static void checkComponents(CuTest *testCase, stList *filteredEdges) {
stHash *nodesToComponents = getComponents(filteredEdges);
//Check all components are smaller than threshold
stList *components = stHash_getValues(nodesToComponents);
for (int64_t i = 0; i < stList_length(components); i++) {
stSortedSet *component = stList_get(components, i);
CuAssertTrue(testCase, stSortedSet_size(component) <= maxComponentSize);
CuAssertTrue(testCase, stSortedSet_size(component) >= 1);
}
//Check no edges can be added from those filtered.
stSortedSet *filteredEdgesSet = stList_getSortedSet(filteredEdges, (int(*)(const void *, const void *)) stIntTuple_cmpFn);
for (int64_t i = 0; i < stList_length(edges); i++) {
stIntTuple *edge = stList_get(edges, i);
if (stSortedSet_search(filteredEdgesSet, edge) == NULL) {
stIntTuple *node1 = stIntTuple_construct1( stIntTuple_get(edge, 1));
stIntTuple *node2 = stIntTuple_construct1( stIntTuple_get(edge, 2));
stSortedSet *component1 = stHash_search(nodesToComponents, node1);
stSortedSet *component2 = stHash_search(nodesToComponents, node2);
CuAssertTrue(testCase, component1 != NULL && component2 != NULL);
CuAssertTrue(testCase, component1 != component2);
CuAssertTrue(testCase, stSortedSet_size(component1) + stSortedSet_size(component2) > maxComponentSize);
stIntTuple_destruct(node1);
stIntTuple_destruct(node2);
}
}
stSortedSet_destruct(filteredEdgesSet);
//Cleanup the components
stSortedSet *componentsSet = stList_getSortedSet(components, NULL);
stList_destruct(components);
stSortedSet_setDestructor(componentsSet, (void(*)(void *)) stSortedSet_destruct);
stSortedSet_destruct(componentsSet);
stHash_destruct(nodesToComponents);
}
static void test_stSet_search(CuTest* testCase) {
testSetup();
stIntTuple *i = stIntTuple_construct1( 0);
stIntTuple *j = stIntTuple_construct2(10, 0);
stIntTuple *k = stIntTuple_construct1( 5);
//Check search by memory address
CuAssertTrue(testCase, stSet_search(set0, one) == one);
CuAssertTrue(testCase, stSet_search(set0, two) == two);
CuAssertTrue(testCase, stSet_search(set0, three) == three);
CuAssertTrue(testCase, stSet_search(set0, four) == four);
CuAssertTrue(testCase, stSet_search(set0, five) == five);
CuAssertTrue(testCase, stSet_search(set0, six) == six);
//Check not present
CuAssertTrue(testCase, stSet_search(set0, i) == NULL);
CuAssertTrue(testCase, stSet_search(set0, j) == NULL);
CuAssertTrue(testCase, stSet_search(set0, k) == NULL);
//Check search by memory address
CuAssertTrue(testCase, stSet_search(set1, one) == one);
CuAssertTrue(testCase, stSet_search(set1, two) == two);
CuAssertTrue(testCase, stSet_search(set1, three) == three);
CuAssertTrue(testCase, stSet_search(set1, four) == four);
CuAssertTrue(testCase, stSet_search(set1, five) == five);
CuAssertTrue(testCase, stSet_search(set1, six) == six);
//Check not present
CuAssertTrue(testCase, stSet_search(set1, j) == NULL);
//Check is searching by memory
CuAssertTrue(testCase, stSet_search(set1, i) == one);
CuAssertTrue(testCase, stSet_search(set1, k) == six);
stIntTuple_destruct(i);
stIntTuple_destruct(j);
stIntTuple_destruct(k);
testTeardown();
}
static stHash *getComponents(stList *filteredEdges) {
/*
* A kind of stupid reimplementation of the greedy function, done just to trap typos.
*/
stHash *nodesToComponents = stHash_construct3((uint64_t(*)(const void *)) stIntTuple_hashKey,
(int(*)(const void *, const void *)) stIntTuple_equalsFn, NULL, NULL);
for (int64_t i = 0; i < stList_length(nodes); i++) {
stIntTuple *node = stList_get(nodes, i);
stSortedSet *component = stSortedSet_construct();
stSortedSet_insert(component, node);
stHash_insert(nodesToComponents, node, component);
}
for (int64_t i = 0; i < stList_length(filteredEdges); i++) {
stIntTuple *edge = stList_get(filteredEdges, i);
stIntTuple *node1 = stIntTuple_construct1( stIntTuple_get(edge, 1));
stIntTuple *node2 = stIntTuple_construct1( stIntTuple_get(edge, 2));
stSortedSet *component1 = stHash_search(nodesToComponents, node1);
stSortedSet *component2 = stHash_search(nodesToComponents, node2);
assert(component1 != NULL && component2 != NULL);
if (component1 != component2) {
stSortedSet *component3 = stSortedSet_getUnion(component1, component2);
stSortedSetIterator *setIt = stSortedSet_getIterator(component3);
stIntTuple *node3;
while ((node3 = stSortedSet_getNext(setIt)) != NULL) {
stHash_insert(nodesToComponents, node3, component3);
}
stSortedSet_destructIterator(setIt);
stSortedSet_destruct(component1);
stSortedSet_destruct(component2);
}
stIntTuple_destruct(node1);
stIntTuple_destruct(node2);
}
return nodesToComponents;
}
static void checkIsValidReference(CuTest *testCase, stList *reference,
double totalScore) {
stList *chosenEdges = convertReferenceToAdjacencyEdges(reference);
//Check that everyone has a partner.
CuAssertIntEquals(testCase, nodeNumber, stList_length(chosenEdges) * 2);
stSortedSet *nodes = stSortedSet_construct3((int(*)(const void *, const void *)) stIntTuple_cmpFn,
(void(*)(void *)) stIntTuple_destruct);
for (int64_t i = 0; i < nodeNumber; i++) {
stSortedSet_insert(nodes, stIntTuple_construct1( i));
}
checkEdges(chosenEdges, nodes, 1, 0);
//Check that the score is correct
double totalScore2 = calculateZScoreOfReference(reference, nodeNumber, zMatrix);
CuAssertDblEquals(testCase, totalScore2, totalScore, 0.000001);
//Check that the stubs are properly connected.
stList *allEdges = stList_copy(chosenEdges, NULL);
stList_appendAll(allEdges, stubs);
stList_appendAll(allEdges, chains);
stList *components = getComponents(allEdges);
CuAssertIntEquals(testCase, stList_length(stubs), stList_length(reference));
CuAssertIntEquals(testCase, stList_length(stubs), stList_length(components));
//Cleanup
stList_destruct(components);
stSortedSet_destruct(nodes);
stList_destruct(allEdges);
stList_destruct(chosenEdges);
}
static stList *getEdgesThatBridgeComponents(stList *components,
stHash *nodesToNonZeroWeightedAdjacencyEdges) {
/*
* Get set of adjacency edges that bridge between (have a node in two) components.
*/
stList *bridgingAdjacencyEdges = stList_construct();
for (int64_t i = 0; i < stList_length(components); i++) {
stSortedSet *componentNodes = getNodeSetOfEdges(
stList_get(components, i));
stSortedSetIterator *it = stSortedSet_getIterator(componentNodes);
stIntTuple *node;
while ((node = stSortedSet_getNext(it)) != NULL) {
stList *edges = stHash_search(nodesToNonZeroWeightedAdjacencyEdges,
node);
if (edges != NULL) {
for (int64_t j = 0; j < stList_length(edges); j++) {
stIntTuple *edge = stList_get(edges, j);
stIntTuple *node1 = stIntTuple_construct1(
stIntTuple_get(edge, 0));
stIntTuple *node2 = stIntTuple_construct1(
stIntTuple_get(edge, 1));
assert(
stSortedSet_search(componentNodes, node1) != NULL
|| stSortedSet_search(componentNodes, node2)
!= NULL);
if (stSortedSet_search(componentNodes, node1) == NULL
|| stSortedSet_search(componentNodes, node2)
== NULL) {
stList_append(bridgingAdjacencyEdges, edge);
}
stIntTuple_destruct(node1);
stIntTuple_destruct(node2);
}
}
}
stSortedSet_destructIterator(it);
stSortedSet_destruct(componentNodes);
}
return bridgingAdjacencyEdges;
}
static void test_stPosetAlignment_addAndIsPossible(CuTest *testCase) {
for(int64_t trial=0; trial<100; trial++) {
setup();
//Make random number of sequences.
stList *sequenceLengths = stList_construct3(0, (void (*)(void *))stIntTuple_destruct);
for(int64_t i=0; i<sequenceNumber; i++) {
stList_append(sequenceLengths, stIntTuple_construct1( st_randomInt(0, MAX_SEQUENCE_SIZE)));
}
//Propose random alignment pairs...
stList *pairs = stList_construct3(0, (void(*)(void *))stIntTuple_destruct);
int64_t maxAlignedPairs = st_randomInt(0, MAX_ALIGNMENTS);
if(sequenceNumber > 0) {
for(int64_t i=0; i<maxAlignedPairs; i++) {
int64_t seq1 = st_randomInt(0, sequenceNumber);
int64_t seqLength1 = stIntTuple_get(stList_get(sequenceLengths, seq1), 0);
if(seqLength1 == 0) {
continue;
}
int64_t position1 = st_randomInt(0, seqLength1);
int64_t seq2 = st_randomInt(0, sequenceNumber);
int64_t seqLength2 = stIntTuple_get(stList_get(sequenceLengths, seq1), 0);
if(seqLength2 == 0) {
continue;
}
int64_t position2 = st_randomInt(0, seqLength2);
if(seq1 != seq2) {
stList_append(pairs, stIntTuple_construct4( seq1, position1, seq2, position2));
if(stPosetAlignment_isPossible(posetAlignment, seq1, position1, seq2, position2)) {
st_logInfo("In %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " \n", seq1, position1, seq2, position2);
//For each accepted pair check it doesn't create a cycle.
CuAssertTrue(testCase, !containsACycle(pairs, sequenceNumber));
CuAssertTrue(testCase, stPosetAlignment_add(posetAlignment, seq1, position1, seq2, position2));
}
else {
st_logInfo("Out %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " \n", seq1, position1, seq2, position2);
//For each rejected pair check it creates a cycle..
CuAssertTrue(testCase, containsACycle(pairs, sequenceNumber));
CuAssertTrue(testCase, !stPosetAlignment_isPossible(posetAlignment, seq1, position1, seq2, position2));
stIntTuple_destruct(stList_pop(pairs)); //remove the pair which created the cycle.
CuAssertTrue(testCase, !containsACycle(pairs, sequenceNumber)); //Check we're back to being okay..
}
}
}
}
//Cleanup
stList_destruct(sequenceLengths);
stList_destruct(pairs);
teardown();
st_logInfo("Passed a random ordering test with %" PRIi64 " sequences and %" PRIi64 " aligned pairs\n", sequenceNumber, maxAlignedPairs);
}
}
stHash *buildContigPathToContigPathLengthHash(
stList *maximalHaplotypePaths) {
stHash *maximalHaplotypesToMaximalHaplotypePathLengths = stHash_construct();
for (int64_t i = 0; i < stList_length(maximalHaplotypePaths); i++) {
stList *maximalHaplotypePath = stList_get(maximalHaplotypePaths, i);
int64_t k = contigPathLength(maximalHaplotypePath);
stHash_insert(maximalHaplotypesToMaximalHaplotypePathLengths,
maximalHaplotypePath, stIntTuple_construct1( k));
}
return maximalHaplotypesToMaximalHaplotypePathLengths;
}
static void testSetup() {
// compare by value of memory address
set0 = stSet_construct();
// compare by value of ints.
set1 = stSet_construct3((uint64_t(*)(const void *)) stIntTuple_hashKey,
(int(*)(const void *, const void *)) stIntTuple_equalsFn,
(void(*)(void *)) stIntTuple_destruct);
one = stIntTuple_construct1( 0);
two = stIntTuple_construct1( 1);
three = stIntTuple_construct1( 2);
four = stIntTuple_construct1( 3);
five = stIntTuple_construct1( 4);
six = stIntTuple_construct1( 5);
stSet_insert(set0, one);
stSet_insert(set0, two);
stSet_insert(set0, three);
stSet_insert(set0, four);
stSet_insert(set0, five);
stSet_insert(set0, six);
stSet_insert(set1, one);
stSet_insert(set1, two);
stSet_insert(set1, three);
stSet_insert(set1, four);
stSet_insert(set1, five);
stSet_insert(set1, six);
}
static void getComponentsP(stHash *nodesToEdges, int64_t node,
stSortedSet *component) {
stIntTuple *key = stIntTuple_construct1( node);
stList *edges = stHash_search(nodesToEdges, key);
if (edges != NULL) {
stHash_remove(nodesToEdges, key);
for (int64_t i = 0; i < stList_length(edges); i++) {
stIntTuple *edge = stList_get(edges, i);
if (stSortedSet_search(component, edge) == NULL) {
stSortedSet_insert(component, edge);
}
/*
* Recursion on stack could equal the total number of nodes.
*/
getComponentsP(nodesToEdges, stIntTuple_get(edge, 0),
component);
getComponentsP(nodesToEdges, stIntTuple_get(edge, 1),
component);
}
stList_destruct(edges);
}
stIntTuple_destruct(key);
}
static void setup() {
teardown();
//Make nodes
nodes = stList_construct3(0, (void(*)(void *)) stIntTuple_destruct);
int64_t nodeNumber = st_randomInt(0, 1000);
for (int64_t i = 0; i < nodeNumber; i++) {
stList_append(nodes, stIntTuple_construct1( i));
}
//Make edges
edges = stList_construct3(0, (void(*)(void *)) stIntTuple_destruct);
float edgeProb = st_random();
for (int64_t i = 0; i < nodeNumber; i++) {
for (int64_t j = i; j < nodeNumber; j++) {
if (st_random() <= edgeProb) {
stList_append(edges, stIntTuple_construct3( st_randomInt(1, 100), i, j));
}
}
}
//Max component size
maxComponentSize = 1 + log(nodeNumber) * 10; //(st_randomInt(0, nodeNumber+1);
}
void cactusDisk_write(CactusDisk *cactusDisk) {
Flower *flower;
int64_t recordSize;
stList *removeRequests = stList_construct3(0, (void (*)(void *)) stIntTuple_destruct);
st_logDebug("Starting to write the cactus to disk\n");
stSortedSetIterator *it = stSortedSet_getIterator(cactusDisk->flowers);
//Sort flowers to update.
while ((flower = stSortedSet_getNext(it)) != NULL) {
cactusDisk_addUpdateRequest(cactusDisk, flower);
}
stSortedSet_destructIterator(it);
st_logDebug("Got the flowers to update\n");
//Remove nets that are marked for deletion..
it = stSortedSet_getIterator(cactusDisk->flowerNamesMarkedForDeletion);
char *nameString;
while ((nameString = stSortedSet_getNext(it)) != NULL) {
Name name = cactusMisc_stringToName(nameString);
if (containsRecord(cactusDisk, name)) {
stList_append(cactusDisk->updateRequests, stKVDatabaseBulkRequest_constructUpdateRequest(name, &name, 0)); //We set it to null in the first atomic operation.
stList_append(removeRequests, stIntTuple_construct1(name));
}
}
stSortedSet_destructIterator(it);
st_logDebug("Avoided updating nets marked for deletion\n");
// Insert and/or update meta-sequences.
it = stSortedSet_getIterator(cactusDisk->metaSequences);
MetaSequence *metaSequence;
while ((metaSequence = stSortedSet_getNext(it)) != NULL) {
void *vA =
binaryRepresentation_makeBinaryRepresentation(metaSequence,
(void (*)(void *, void (*)(const void * ptr, size_t size, size_t count))) metaSequence_writeBinaryRepresentation,
&recordSize);
//Compression
vA = compress(vA, &recordSize);
if (!containsRecord(cactusDisk, metaSequence_getName(metaSequence))) {
stList_append(cactusDisk->updateRequests,
stKVDatabaseBulkRequest_constructInsertRequest(metaSequence_getName(metaSequence), vA, recordSize));
} else {
stList_append(cactusDisk->updateRequests,
stKVDatabaseBulkRequest_constructUpdateRequest(metaSequence_getName(metaSequence), vA, recordSize));
}
free(vA);
}
stSortedSet_destructIterator(it);
st_logDebug("Got the sequences we are going to add to the database.\n");
if (!containsRecord(cactusDisk, CACTUS_DISK_PARAMETER_KEY)) { //We only write the parameters once.
//Finally the database info.
void *cactusDiskParameters =
binaryRepresentation_makeBinaryRepresentation(cactusDisk,
(void (*)(void *, void (*)(const void * ptr, size_t size, size_t count))) cactusDisk_writeBinaryRepresentation,
&recordSize);
//Compression
cactusDiskParameters = compress(cactusDiskParameters, &recordSize);
stList_append(cactusDisk->updateRequests,
stKVDatabaseBulkRequest_constructInsertRequest(CACTUS_DISK_PARAMETER_KEY, cactusDiskParameters,
recordSize));
free(cactusDiskParameters);
}
st_logDebug("Checked if need to write the initial parameters\n");
if (stList_length(cactusDisk->updateRequests) > 0) {
st_logDebug("Going to write %" PRIi64 " updates\n", stList_length(cactusDisk->updateRequests));
stTry
{
st_logDebug("Writing %" PRIi64 " updates\n", stList_length(cactusDisk->updateRequests));
assert(stList_length(cactusDisk->updateRequests) > 0);
stKVDatabase_bulkSetRecords(cactusDisk->database, cactusDisk->updateRequests);
}
stCatch(except)
{
stThrowNewCause(except, ST_KV_DATABASE_EXCEPTION_ID,
"Failed when trying to set records in updating the cactus disk");
}stTryEnd
;
}
static stHash *getScaffoldPathsP(stList *haplotypePaths, stHash *haplotypePathToScaffoldPathHash,
stList *haplotypeEventStrings, stList *contaminationEventStrings, CapCodeParameters *capCodeParameters) {
stHash *haplotypeToMaximalHaplotypeLengthHash = buildContigPathToContigPathLengthHash(haplotypePaths);
stHash *segmentToMaximalHaplotypePathHash = buildSegmentToContigPathHash(haplotypePaths);
for (int64_t i = 0; i < stList_length(haplotypePaths); i++) {
stSortedSet *bucket = stSortedSet_construct();
stHash_insert(haplotypePathToScaffoldPathHash, stList_get(haplotypePaths, i), bucket);
stSortedSet_insert(bucket, stList_get(haplotypePaths, i));
}
for (int64_t i = 0; i < stList_length(haplotypePaths); i++) {
stList *haplotypePath = stList_get(haplotypePaths, i);
assert(stList_length(haplotypePath) > 0);
Segment *_5Segment = stList_get(haplotypePath, 0);
if (!segment_getStrand(_5Segment)) {
_5Segment = segment_getReverse(stList_get(haplotypePath, stList_length(haplotypePath) - 1));
}
assert(segment_getStrand(_5Segment));
if (getAdjacentCapsSegment(segment_get5Cap(_5Segment)) != NULL) {
assert(!trueAdjacency(segment_get5Cap(_5Segment), haplotypeEventStrings));
}
int64_t insertLength;
int64_t deleteLength;
Cap *otherCap;
enum CapCode _5CapCode = getCapCode(segment_get5Cap(_5Segment), &otherCap, haplotypeEventStrings, contaminationEventStrings, &insertLength, &deleteLength, capCodeParameters);
if (_5CapCode == SCAFFOLD_GAP || _5CapCode == AMBIGUITY_GAP) {
assert(stHash_search(haplotypeToMaximalHaplotypeLengthHash, haplotypePath) != NULL);
int64_t j = stIntTuple_get(stHash_search(haplotypeToMaximalHaplotypeLengthHash, haplotypePath), 0);
Segment *adjacentSegment = getAdjacentCapsSegment(segment_get5Cap(_5Segment));
assert(adjacentSegment != NULL);
while (!hasCapInEvents(cap_getEnd(segment_get5Cap(adjacentSegment)), haplotypeEventStrings)) { //is not a haplotype end
adjacentSegment = getAdjacentCapsSegment(segment_get5Cap(adjacentSegment));
assert(adjacentSegment != NULL);
}
assert(adjacentSegment != NULL);
assert(hasCapInEvents(cap_getEnd(segment_get5Cap(adjacentSegment)), haplotypeEventStrings)); //is a haplotype end
stList *adjacentHaplotypePath = stHash_search(segmentToMaximalHaplotypePathHash, adjacentSegment);
if (adjacentHaplotypePath == NULL) {
adjacentHaplotypePath = stHash_search(segmentToMaximalHaplotypePathHash, segment_getReverse(
adjacentSegment));
}
assert(adjacentHaplotypePath != NULL);
assert(adjacentHaplotypePath != haplotypePath);
assert(stHash_search(haplotypeToMaximalHaplotypeLengthHash, adjacentHaplotypePath) != NULL);
int64_t k = stIntTuple_get(stHash_search(haplotypeToMaximalHaplotypeLengthHash, adjacentHaplotypePath), 0);
//Now merge the buckets and make new int tuples..
stSortedSet *bucket1 = stHash_search(haplotypePathToScaffoldPathHash, haplotypePath);
stSortedSet *bucket2 = stHash_search(haplotypePathToScaffoldPathHash, adjacentHaplotypePath);
assert(bucket1 != NULL);
assert(bucket2 != NULL);
assert(bucket1 != bucket2);
stSortedSet *bucket3 = stSortedSet_getUnion(bucket1, bucket2);
stSortedSetIterator *bucketIt = stSortedSet_getIterator(bucket3);
stList *l;
while ((l = stSortedSet_getNext(bucketIt)) != NULL) {
//Do the bucket first
assert(stHash_search(haplotypePathToScaffoldPathHash, l) == bucket1 || stHash_search(haplotypePathToScaffoldPathHash, l) == bucket2);
stHash_remove(haplotypePathToScaffoldPathHash, l);
stHash_insert(haplotypePathToScaffoldPathHash, l, bucket3);
//Now the length
stIntTuple *m = stHash_remove(haplotypeToMaximalHaplotypeLengthHash, l);
assert(m != NULL);
assert(stIntTuple_get(m, 0) == j || stIntTuple_get(m, 0) == k);
stHash_insert(haplotypeToMaximalHaplotypeLengthHash, l, stIntTuple_construct1( j + k));
stIntTuple_destruct(m);
}
assert(stHash_search(haplotypePathToScaffoldPathHash, haplotypePath) == bucket3);
assert(stHash_search(haplotypePathToScaffoldPathHash, adjacentHaplotypePath) == bucket3);
stSortedSet_destructIterator(bucketIt);
}
}
stHash_destruct(segmentToMaximalHaplotypePathHash);
return haplotypeToMaximalHaplotypeLengthHash;
}
