesBeforeFilter) {
String msg = String.format("Filtered out node %s which we shouldn't have", GeneNetwork.getNodeKeyData(nonGene, GeneNetwork.LABEL));
assertTrue(msg, nonGenesAfterFilter.contains(nonGene));
}
}
private static Predicate evenPred = new Predicate() {
public boolean apply(Object object) {
Node node = (Node) object;
NamedNodeMap map = node.getAttributes();
if (map == null) {
return false;
}
int id = Integer.parseInt(map.getNamedItem("id").getTextContent());
return id % 2 == 0;
}
};
@Test
public void testReset() throws Exception {
network.loadNetwork(testpath);
int initSize = network.vertexSet().size();
int nonQuery = network.filterGenes(GeneNetwork.inQuery);
assertTrue("Bad test setup, all genes in query", nonQuery > 0);
assertEquals(initSize - network.vertexSet().size(), nonQuery);
network.reset();
assertEquals("Resetting network did not bring back everything", initSize, network.vertexSet().size());
}
@Test
public void testFilterKeepsQuery() throws Exception {
network.loadNetwork(testpath);
network.filterGenes(GeneNetwork.inQuery);
Collection<Node> queryGenes = network.geneVertexes();
network.reset();
int numRemoved = network.filterGenes(evenPred);
assertTrue(numRemoved > 0);
Collection<Node> randomGenes = network.geneVertexes();
Set<Node> randomGeneSet = new HashSet<Node>(randomGenes);
for (Node queryGene : queryGenes) {
assertTrue(randomGeneSet.contains(queryGene));
}
}
@Test
public void testFilterKeepsEdges() throws Exception {
network.loadNetwork(testpath);
int initSize = network.vertexSet().size();
int numRemoved = network.filterGenes(evenPred);
assertTrue(numRemoved > 0);
//Test that we can get the filtered edges of a node
Set<Node> keptNodes = new HashSet<Node>();
for (Node n : network.geneVertexes()) {
for (Node e : network.edgesOf(n)) {
keptNodes.add(network.getEdgeSource(e));
keptNodes.add(network.getEdgeTarget(e));
}
}
assertEquals(network.geneVertexes().size(), keptNodes.size());
assertTrue("Filtering not performed", keptNodes.size() < initSize);
}
@Test
public void testFilterOnEvidence() throws Exception {
GeneNetwork geneNetwork = new GeneNetwork();
assertTrue(geneNetwork.loadNetwork(TestUtils.DATA_DIR + "xml/egfr_brca1.xml.gz") > 0);
final String badname = "NA";
Predicate<Node> has_evidence = new Predicate<Node>() {
public boolean apply(Node object) {
String label = GeneNetwork.getNodeKeyData(object, "EXPERIMENTAL_TYPE");
return label != null && !label.equals(badname);
}
};
//Perform the filtering.
//This is true if any modifications are made.
assertTrue(geneNetwork.filterEdges(has_evidence) > 0);
for (Node e : geneNetwork.edgeSet()) {
assertTrue(has_evidence.apply(e));
}
}
@Test
public void testOutputNoGzip() throws Exception {
String networkPath = TestUtils.DATA_DIR + "xml/egfr_brca1.xml.gz";
//String networkPath = testpath;
assertTrue(network.loadNetwork(networkPath) > 0);
String outPath = TestUtils.DATA_DIR + "out/test.xml";
tstOutputNetwork(network, outPath);
}
@Test
public void testOutputGzip() throws Exception {
String networkPath = TestUtils.DATA_DIR + "xml/egfr_brca1.xml.gz";
assertTrue(network.loadNetwork(networkPath) > 0);
String outPath = TestUtils.DATA_DIR + "out/test.xml.gz";
tstOutputNetwork(network, outPath);
}
public void tstOutputNetwork(GeneNetwork network, String outPath) throws Exception {
Set<Node> nodes = network.vertexSet();
Set<String> nodeNames = new HashSet<String>();
for (Node node : nodes) {
nodeNames.add(GeneNetwork.getNodeKeyData(node, "label"));
}
assertTrue(network.exportGraph(outPath) > 0);
GeneNetwork at = new GeneNetwork();
assertTrue(at.loadNetwork(outPath) > 0);
//Check that node set matches
Set<Node> outNodes = at.vertexSet();
assertEquals("Output has a different number of nodes than input", nodes.size(), outNodes.size());
for (Node oNode : outNodes) {
String nodeName = GeneNetwork.getNodeKeyData(oNode, "label");
assertTrue(nodeNames.contains(nodeName));
}
}
private void doTestAnnotation(GeneNetwork network) throws Exception {
//Load some tracks
String dataPath = TestUtils.DATA_DIR + "seg/Broad.080528.subtypes.seg.gz";
ResourceLocator locator = new ResourceLocator(dataPath);
List<Track> tracks = new TrackLoader().load(locator, genome);
network.annotateAll(tracks);
}
@Test
public void testAnnotateAll() throws Exception {
String networkPath = TestUtils.DATA_DIR + "xml/egfr_brca1.xml.gz";
assertTrue(network.loadNetwork(networkPath) > 0);
doTestAnnotation(network);
//Check data
Set<Node> nodes = network.vertexSet();
for (Node node : nodes) {
for (String key : GeneNetwork.attributeMap.keySet()) {
String data = GeneNetwork.getNodeKeyData(node, key);
String name = GeneNetwork.getNodeKeyData(node, GeneNetwork.LABEL);
if (!"CHMP3".equalsIgnoreCase(name)) {
assertNotNull(data);
}
}
}
//Check schema
Document doc = network.createDocument();
Node gml = doc.getFirstChild();
for (String key : GeneNetwork.attributeMap.keySet()) {
String data = GeneNetwork.getNodeAttrValue(gml, "id", key);
assertNotNull(data);
}
}
@Test
public void testPruneGraph() throws Exception {
int size = 10;
int exp_edges = size - 1 + size - 1;
Document document = DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument();
SimpleVertexFactory vFactory = new SimpleVertexFactory(document);
SimpleEdgeFactory eFactory = new SimpleEdgeFactory(document);
GeneNetwork graph = new GeneNetwork(eFactory);
//Bicycle wheel. Hub node is connected to all elements, outer elements
//conected in ring.
WheelGraphGenerator<Node, Node> wgg = new WheelGraphGenerator<Node, Node>(size);
wgg.generateGraph(graph, vFactory, null);
assertEquals(size, graph.vertexSet().size());
assertEquals(exp_edges, graph.edgeSet().size());
//Find first spoke vertex, isolate it.
Node toRem = null;
for (Node v : graph.vertexSet()) {
if (graph.edgesOf(v).size() < size && graph.edgesOf(v).size() == 3) {
//Spoke vertex
toRem = v;
break;
}
}
Set<Node> edges = new HashSet<Node>();
for (Node e : graph.edgesOf(toRem)) {
edges.add(e);
}
assertTrue(graph.removeAllEdges(edges));
//At this point, toRem should be isolated
assertEquals(0, graph.edgesOf(toRem).size());
assertTrue(graph.pruneGraph());
assertEquals(size - 1, graph.vertexSet().size());
assertEquals(exp_edges - 3, graph.edgeSet().size());
assertFalse(graph.containsVertex(toRem));
}
@Test
public void testOutputForcBioView() throws Exception {
assertTrue(network.loadNetwork(testpath) > 0);
String outPath = network.outputForcBioView();
//Now attempt to read back in
//Note that this will not work if the output is in plaintext,
//because it contains the XML header in the middle of the document
Document inDoc = Utilities.createDOMDocumentFromXmlStream(new FileInputStream(outPath));
String b64data = inDoc.getElementsByTagName("textarea").item(0).getTextContent().trim();
byte[] gzippedInput = Base64Coder.decode(b64data);
BufferedReader bufIn = null;
try {
InputStream plainData = new GZIPInputStream(new ByteArrayInputStream(gzippedInput));
bufIn = new BufferedReader(new InputStreamReader(plainData));
} catch (IOException e) {
bufIn = new BufferedReader(new InputStreamReader(new ByteArrayInputStream(gzippedInput)));
}
int count = 0;
String[] outLines = Utilities.getString(network.createDocument()).split(FileUtils.LINE_SEPARATOR);
String line;
while ((line = bufIn.readLine()) != null) {
assertEquals(outLines[count], line);
count++;
}
}
/**
* Generates nodes with unique IDs, all with <data key="type">Protein</data>
* attributes
*/
public static class SimpleVertexFactory implements VertexFactory<Node> {
private long vCounter = 0;
private final Document document;
public SimpleVertexFactory(Document document) {
this.document = document;
}
public Node createVertex() {
Element v = document.createElement("v" + vCounter);
v.setAttribute("id", "" + vCounter);
//<data key="type">Protein</data>
Element type = document.createElement("data");
type.setAttribute("key", "type");
type.setTextContent("Protein");
v.appendChild(type);
vCounter++;
return v;
}
}
public static class SimpleEdgeFactory implements EdgeFactory<Node, Node> {
private long eCounter = 0;
private final Document document;
public SimpleEdgeFactory(Document document) {
this.document = document;
}
public Node createEdge(Node sourceVertex, Node targetVertex) {
Element e = document.createElement("e" + eCounter);
e.setAttribute("source", sourceVertex.getAttributes().getNamedItem("id").toString());
e.setAttribute("target", targetVertex.getAttributes().getNamedItem("id").toString());
eCounter++;
return e;
}
}
}