void eskoTest() {
unsigned step = 5000;
unsigned start = 30000;
unsigned stop = 120000;
unsigned topL = 300;
unsigned minL = 5;
for(unsigned cur = start; cur <= stop; cur += step) {
takeSubstring("data/ecoli.seq", "data/tmpseq.seq", 0, cur);
//takeSubstring("data/ecoli.seq", "data/tmpseq.seq", 278000, 293000);
unsigned L = minL;
unsigned R = topL;
unsigned ans = 0U;
while (L<=R) {
unsigned mid = (L+R) / 2U;
unsigned k = mid-1U;
Genome genome;
readGenome("data/tmpseq.seq", mid, genome);
AhoCorasick * aho = new AhoCorasick(genome);
aho->filterOverlaps(k);
NFA_Automata nfa(*aho, genome, 0, genome.generatedReads.size()-1);
delete aho;
DFA_Automata dfa(nfa, genome, 1.0);
// remove reads since we do not need them anymore
genome.generatedReads.resize(0);
genome.sequence.resize(0);
bool uniqueOk = dfa.isCOAUnique();
if (uniqueOk) {
R = mid-1U;
ans = mid;
} else {
L = mid+1U;
}
}
std::string seq = readSequence("data/tmpseq.seq");
std::cout << cur << " " << ans << " " << getLongestSingleRepeat(seq) << std::endl;
}
}
void run(size_t kFilter) {
Genome genome;
readGenome(sequenceFile, patternLen, genome);
size_t numReads = genome.generatedReads.size();
std::cout << genome.generatedReads.size() << " " <<
(genome.generatedReads[0].second - genome.generatedReads[0].first) << " " << kFilter << std::endl;
std::cout << "Longest repeat: " << getLongestSingleRepeat(genome.sequence) << std::endl;
double coverage = double(numReads * patternLen) / double(genome.sequence.length());
cout << numReads * patternLen << endl;
std::cout << "START AC" << std::endl;
AhoCorasick * aho = new AhoCorasick(genome);
std::cout << "END AC" << std::endl;
std::cout << "START FILTER" << std::endl;
aho->filterOverlaps(kFilter);
std::cout << "END FILTER" << std::endl;
std::cout << "START NFA" << std::endl;
NFA_Automata nfa(*aho, genome, 0, genome.generatedReads.size()-1);
std::cout << "END NFA" << std::endl;
// aho corasick is not needed once we have the nfa FOR NOW
delete aho;
#ifdef DEBUG
nfa.saveAsSVG("data/nfa.svg");
#endif
// create DFA from NFA
std::cout << "START DFA" << std::endl;
DFA_Automata dfa(nfa, genome, coverage);
// remove reads since we do not need them anymore
genome.generatedReads.resize(0);
genome.sequence.resize(0);
std::cout << "END DFA" << std::endl;
std::cout << "UNIQUE ? " << dfa.isCOAUnique() << std::endl;
size_t numLoops = Statistics::getNumberOfLoops(&dfa);
std::cout << numLoops << std::endl;
#ifdef DEBUG
dfa.saveAsSVG("data/dfa.svg");
#endif
}
void runBenchmark() {
typedef std::pair<size_t, size_t> TestEntry;
std::ofstream out("data/benchmark.txt");
int tval = 17;
out << std::setw(tval) << std::left << "AC_SZ" << std::setw(tval) << std::left
<< "AC_TIME" << std::setw(tval) << std::left
<< "NOA_ST_SZ" << std::setw(tval) << std::left
<< "NOA_TR_SZ" << std::setw(tval) << std::left
<< "NOA_TIME" << std::setw(tval) << std::left
<< "DOA_ST_SZ" << std::setw(tval) << std::left
<< "DOA_TR_SZ" << std::setw(tval) << std::left
<< "DOA_PATH_LEN" << std::setw(tval) << std::left
<< "DOA_CNT_LOOPS" << std::setw(tval) << std::left
<< "DOA_CNT_BR_NO" << std::setw(tval) << std::left
<< "DOA_CNT_ACC_NO" << std::setw(tval) << std::left
<< "DOA_TIME" << std::setw(tval) << std::left
<< "REF_LEN" << std::setw(tval) << std::left
<< "NUM_READS" << std::setw(tval) << std::left
<< "LEN_READS" << std::setw(tval) << std::left
<< "SAMPLING" << std::endl;
// seq length
size_t R = 300000;
std::vector<TestEntry> Tests = {
TestEntry(60, 180),
TestEntry(120, 180),
};
std::vector<double> Results;
Timer timer;
double dt;
size_t statesCnt;
size_t transitionsCnt;
for (size_t i = 0; i < Tests.size(); ++i) {
// prepare and load test
size_t K = Tests[i].first;
size_t L = Tests[i].second;
takePrefix("data/ecoli.seq", "data/test.seq", R);
Genome genome;
readGenome("data/test.seq", L, genome);
double coverage = double(genome.generatedReads.size() * L) / double(R);
TestResult tmpResult;
// general stuff
tmpResult.REFERENCE_LENGTH = R;
tmpResult.NUMBER_OF_READS = genome.generatedReads.size();
tmpResult.SAMPLING = SAMPLING_TYPE::REGULARLY_SPACED;
tmpResult.TOTAL_LENGTH_OF_READS = genome.generatedReads.size() * L;
std::cout << "START AHO" << std::endl;
// AHO
timer.start();
AhoCorasick * aho = new AhoCorasick(genome);
aho->filterOverlaps(K);
timer.stop();
dt = timer.getElapsedTimeInSeconds();
Statistics::calcACStates(aho, statesCnt);
tmpResult.AHO_SIZE = statesCnt;
tmpResult.AHO_TIME = dt;
std::cout << "END AHO" << std::endl;
std::cout << "START NFA" << std::endl;
timer.start();
NFA_Automata * nfa = new NFA_Automata(*aho, genome, 0, genome.generatedReads.size()-1);
timer.stop();
dt = timer.getElapsedTimeInSeconds();
std::cout << "END NFA" << std::endl;
// clean not needed data
Statistics::calcNOAStatesAndTransitions(nfa, statesCnt, transitionsCnt);
tmpResult.NOA_STATES_SIZE = statesCnt;
tmpResult.NOA_TRANSITIONS_SIZE = transitionsCnt;
tmpResult.NOA_TIME = dt;
// clean not needed data
delete aho;
std::cout << "START DFA" << std::endl;
timer.start();
DFA_Automata * dfa = new DFA_Automata(*nfa, genome, coverage);
timer.stop();
dt = timer.getElapsedTimeInSeconds();
std::cout << "END DFA" << std::endl;
// clean not needed data
delete nfa;
genome.generatedReads.clear();
genome.sequence.clear();
std::cout << "START STORING REV EDGES" << std::endl;
dfa->storeReversedEdges();
std::cout << "END STORING REV EDGES" << std::endl;
std::cout << "START UPDATING UNITIG PATHS" << std::endl;
dfa->updateUnitigPaths();
std::cout << "STOP UPDATING UNITIG PATHS" << std::endl;
Statistics::calcDOAStatesAndTransitions(dfa, statesCnt, transitionsCnt);
tmpResult.DOA_STATES_SIZE = statesCnt;
tmpResult.DOA_TRANSITIONS_SIZE = transitionsCnt;
tmpResult.DOA_TIME = dt;
tmpResult.DOA_NUMBER_BRANCHING_NODES = Statistics::getBranchingNodesCount(dfa);
tmpResult.DOA_NUMBER_OF_LOOPS = Statistics::getNumberOfLoops(dfa);
tmpResult.DOA_SHORTEST_PATH_LEN = Statistics::getShortestPathToAllAcceptStates(dfa);
tmpResult.DOA_NUM_ACCEPT_STATES = dfa->m_acceptStates.size();
// clean not needed data
delete dfa;
std::cout << "DONE" << std::endl;
out << tmpResult << std::endl;
}
out.close();
}
