/*
* Function does the actual depth first search to detect if the thing has an acyclic ordering.
*/
static int64_t dfs(stHash *adjacencyList, stIntTuple *seqPos,
stSortedSet *started, stSortedSet *done) {
if(stSortedSet_search(started, seqPos) != NULL) {
if(stSortedSet_search(done, seqPos) == NULL) {
//We have detected a cycle
//st_logInfo("I have cycle %" PRIi64 " %" PRIi64 "\n", stIntTuple_getPosition(seqPos, 0), stIntTuple_getPosition(seqPos, 1));
return 1;
}
//We have already explored this area, but no cycle.
return 0;
}
stSortedSet_insert(started, seqPos);
int64_t cycle =0;
stIntTuple *nextSeqPos = stIntTuple_construct2( stIntTuple_get(seqPos, 0), stIntTuple_get(seqPos, 1) + 1);
stSortedSet *column = stHash_search(adjacencyList, nextSeqPos);
if(column != NULL) { //It is in the adjacency list, so we can do the recursion
assert(stSortedSet_search(column, nextSeqPos) != NULL);
stSortedSetIterator *it = stSortedSet_getIterator(column);
stIntTuple *seqPos2;
while((seqPos2 = stSortedSet_getNext(it)) != NULL) {
cycle = cycle || dfs(adjacencyList, seqPos2, started, done);
}
stSortedSet_destructIterator(it);
}
stIntTuple_destruct(nextSeqPos);
stSortedSet_insert(done, seqPos);
return cycle;
}
static void test_stSortedSet_searchGreaterThan(CuTest* testCase) {
sonLibSortedSetTestSetup();
for(int32_t i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
//static int32_t sortedInput[] = { -10, -1, 1, 3, 5, 10, 12 };
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, -11)) == stSortedSet_search(sortedSet, stIntTuple_construct(1, -10)));
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, -10)) == stSortedSet_search(sortedSet, stIntTuple_construct(1, -1)));
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, -5)) == stSortedSet_search(sortedSet, stIntTuple_construct(1, -1)));
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, 1)) == stSortedSet_search(sortedSet, stIntTuple_construct(1, 3)));
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, 13)) == NULL);
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet,
stIntTuple_construct(1, 12)) == NULL);
for(int32_t i=0; i<100; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, st_randomInt(-1000, 1000)));
}
stList *list = stSortedSet_getList(sortedSet);
for(int32_t i=1; i<stList_length(list); i++) {
stIntTuple *p = stList_get(list, i-1);
stIntTuple *j = stList_get(list, i);
stIntTuple *k = stIntTuple_construct(1, st_randomInt(stIntTuple_getPosition(p, 0), stIntTuple_getPosition(j, 0)));
CuAssertTrue(testCase, stSortedSet_searchGreaterThan(sortedSet, k) == j);
stIntTuple_destruct(k);
}
stList_destruct(list);
sonLibSortedSetTestTeardown();
}
static void test_stSortedSetEquals(CuTest* testCase) {
sonLibSortedSetTestSetup();
CuAssertTrue(testCase, stSortedSet_equals(sortedSet, sortedSet));
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
CuAssertTrue(testCase, stSortedSet_equals(sortedSet, sortedSet));
CuAssertTrue(testCase, !stSortedSet_equals(sortedSet, sortedSet2));
for(i=1; i<size; i++) { //first argument is unique in input..
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, input[i]));
}
CuAssertTrue(testCase, !stSortedSet_equals(sortedSet, sortedSet2));
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, input[0]));
CuAssertTrue(testCase, stSortedSet_equals(sortedSet, sortedSet2));
stSortedSet *sortedSet3 = stSortedSet_construct(); //diff comparator..
CuAssertTrue(testCase, !stSortedSet_equals(sortedSet, sortedSet3));
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet3, stIntTuple_construct(1, input[i]));
}
CuAssertTrue(testCase, !stSortedSet_equals(sortedSet, sortedSet3));
stSortedSet_destruct(sortedSet3);
sonLibSortedSetTestTeardown();
}
static void test_stSortedSetIterator(CuTest* testCase) {
sonLibSortedSetTestSetup();
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
stSortedSetIterator *iterator = stSortedSet_getIterator(sortedSet);
CuAssertTrue(testCase, iterator != NULL);
for(i=0; i<sortedSize; i++) {
CuAssertIntEquals(testCase, sortedInput[i], stIntTuple_getPosition(stSortedSet_getNext(iterator), 0));
}
CuAssertTrue(testCase, stSortedSet_getNext(iterator) == NULL);
stSortedSetIterator *iterator2 = stSortedSet_copyIterator(iterator);
CuAssertTrue(testCase, iterator2 != NULL);
for(i=0; i<sortedSize; i++) {
CuAssertIntEquals(testCase, sortedInput[sortedSize - 1 - i], stIntTuple_getPosition(stSortedSet_getPrevious(iterator), 0));
CuAssertIntEquals(testCase, sortedInput[sortedSize - 1 - i], stIntTuple_getPosition(stSortedSet_getPrevious(iterator2), 0));
}
CuAssertTrue(testCase, stSortedSet_getPrevious(iterator) == NULL);
CuAssertTrue(testCase, stSortedSet_getPrevious(iterator2) == NULL);
stSortedSet_destructIterator(iterator);
stSortedSet_destructIterator(iterator2);
sonLibSortedSetTestTeardown();
}
static void test_stSortedSetIterator_getIteratorFrom(CuTest* testCase) {
sonLibSortedSetTestSetup();
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
stSortedSetIterator *iterator = stSortedSet_getIterator(sortedSet);
CuAssertTrue(testCase, iterator != NULL);
for(i=0; i<sortedSize; i++) {
stSortedSetIterator *it = stSortedSet_getIteratorFrom(sortedSet, stIntTuple_construct(1, sortedInput[i]));
stIntTuple *intTuple = stSortedSet_getNext(it);
CuAssertTrue(testCase, intTuple != NULL);
CuAssertIntEquals(testCase, sortedInput[i], stIntTuple_getPosition(intTuple, 0));
stSortedSet_destructIterator(it);
}
stTry {
stSortedSet_getIteratorFrom(sortedSet, stIntTuple_construct(1, 7)); //This number if not in the input.
CuAssertTrue(testCase, 0);
} stCatch(except) {
CuAssertTrue(testCase, stExcept_getId(except) == SORTED_SET_EXCEPTION_ID);
}
stTryEnd
sonLibSortedSetTestTeardown();
}
void test_stList_filter(CuTest *testCase) {
setup();
stSortedSet *set = stSortedSet_construct();
stSortedSet_insert(set, strings[0]);
stSortedSet_insert(set, strings[4]);
stList *list2 = stList_filterToExclude(list, set);
stList *list3 = stList_filterToInclude(list, set);
CuAssertTrue(testCase,stList_length(list2) == 3);
CuAssertTrue(testCase,stList_length(list3) == 2);
CuAssertTrue(testCase,stList_get(list2, 0) == strings[1]);
CuAssertTrue(testCase,stList_get(list2, 1) == strings[2]);
CuAssertTrue(testCase,stList_get(list2, 2) == strings[3]);
CuAssertTrue(testCase,stList_get(list3, 0) == strings[0]);
CuAssertTrue(testCase,stList_get(list3, 1) == strings[4]);
teardown();
}
static void getMetaSequencesForEventsP(stSortedSet *metaSequences,
Flower *flower, stList *eventStrings) {
//Iterate over the sequences in the flower.
Flower_SequenceIterator *seqIt = flower_getSequenceIterator(flower);
Sequence *sequence;
while ((sequence = flower_getNextSequence(seqIt)) != NULL) {
MetaSequence *metaSequence = sequence_getMetaSequence(sequence);
if (stringIsInList(event_getHeader(sequence_getEvent(sequence)),
eventStrings) == 0) {
if (stSortedSet_search(metaSequences, metaSequence) == NULL) {
stSortedSet_insert(metaSequences, metaSequence);
}
}
}
flower_destructSequenceIterator(seqIt);
//Recurse over the flowers
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getMetaSequencesForEventsP(metaSequences,
group_getNestedFlower(group), eventStrings);
}
}
flower_destructGroupIterator(groupIt);
}
/*
* Uses the functions above to build an adjacency list, then by DFS attempts to create
* a valid topological sort, returning non-zero if the graph contains a cycle.
*/
static int64_t containsACycle(stList *pairs, int64_t sequenceNumber) {
//Build an adjacency list structure..
stHash *adjacencyList = buildAdjacencyList(pairs, sequenceNumber);
//Do a topological sort of the adjacency list
stSortedSet *started = stSortedSet_construct3((int (*)(const void *, const void *))stIntTuple_cmpFn, NULL);
stSortedSet *done = stSortedSet_construct3((int (*)(const void *, const void *))stIntTuple_cmpFn, NULL);
int64_t cyclic = 0;
for(int64_t seq=0; seq<sequenceNumber; seq++) {
stIntTuple *seqPos = stIntTuple_construct2( seq, 0); //The following hacks avoid memory cleanup..
stSortedSet *column = stHash_search(adjacencyList, seqPos);
assert(column != NULL);
stIntTuple *seqPos2 = stSortedSet_search(column, seqPos);
assert(seqPos2 != NULL);
cyclic = cyclic || dfs(adjacencyList, seqPos2, started, done);
stIntTuple_destruct(seqPos);
}
//cleanup
stHashIterator *it = stHash_getIterator(adjacencyList);
stIntTuple *seqPos;
stSortedSet *columns = stSortedSet_construct2((void (*)(void *))stSortedSet_destruct);
while((seqPos = stHash_getNext(it)) != NULL) {
stSortedSet *column = stHash_search(adjacencyList, seqPos);
assert(column != NULL);
stSortedSet_insert(columns, column);
}
stHash_destructIterator(it);
stHash_destruct(adjacencyList);
stSortedSet_destruct(columns);
stSortedSet_destruct(started);
stSortedSet_destruct(done);
return cyclic;
}
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 void test_stSortedSetIntersection(CuTest* testCase) {
sonLibSortedSetTestSetup();
//Check intersection of empty sets is okay..
stSortedSet *sortedSet3 = stSortedSet_getIntersection(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_size(sortedSet3) == 0);
stSortedSet_destruct(sortedSet3);
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
//Check intersection of empty and non-empty set is empty.
sortedSet3 = stSortedSet_getIntersection(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_size(sortedSet3) == 0);
stSortedSet_destruct(sortedSet3);
//Check intersection of two non-empty, overlapping sets in correct.
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 0));
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 1));
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 5));
sortedSet3 = stSortedSet_getIntersection(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_size(sortedSet3) == 2);
stIntTuple *intTuple = stIntTuple_construct(1, 1);
CuAssertTrue(testCase, stSortedSet_search(sortedSet3, intTuple) != NULL);
stIntTuple_destruct(intTuple);
intTuple = stIntTuple_construct(1, 5);
CuAssertTrue(testCase, stSortedSet_search(sortedSet3, intTuple) != NULL);
stIntTuple_destruct(intTuple);
stSortedSet_destruct(sortedSet3);
//Check we get an exception with sorted sets with different comparators.
stSortedSet *sortedSet4 = stSortedSet_construct();
stTry {
stSortedSet_getIntersection(sortedSet, sortedSet4);
} stCatch(except) {
CuAssertTrue(testCase, stExcept_getId(except) == SORTED_SET_EXCEPTION_ID);
}
stTryEnd
stSortedSet_destruct(sortedSet4);
sonLibSortedSetTestTeardown();
}
stSortedSet *stSortedSet_copyConstruct(stSortedSet *sortedSet, void (*destructElementFn)(void *)) {
stSortedSet *sortedSet2 = stSortedSet_construct3(stSortedSet_getComparator(sortedSet)->compareFn, destructElementFn);
stSortedSetIterator *it = stSortedSet_getIterator(sortedSet);
void *o;
while((o = stSortedSet_getNext(it)) != NULL) {
stSortedSet_insert(sortedSet2, o);
}
stSortedSet_destructIterator(it);
return sortedSet2;
}
/*
* This builds an adjacency list structure for the the sequences. Every sequence-position
* has a column in the hash with which it can be aligned with.
*/
static stHash *buildAdjacencyList(stList *pairs, int64_t sequenceNumber) {
stHash *hash = stHash_construct3((uint64_t (*)(const void *))stIntTuple_hashKey,
(int (*)(const void *, const void *))stIntTuple_equalsFn,
(void (*)(void *))stIntTuple_destruct, NULL);
for(int64_t seq=0; seq<sequenceNumber; seq++) {
for(int64_t position=0; position<MAX_SEQUENCE_SIZE; position++) {
stIntTuple *seqPos = stIntTuple_construct2( seq, position);
stSortedSet *column = stSortedSet_construct3((int (*)(const void *, const void *))stIntTuple_cmpFn, NULL);
stSortedSet_insert(column, seqPos);
stHash_insert(hash, seqPos, column);
}
}
stListIterator *it = stList_getIterator(pairs);
stIntTuple *pair;
while((pair = stList_getNext(it)) != NULL) {
stIntTuple *seqPos1 = stIntTuple_construct2( stIntTuple_get(pair, 0), stIntTuple_get(pair, 1));
stIntTuple *seqPos2 = stIntTuple_construct2( stIntTuple_get(pair, 2), stIntTuple_get(pair, 3));
stSortedSet *column1 = stHash_search(hash, seqPos1);
assert(column1 != NULL);
stSortedSet *column2 = stHash_search(hash, seqPos2);
assert(column2 != NULL);
if(column1 != column2) { //Merge the columns
stSortedSetIterator *it2 = stSortedSet_getIterator(column2);
stIntTuple *seqPos3;
while((seqPos3 = stSortedSet_getNext(it2)) != NULL) {
assert(stSortedSet_search(column1, seqPos3) == NULL);
stSortedSet_insert(column1, seqPos3);
assert(stHash_search(hash, seqPos3) == column2);
stHash_insert(hash, seqPos3, column1);
assert(stHash_search(hash, seqPos3) == column1);
}
stSortedSet_destructIterator(it2);
stSortedSet_destruct(column2);
}
//Cleanup loop.
stIntTuple_destruct(seqPos1);
stIntTuple_destruct(seqPos2);
}
stList_destructIterator(it);
return hash;
}
static void test_stSortedSet_copyConstruct(CuTest* testCase) {
sonLibSortedSetTestSetup();
CuAssertTrue(testCase, sortedSet != NULL);
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
stSortedSet *sortedSet2 = stSortedSet_copyConstruct(sortedSet, NULL);
CuAssertTrue(testCase, stSortedSet_size(sortedSet2) == stSortedSet_size(sortedSet));
CuAssertTrue(testCase, stSortedSet_equals(sortedSet2, sortedSet));
stSortedSet_destruct(sortedSet2);
sonLibSortedSetTestTeardown();
}
stSortedSet *stSortedSet_getUnion(stSortedSet *sortedSet1, stSortedSet *sortedSet2) {
if(!stSortedSet_comparatorsEqual(sortedSet1, sortedSet2)) {
stThrowNew(SORTED_SET_EXCEPTION_ID, "Comparators are not equal for creating the union of two sorted sets");
}
stSortedSet *sortedSet3 = stSortedSet_construct3(stSortedSet_getComparator(sortedSet1)->compareFn, NULL);
//Add those from sortedSet1
stSortedSetIterator *it= stSortedSet_getIterator(sortedSet1);
void *o;
while((o = stSortedSet_getNext(it)) != NULL) {
stSortedSet_insert(sortedSet3, o);
}
stSortedSet_destructIterator(it);
//Add those from sortedSet2
it= stSortedSet_getIterator(sortedSet2);
while((o = stSortedSet_getNext(it)) != NULL) {
stSortedSet_insert(sortedSet3, o);
}
stSortedSet_destructIterator(it);
return sortedSet3;
}
static void getMaximalHaplotypePathsP3(Segment *segment,
stList *maximalHaplotypePath, stSortedSet *segmentSet, stList *eventStrings) {
stList_append(maximalHaplotypePath, segment);
assert(stSortedSet_search(segmentSet, segment) == NULL);
assert(stSortedSet_search(segmentSet, segment_getReverse(segment)) == NULL);
stSortedSet_insert(segmentSet, segment);
Cap *_3Cap = segment_get3Cap(segment);
if (trueAdjacency(_3Cap, eventStrings)) { //Continue on..
Segment *otherSegment = getAdjacentCapsSegment(_3Cap);
if (otherSegment != NULL) {
getMaximalHaplotypePathsP3(otherSegment, maximalHaplotypePath,
segmentSet, eventStrings);
}
}
}
void testCactusDisk_getUniqueID_Unique(CuTest* testCase) {
cactusDiskTestSetup();
stSortedSet *uniqueNames = stSortedSet_construct3(testCactusDisk_getUniqueID_UniqueP, free);
for (int64_t i = 0; i < 100000; i++) { //Gets a billion ids, checks we are good.
Name uniqueName = cactusDisk_getUniqueID(cactusDisk);
CuAssertTrue(testCase, uniqueName > 0);
CuAssertTrue(testCase, uniqueName < INT64_MAX);
CuAssertTrue(testCase, uniqueName != NULL_NAME);
char *cA = cactusMisc_nameToString(uniqueName);
CuAssertTrue(testCase, stSortedSet_search(uniqueNames, cA) == NULL);
CuAssertTrue(testCase, cactusMisc_stringToName(cA) == uniqueName);
stSortedSet_insert(uniqueNames, cA);
}
stSortedSet_destruct(uniqueNames);
cactusDiskTestTeardown();
}
static stSortedSet *getEventStrings(End *end, stList *eventStrings) {
stSortedSet *eventStringsSet = stSortedSet_construct3(
(int(*)(const void *, const void *)) strcmp, NULL);
End_InstanceIterator *instanceIt = end_getInstanceIterator(end);
Cap *cap;
while ((cap = end_getNext(instanceIt)) != NULL) {
const char *header = event_getHeader(cap_getEvent(cap));
for(int64_t i=0; i<stList_length(eventStrings); i++) {
if(strcmp(stList_get(eventStrings, i), header) == 0) {
stSortedSet_insert(eventStringsSet,
(void *) header);
}
}
}
end_destructInstanceIterator(instanceIt);
return eventStringsSet;
}
static void getMAFBlock2(Block *block, FILE *fileHandle) {
if (block_getLength(block) >= minimumBlockLength) {
//Calculate bases in the reference and other reference sequence
Block_InstanceIterator *instanceIt = block_getInstanceIterator(block);
bool includesReference = 0;
bool includesOtherReference = 0;
Segment *segment;
while ((segment = block_getNext(instanceIt)) != NULL) {
const char *segmentEvent = event_getHeader(
segment_getEvent(segment));
if (strcmp(segmentEvent, referenceEventString) == 0) {
includesReference = 1;
} else if (strcmp(segmentEvent, otherReferenceEventString) == 0) {
includesOtherReference = 1;
}
}
block_destructInstanceIterator(instanceIt);
if (ignoreOtherReferenceBlocks && includesOtherReference) {
return;
}
stSortedSet *otherSampleEvents = stSortedSet_construct3(
(int(*)(const void *, const void *)) strcmp, NULL);
instanceIt = block_getInstanceIterator(block);
int32_t sampleNumber = 0;
while ((segment = block_getNext(instanceIt)) != NULL) {
const char *segmentEvent = event_getHeader(
segment_getEvent(segment));
if (strcmp(segmentEvent, sampleEventString) == 0) {
sampleNumber++;
} else if (strcmp(segmentEvent, referenceEventString) != 0) {
stSortedSet_insert(otherSampleEvents, (void *) segmentEvent);
}
}
block_destructInstanceIterator(instanceIt);
baseCoverages[stSortedSet_size(otherSampleEvents)] += block_getLength(
block) * sampleNumber;
stSortedSet_destruct(otherSampleEvents);
referenceBases += includesReference ? block_getLength(block)
* sampleNumber : 0;
otherReferenceBases += includesOtherReference ? block_getLength(block)
* sampleNumber : 0;
}
}
stSortedSet *stSortedSet_getDifference(stSortedSet *sortedSet1, stSortedSet *sortedSet2) {
if(!stSortedSet_comparatorsEqual(sortedSet1, sortedSet2)) {
stThrowNew(SORTED_SET_EXCEPTION_ID, "Comparators are not equal for creating the sorted set difference");
}
stSortedSet *sortedSet3 = stSortedSet_construct3(stSortedSet_getComparator(sortedSet1)->compareFn, NULL);
//Add those from sortedSet1 only if they are not in sortedSet2
stSortedSetIterator *it= stSortedSet_getIterator(sortedSet1);
void *o;
while((o = stSortedSet_getNext(it)) != NULL) {
if(stSortedSet_search(sortedSet2, o) == NULL) {
stSortedSet_insert(sortedSet3, o);
}
}
stSortedSet_destructIterator(it);
return sortedSet3;
}
static void test_stSortedSet(CuTest* testCase) {
sonLibSortedSetTestSetup();
int32_t i;
CuAssertIntEquals(testCase, 0, stSortedSet_size(sortedSet));
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
CuAssertIntEquals(testCase, sortedSize, stSortedSet_size(sortedSet));
CuAssertIntEquals(testCase, sortedInput[0], stIntTuple_getPosition(stSortedSet_getFirst(sortedSet), 0));
CuAssertIntEquals(testCase, sortedInput[sortedSize-1], stIntTuple_getPosition(stSortedSet_getLast(sortedSet), 0));
for(i=0; i<sortedSize; i++) {
CuAssertIntEquals(testCase, sortedSize-i, stSortedSet_size(sortedSet));
stIntTuple *tuple = stIntTuple_construct(1, sortedInput[i]);
CuAssertTrue(testCase, stIntTuple_getPosition(stSortedSet_search(sortedSet, tuple), 0) == sortedInput[i]);
stSortedSet_remove(sortedSet, tuple);
CuAssertTrue(testCase, stSortedSet_search(sortedSet, tuple) == NULL);
stIntTuple_destruct(tuple);
}
sonLibSortedSetTestTeardown();
}
void testCactusDisk_getUniqueID_UniqueIntervals(CuTest* testCase) {
cactusDiskTestSetup();
stSortedSet *uniqueNames = stSortedSet_construct3(testCactusDisk_getUniqueID_UniqueP, free);
for (int64_t i = 0; i < 10; i++) { //Gets a billion ids, checks we are good.
int64_t intervalSize = st_randomInt(0, 100000);
Name uniqueName = cactusDisk_getUniqueIDInterval(cactusDisk, intervalSize);
for(int64_t j=0; j<intervalSize; j++) {
CuAssertTrue(testCase, uniqueName > 0);
CuAssertTrue(testCase, uniqueName < INT64_MAX);
CuAssertTrue(testCase, uniqueName != NULL_NAME);
char *cA = cactusMisc_nameToString(uniqueName);
CuAssertTrue(testCase, stSortedSet_search(uniqueNames, cA) == NULL);
CuAssertTrue(testCase, cactusMisc_stringToName(cA) == uniqueName);
stSortedSet_insert(uniqueNames, cA);
uniqueName++;
}
}
stSortedSet_destruct(uniqueNames);
cactusDiskTestTeardown();
}
/*
* Adds a prime less than (or equals) constraint to the list of prime constraints, removing any redundant constraints in the process.
* The or equals is specified by making the lessThanOrEquals argument non-zero.
*/
void addConstraint_lessThan(stPosetAlignment *posetAlignment, int64_t sequence1, int64_t position1, int64_t sequence2, int64_t position2, int64_t lessThanOrEquals) {
stSortedSet *constraintList = getConstraintList(posetAlignment, sequence1, sequence2);
assert(position1 != INT64_MAX);
assert(position2 != INT64_MAX);
stIntTuple *constraint1 = stIntTuple_construct3( position1, position2, lessThanOrEquals);
stIntTuple *constraint2;
while((constraint2 = stSortedSet_searchLessThanOrEqual(constraintList, constraint1)) != NULL) {
assert(stIntTuple_get(constraint2, 0) <= position1);
if(stIntTuple_get(constraint2, 1) >= position2) {
if(stIntTuple_get(constraint2, 1) == position2) { //Check we are not removing an equivalent or more severe constraint.
assert((!lessThanOrEquals && stIntTuple_get(constraint2, 2)) || stIntTuple_get(constraint2, 0) < position1);
}
stSortedSet_remove(constraintList, constraint2);
stIntTuple_destruct(constraint2);
}
else {
assert(stIntTuple_get(constraint2, 0) < position1); //Check the constraint does not overshadow our proposed constraint.
break;
}
}
stSortedSet_insert(constraintList, constraint1);
}
static void getOrderedSegmentsP(Flower *flower,
stSortedSet *segments) {
Flower_SegmentIterator *segmentIt = flower_getSegmentIterator(flower);
Segment *segment;
while ((segment = flower_getNextSegment(segmentIt)) != NULL) {
if (!segment_getStrand(segment)) {
segment = segment_getReverse(segment);
}
assert(stSortedSet_search(segments, segment) == NULL);
stSortedSet_insert(segments, segment);
}
flower_destructSegmentIterator(segmentIt);
//Recurse over the flowers
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while ((group = flower_getNextGroup(groupIt)) != NULL) {
if (group_getNestedFlower(group) != NULL) {
getOrderedSegmentsP(group_getNestedFlower(group),
segments);
}
}
flower_destructGroupIterator(groupIt);
}
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);
}
stSortedSet *loadEndAlignmentFromDisk(Flower *flower, FILE *fileHandle, End **end) {
stSortedSet *endAlignment =
stSortedSet_construct3((int (*)(const void *, const void *))alignedPair_cmpFn,
(void (*)(void *))alignedPair_destruct);
char *line = stFile_getLineFromFile(fileHandle);
if(line == NULL) {
*end = NULL;
return NULL;
}
Name flowerName;
int64_t lineNumber;
int64_t i = sscanf(line, "%" PRIi64 " %" PRIi64 "", &flowerName, &lineNumber);
if(i != 2 || lineNumber < 0) {
st_errAbort("We encountered a mis-specified name in loading the first line of an end alignment from the disk: '%s'\n", line);
}
free(line);
*end = flower_getEnd(flower, flowerName);
if(*end == NULL) {
st_errAbort("We encountered an end name that is not in the database: '%s'\n", line);
}
for(int64_t i=0; i<lineNumber; i++) {
line = stFile_getLineFromFile(fileHandle);
if(line == NULL) {
st_errAbort("Got a null line when parsing an end alignment\n");
}
int64_t sI1, sI2;
int64_t p1, st1, p2, st2, score1, score2;
int64_t i = sscanf(line, "%" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 " %" PRIi64 "", &sI1, &p1, &st1, &score1, &sI2, &p2, &st2, &score2);
(void)i;
if(i != 8) {
st_errAbort("We encountered a mis-specified name in loading an end alignment from the disk: '%s'\n", line);
}
stSortedSet_insert(endAlignment, alignedPair_construct(sI1, p1, st1, sI2, p2, st2, score1, score2));
free(line);
}
return endAlignment;
}
static void test_stSortedSetDifference(CuTest* testCase) {
sonLibSortedSetTestSetup();
//Check difference of empty sets is okay..
stSortedSet *sortedSet3 = stSortedSet_getDifference(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_size(sortedSet3) == 0);
stSortedSet_destruct(sortedSet3);
int32_t i;
for(i=0; i<size; i++) {
stSortedSet_insert(sortedSet, stIntTuple_construct(1, input[i]));
}
//Check difference of non-empty set / empty set is the non-empty.
sortedSet3 = stSortedSet_getDifference(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_equals(sortedSet, sortedSet3));
stSortedSet_destruct(sortedSet3);
//Check difference of two non-empty, overlapping sets in correct.
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 0));
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 1));
stSortedSet_insert(sortedSet2, stIntTuple_construct(1, 5));
sortedSet3 = stSortedSet_getDifference(sortedSet, sortedSet2);
CuAssertTrue(testCase, stSortedSet_size(sortedSet3) == stSortedSet_size(sortedSet) - 2);
CuAssertTrue(testCase, !stSortedSet_equals(sortedSet, sortedSet3));
stSortedSet_insert(sortedSet3, stIntTuple_construct(1, 1));
stSortedSet_insert(sortedSet3, stIntTuple_construct(1, 5));
CuAssertTrue(testCase, stSortedSet_equals(sortedSet, sortedSet3));
stSortedSet_destruct(sortedSet3);
//Check we get an exception when merging sorted sets with different comparators.
stSortedSet *sortedSet4 = stSortedSet_construct();
stTry {
stSortedSet_getDifference(sortedSet, sortedSet4);
CuAssertTrue(testCase, 0);
} stCatch(except) {
CuAssertTrue(testCase, stExcept_getId(except) == SORTED_SET_EXCEPTION_ID);
}
stTryEnd
stSortedSet_destruct(sortedSet4);
sonLibSortedSetTestTeardown();
}
stSortedSet *makeEndAlignment(StateMachine *sM, End *end, int64_t spanningTrees, int64_t maxSequenceLength,
bool useProgressiveMerging, float gapGamma,
PairwiseAlignmentParameters *pairwiseAlignmentBandingParameters) {
//Make an alignment of the sequences in the ends
//Get the adjacency sequences to be aligned.
Cap *cap;
End_InstanceIterator *it = end_getInstanceIterator(end);
stList *sequences = stList_construct3(0, (void (*)(void *))adjacencySequence_destruct);
stList *seqFrags = stList_construct3(0, (void (*)(void *))seqFrag_destruct);
stHash *endInstanceNumbers = stHash_construct2(NULL, free);
while((cap = end_getNext(it)) != NULL) {
if(cap_getSide(cap)) {
cap = cap_getReverse(cap);
}
AdjacencySequence *adjacencySequence = adjacencySequence_construct(cap, maxSequenceLength);
stList_append(sequences, adjacencySequence);
assert(cap_getAdjacency(cap) != NULL);
End *otherEnd = end_getPositiveOrientation(cap_getEnd(cap_getAdjacency(cap)));
stList_append(seqFrags, seqFrag_construct(adjacencySequence->string, 0, end_getName(otherEnd)));
//Increase count of seqfrags with a given end.
int64_t *c = stHash_search(endInstanceNumbers, otherEnd);
if(c == NULL) {
c = st_calloc(1, sizeof(int64_t));
assert(*c == 0);
stHash_insert(endInstanceNumbers, otherEnd, c);
}
(*c)++;
}
end_destructInstanceIterator(it);
//Get the alignment.
MultipleAlignment *mA = makeAlignment(sM, seqFrags, spanningTrees, 100000000, useProgressiveMerging, gapGamma, pairwiseAlignmentBandingParameters);
//Build an array of weights to reweight pairs in the alignment.
int64_t *pairwiseAlignmentsPerSequenceNonCommonEnds = st_calloc(stList_length(seqFrags), sizeof(int64_t));
int64_t *pairwiseAlignmentsPerSequenceCommonEnds = st_calloc(stList_length(seqFrags), sizeof(int64_t));
//First build array on number of pairwise alignments to each sequence, distinguishing alignments between sequences sharing
//common ends.
for(int64_t i=0; i<stList_length(mA->chosenPairwiseAlignments); i++) {
stIntTuple *pairwiseAlignment = stList_get(mA->chosenPairwiseAlignments, i);
int64_t seq1 = stIntTuple_get(pairwiseAlignment, 1);
int64_t seq2 = stIntTuple_get(pairwiseAlignment, 2);
assert(seq1 != seq2);
SeqFrag *seqFrag1 = stList_get(seqFrags, seq1);
SeqFrag *seqFrag2 = stList_get(seqFrags, seq2);
int64_t *pairwiseAlignmentsPerSequence = seqFrag1->rightEndId == seqFrag2->rightEndId
? pairwiseAlignmentsPerSequenceCommonEnds : pairwiseAlignmentsPerSequenceNonCommonEnds;
pairwiseAlignmentsPerSequence[seq1]++;
pairwiseAlignmentsPerSequence[seq2]++;
}
//Now calculate score adjustments.
double *scoreAdjustmentsNonCommonEnds = st_malloc(stList_length(seqFrags) * sizeof(double));
double *scoreAdjustmentsCommonEnds = st_malloc(stList_length(seqFrags) * sizeof(double));
for(int64_t i=0; i<stList_length(seqFrags); i++) {
SeqFrag *seqFrag = stList_get(seqFrags, i);
End *otherEnd = flower_getEnd(end_getFlower(end), seqFrag->rightEndId);
assert(otherEnd != NULL);
assert(stHash_search(endInstanceNumbers, otherEnd) != NULL);
int64_t commonInstanceNumber = *(int64_t *)stHash_search(endInstanceNumbers, otherEnd);
int64_t nonCommonInstanceNumber = stList_length(seqFrags) - commonInstanceNumber;
assert(commonInstanceNumber > 0 && nonCommonInstanceNumber >= 0);
assert(pairwiseAlignmentsPerSequenceNonCommonEnds[i] <= nonCommonInstanceNumber);
assert(pairwiseAlignmentsPerSequenceNonCommonEnds[i] >= 0);
assert(pairwiseAlignmentsPerSequenceCommonEnds[i] < commonInstanceNumber);
assert(pairwiseAlignmentsPerSequenceCommonEnds[i] >= 0);
//scoreAdjustmentsNonCommonEnds[i] = ((double)nonCommonInstanceNumber + commonInstanceNumber - 1)/(pairwiseAlignmentsPerSequenceNonCommonEnds[i] + pairwiseAlignmentsPerSequenceCommonEnds[i]);
//scoreAdjustmentsCommonEnds[i] = scoreAdjustmentsNonCommonEnds[i];
if(pairwiseAlignmentsPerSequenceNonCommonEnds[i] > 0) {
scoreAdjustmentsNonCommonEnds[i] = ((double)nonCommonInstanceNumber)/pairwiseAlignmentsPerSequenceNonCommonEnds[i];
assert(scoreAdjustmentsNonCommonEnds[i] >= 1.0);
assert(scoreAdjustmentsNonCommonEnds[i] <= nonCommonInstanceNumber);
}
else {
scoreAdjustmentsNonCommonEnds[i] = INT64_MIN;
}
if(pairwiseAlignmentsPerSequenceCommonEnds[i] > 0) {
scoreAdjustmentsCommonEnds[i] = ((double)commonInstanceNumber-1)/pairwiseAlignmentsPerSequenceCommonEnds[i];
assert(scoreAdjustmentsCommonEnds[i] >= 1.0);
assert(scoreAdjustmentsCommonEnds[i] <= commonInstanceNumber-1);
}
else {
scoreAdjustmentsCommonEnds[i] = INT64_MIN;
}
}
//Convert the alignment pairs to an alignment of the caps..
stSortedSet *sortedAlignment =
stSortedSet_construct3((int (*)(const void *, const void *))alignedPair_cmpFn,
(void (*)(void *))alignedPair_destruct);
while(stList_length(mA->alignedPairs) > 0) {
stIntTuple *alignedPair = stList_pop(mA->alignedPairs);
assert(stIntTuple_length(alignedPair) == 5);
int64_t seqIndex1 = stIntTuple_get(alignedPair, 1);
int64_t seqIndex2 = stIntTuple_get(alignedPair, 3);
AdjacencySequence *i = stList_get(sequences, seqIndex1);
AdjacencySequence *j = stList_get(sequences, seqIndex2);
assert(i != j);
int64_t offset1 = stIntTuple_get(alignedPair, 2);
int64_t offset2 = stIntTuple_get(alignedPair, 4);
int64_t score = stIntTuple_get(alignedPair, 0);
if(score <= 0) { //Happens when indel probs are included
score = 1; //This is the minimum
}
assert(score > 0 && score <= PAIR_ALIGNMENT_PROB_1);
SeqFrag *seqFrag1 = stList_get(seqFrags, seqIndex1);
SeqFrag *seqFrag2 = stList_get(seqFrags, seqIndex2);
assert(seqFrag1 != seqFrag2);
double *scoreAdjustments = seqFrag1->rightEndId == seqFrag2->rightEndId ? scoreAdjustmentsCommonEnds : scoreAdjustmentsNonCommonEnds;
assert(scoreAdjustments[seqIndex1] != INT64_MIN);
assert(scoreAdjustments[seqIndex2] != INT64_MIN);
AlignedPair *alignedPair2 = alignedPair_construct(
i->subsequenceIdentifier, i->start + (i->strand ? offset1 : -offset1), i->strand,
j->subsequenceIdentifier, j->start + (j->strand ? offset2 : -offset2), j->strand,
score*scoreAdjustments[seqIndex1], score*scoreAdjustments[seqIndex2]); //Do the reweighting here.
assert(stSortedSet_search(sortedAlignment, alignedPair2) == NULL);
assert(stSortedSet_search(sortedAlignment, alignedPair2->reverse) == NULL);
stSortedSet_insert(sortedAlignment, alignedPair2);
stSortedSet_insert(sortedAlignment, alignedPair2->reverse);
stIntTuple_destruct(alignedPair);
}
//Cleanup
stList_destruct(seqFrags);
stList_destruct(sequences);
free(pairwiseAlignmentsPerSequenceNonCommonEnds);
free(pairwiseAlignmentsPerSequenceCommonEnds);
free(scoreAdjustmentsNonCommonEnds);
free(scoreAdjustmentsCommonEnds);
multipleAlignment_destruct(mA);
stHash_destruct(endInstanceNumbers);
return sortedAlignment;
}
int main(int argc, char *argv[]) {
st_setLogLevelFromString(argv[1]);
st_logDebug("Set up logging\n");
stKVDatabaseConf *kvDatabaseConf = stKVDatabaseConf_constructFromString(argv[2]);
CactusDisk *cactusDisk = cactusDisk_construct(kvDatabaseConf, 0);
stKVDatabaseConf_destruct(kvDatabaseConf);
st_logDebug("Set up the flower disk\n");
Name flowerName = cactusMisc_stringToName(argv[3]);
Flower *flower = cactusDisk_getFlower(cactusDisk, flowerName);
int64_t totalBases = flower_getTotalBaseLength(flower);
int64_t totalEnds = flower_getEndNumber(flower);
int64_t totalFreeEnds = flower_getFreeStubEndNumber(flower);
int64_t totalAttachedEnds = flower_getAttachedStubEndNumber(flower);
int64_t totalCaps = flower_getCapNumber(flower);
int64_t totalBlocks = flower_getBlockNumber(flower);
int64_t totalGroups = flower_getGroupNumber(flower);
int64_t totalChains = flower_getChainNumber(flower);
int64_t totalLinkGroups = 0;
int64_t maxEndDegree = 0;
int64_t maxAdjacencyLength = 0;
int64_t totalEdges = 0;
Flower_EndIterator *endIt = flower_getEndIterator(flower);
End *end;
while((end = flower_getNextEnd(endIt)) != NULL) {
assert(end_getOrientation(end));
if(end_getInstanceNumber(end) > maxEndDegree) {
maxEndDegree = end_getInstanceNumber(end);
}
stSortedSet *ends = stSortedSet_construct();
End_InstanceIterator *capIt = end_getInstanceIterator(end);
Cap *cap;
while((cap = end_getNext(capIt)) != NULL) {
if(cap_getSequence(cap) != NULL) {
Cap *adjacentCap = cap_getAdjacency(cap);
assert(adjacentCap != NULL);
End *adjacentEnd = end_getPositiveOrientation(cap_getEnd(adjacentCap));
stSortedSet_insert(ends, adjacentEnd);
int64_t adjacencyLength = cap_getCoordinate(cap) - cap_getCoordinate(adjacentCap);
if(adjacencyLength < 0) {
adjacencyLength *= -1;
}
assert(adjacencyLength >= 1);
if(adjacencyLength >= maxAdjacencyLength) {
maxAdjacencyLength = adjacencyLength;
}
}
}
end_destructInstanceIterator(capIt);
totalEdges += stSortedSet_size(ends);
if(stSortedSet_search(ends, end) != NULL) { //This ensures we count self edges twice, so that the division works.
totalEdges += 1;
}
stSortedSet_destruct(ends);
}
assert(totalEdges % 2 == 0);
flower_destructEndIterator(endIt);
Flower_GroupIterator *groupIt = flower_getGroupIterator(flower);
Group *group;
while((group = flower_getNextGroup(groupIt)) != NULL) {
if(group_getLink(group) != NULL) {
totalLinkGroups++;
}
}
flower_destructGroupIterator(groupIt);
printf("flower name: %" PRIi64 " total bases: %" PRIi64 " total-ends: %" PRIi64 " total-caps: %" PRIi64 " max-end-degree: %" PRIi64 " max-adjacency-length: %" PRIi64 " total-blocks: %" PRIi64 " total-groups: %" PRIi64 " total-edges: %" PRIi64 " total-free-ends: %" PRIi64 " total-attached-ends: %" PRIi64 " total-chains: %" PRIi64 " total-link groups: %" PRIi64 "\n",
flower_getName(flower), totalBases, totalEnds, totalCaps, maxEndDegree, maxAdjacencyLength, totalBlocks, totalGroups, totalEdges/2, totalFreeEnds, totalAttachedEnds, totalChains, totalLinkGroups);
return 0;
}
void eventTree_addEvent(EventTree *eventTree, Event *event) {
stSortedSet_insert(eventTree->events, event);
}
