void parseArgs(){
_options = getInput();
//_sketchSize = _options->getInt(STR_SIMKA_SKETCH_SIZE);
_nbCores = _options->getInt(STR_NB_CORES);
_inputFilename1 = _options->getStr(STR_SIMKA_URI_INPUT_1);
_inputFilename2 = _options->getStr(STR_SIMKA_URI_INPUT_2);
_outputDir = _options->getStr(STR_URI_OUTPUT);
_start_i = _options->getInt("-start-i");
_start_j = _options->getInt("-start-j");
_n_i = _options->getInt("-n-i");
_n_j = _options->getInt("-n-j");
//_kmerSize = _options->getInt(STR_KMER_SIZE);
if(!System::file().doesExist(_outputDir)){
int ok = System::file().mkdir(_outputDir, -1);
if(ok != 0){
std::cerr << "Error: can't create output directory (" << _outputDir << ")" << std::endl;
exit(1);
}
}
}
int main (int argc, char* argv[])
{
/** We create a command line parser. */
OptionsParser parser ("bankgen");
const char* OUTPUT_PREFIX = "-out";
const char* SEQ_LEN = "-seq-len";
const char* READ_LEN = "-read-len";
const char* OVERLAP_LEN = "-overlap-len";
const char* COVERAGE = "-coverage";
parser.push_back (new OptionOneParam (OUTPUT_PREFIX, "output prefix", true));
parser.push_back (new OptionOneParam (SEQ_LEN, "sequence length", false, "1000000"));
parser.push_back (new OptionOneParam (READ_LEN, "read length", false, "150" ));
parser.push_back (new OptionOneParam (OVERLAP_LEN, "overlap between two reads", false, "50" ));
parser.push_back (new OptionOneParam (COVERAGE, "coverage", false, "3" ));
try
{
/** We parse the user options. */
IProperties* options = parser.parse (argc, argv);
/** We create the random sequence. */
IBank* randomBank = new BankRandom (1, options->getInt(SEQ_LEN));
LOCAL (randomBank);
/** We create the reads bank. */
IBank* readsBank = new BankSplitter (
randomBank,
options->getInt(READ_LEN),
options->getInt(OVERLAP_LEN),
options->getInt(COVERAGE)
);
LOCAL (readsBank);
/** We save the random bank. */
SaveAsFasta (randomBank, options->getStr(OUTPUT_PREFIX) + "_sequence.fa");
/** We save the reads bank. */
SaveAsFasta (readsBank, options->getStr(OUTPUT_PREFIX) + "_reads.fa");
}
catch (OptionFailure& e)
{
e.getParser().displayErrors (stdout);
e.getParser().displayHelp (stdout);
return EXIT_FAILURE;
}
return EXIT_SUCCESS;
}
int main (int argc, char* argv[])
{
/** We create a command line parser. */
OptionsParser parser ("GraphStats");
parser.push_back (new OptionOneParam (STR_URI_GRAPH, "graph input", true));
parser.push_back (new OptionOneParam (STR_NB_CORES, "nb cores", false, "0"));
try
{
/** We parse the user options. */
IProperties* options = parser.parse (argc, argv);
// We load the graph
Graph graph = Graph::load (options->getStr(STR_URI_GRAPH));
// We set the number of cores to be used. Use all available cores if set to 0.
size_t nbCores = options->getInt(STR_NB_CORES);
// We get an iterator for branching nodes of the graph.
// We use a progress iterator to get some progress feedback
ProgressGraphIterator<BranchingNode,ProgressTimer> itBranching (graph.iterator<BranchingNode>(), "statistics");
// We define some kind of unique identifier for a couple (indegree,outdegree)
typedef pair<size_t,size_t> InOut_t;
// We want to gather some statistics during the iteration.
// Note the use of ThreadObject: this object will be cloned N times (one object per thread) and each clone will
// be reachable within the iteration block through ThreadObject::operator()
ThreadObject <map <InOut_t, size_t> > topology;
// We dispatch the iteration on several cores. Note the usage of lambda expression here.
IDispatcher::Status status = Dispatcher(nbCores).iterate (itBranching, [&] (const BranchingNode& node)
{
// We retrieve the current instance of map <InOut_t,size_t> for the current running thread.
map <InOut_t,size_t>& localTopology = topology();
// We get branching nodes neighbors for the current branching node.
Graph::Vector<BranchingEdge> successors = graph.successors <BranchingEdge> (node);
Graph::Vector<BranchingEdge> predecessors = graph.predecessors<BranchingEdge> (node);
// We increase the occurrences number for the current couple (in/out) neighbors
localTopology [make_pair(predecessors.size(), successors.size())] ++;
});
// Now, the parallel processing is done. We want now to aggregate the information retrieved
// in each thread in a single map.
// We get each map<InOut_t,size_t> object filled in each thread, and we add its data into the "global" map.
// The global map is reachable through the ThreadObject::operator*. The "topology.foreach" will loop over
// all cloned object used in the threads.
topology.foreach ([&] (const map <InOut_t, size_t>& t)
{
// We update the occurrence of the current couple (in/out)
for_each (t.begin(), t.end(), [&] (const pair<InOut_t, size_t>& p) { (*topology)[p.first] += p.second; });
});
// We sort the statistics by decreasing occurrence numbers. Since map have its own ordering, we need to put all
// the data into a vector and sort it with our own sorting criteria.
vector < pair<InOut_t,size_t> > stats;
for (auto it = topology->begin(); it != topology->end(); it++) { stats.push_back (*it); }
sort (stats.begin(), stats.end(), [=] (const pair<InOut_t,size_t>& a, const pair<InOut_t,size_t>& b) { return a.second > b.second; });
printf ("\nThere are %d branching nodes with the following distribution: \n", itBranching.size());
size_t sum=0;
for (size_t i=0; i<stats.size(); i++)
{
sum += stats[i].second;
printf (" [in=%d out=%d] nb=%7d percent=%5.2f distrib=%5.2f\n",
stats[i].first.first,
stats[i].first.second,
stats[i].second,
100.0*(float)stats[i].second / (float)itBranching.size(),
100.0*(float)sum / (float)itBranching.size()
);
}
printf ("\nDone on %d cores in %.2f sec\n\n", status.nbCores, (float)status.time/1000.0);
}
catch (OptionFailure& e)
{
return e.displayErrors (std::cout);
}
catch (Exception& e)
{
std::cerr << "EXCEPTION: " << e.getMessage() << std::endl;
}
return EXIT_SUCCESS;
}
int main (int argc, char* argv[])
{
/** We create a command line parser. */
OptionsParser parser ("KmerTest");
parser.push_back (new OptionOneParam (STR_URI_INPUT, "bank input", true));
parser.push_back (new OptionOneParam (STR_KMER_SIZE, "kmer size", true));
parser.push_back (new OptionOneParam (STR_MINIMIZER_SIZE, "minimizer size", true));
parser.push_back (new OptionNoParam (STR_VERBOSE, "display kmers", false));
try
{
/** We parse the user options. */
IProperties* options = parser.parse (argc, argv);
// We get the kmer and minimizer sizes.
size_t kmerSize = options->getInt(STR_KMER_SIZE);
size_t mmerSize = options->getInt(STR_MINIMIZER_SIZE);
// We define a try/catch block in case some method fails (bad filename for instance)
u_int64_t nbKmers = 0;
bool display = options->get(STR_VERBOSE) != 0;
// We declare a Bank instance defined by a list of filenames
IBank* bank = Bank::open (options->getStr(STR_URI_INPUT));
LOCAL (bank);
// We declare a kmer model and a minimizer model
Model model (kmerSize, mmerSize);
// We get a reference on the minimizer model, which will be useful for dumping
const ModelMinimizer::Model& modelMinimizer = model.getMmersModel();
Kmer<span>::Type checksum;
size_t nbChanged = 0;
size_t nbInvalid = 0;
// We define an iterator that encapsulates the sequences iterator with progress feedback
ProgressIterator<Sequence> iter (*bank, "iterate bank");
// We loop over sequences.
for (iter.first(); !iter.isDone(); iter.next())
{
// Shortcut
Sequence& seq = iter.item();
//! [snippet1_iterate]
// We iterate the kmers (and minimizers) of the current sequence.
model.iterate (seq.getData(), [&] (const Model::Kmer& kmer, size_t idx)
{
nbKmers ++;
if (kmer.hasChanged() == true) { nbChanged++; }
if (kmer.isValid() == false) { nbInvalid++; }
checksum += kmer.minimizer().value();
});
//! [snippet1_iterate]
}
cout << "nbKmers : " << nbKmers << endl;
cout << "nbInvalid : " << nbInvalid << endl;
cout << "nbChanged : " << nbChanged << endl;
cout << "ratio : " << (nbChanged > 0 ? (double)nbKmers / (double)nbChanged : 0) << endl;
cout << "checksum : " << checksum << endl;
}
catch (OptionFailure& e)
{
return e.displayErrors (std::cout);
}
catch (Exception& e)
{
std::cerr << "EXCEPTION: " << e.getMessage() << std::endl;
}
return EXIT_SUCCESS;
}
int main (int argc, char* argv[])
{
/** We create a command line parser. */
OptionsParser parser ("BankSplitter");
parser.push_back (new OptionOneParam (STR_URI_INPUT, "bank reference", true));
parser.push_back (new OptionOneParam (STR_MAX_INPUT_SIZE, "average db size per split", true));
parser.push_back (new OptionOneParam (STR_URI_OUTPUT_DIR, "output directory", false, "."));
parser.push_back (new OptionNoParam (STR_OUTPUT_FASTQ, "fastq output", false));
parser.push_back (new OptionNoParam (STR_OUTPUT_GZ, "gzip output", false));
// We define a try/catch block in case some method fails (bad filename for instance)
try
{
/** We parse the user options. */
IProperties* options = parser.parse (argc, argv);
/** Shortcuts. */
u_int64_t maxDbSize = options->getInt(STR_MAX_INPUT_SIZE);
// We declare an input Bank
IBank* inputBank = Bank::open (options->getStr(STR_URI_INPUT));
LOCAL (inputBank);
// We get the basename of the input bank.
string inputBasename = System::file().getBaseName (options->getStr(STR_URI_INPUT));
/** We set the name of the output directory. */
stringstream ss; ss << inputBasename << "_S" << maxDbSize;
string outputDirName = ss.str();
/** We create the output directory. */
string outputDir = options->getStr(STR_URI_OUTPUT_DIR) + "/" + outputDirName;
System::file().mkdir (outputDir, S_IRWXU);
// We create the album bank.
BankAlbum album (outputDir + "/album.txt");
/** We get estimations about the bank. */
u_int64_t number, totalSize, maxSize;
inputBank->estimate (number, totalSize, maxSize);
u_int64_t estimationNbSeqToIterate = number;
// We create an iterator over the input bank
ProgressIterator<Sequence> itSeq (*inputBank, "split");
// We loop over sequences to get the exact number of sequences.
int64_t nbBanksOutput = -1;
u_int64_t nbSequences = 0;
u_int64_t dbSize = ~0;
bool isFastq = options->get(STR_OUTPUT_FASTQ) != 0;
bool isGzipped = options->get(STR_OUTPUT_GZ) != 0;
IBank* currentBank = 0;
for (itSeq.first(); !itSeq.isDone(); itSeq.next())
{
if (dbSize > maxDbSize)
{
if (currentBank != 0) { currentBank->flush(); currentBank->finalize(); }
nbBanksOutput ++;
/** We build the uri of the current bank. */
stringstream ss; ss << inputBasename << "_" << nbBanksOutput << (isFastq ? ".fastq" : ".fasta");
if (isGzipped) { ss << ".gz"; }
/** We create a new bank and put it in the album. */
currentBank = album.addBank (outputDir, ss.str(), isFastq, isGzipped);
/** We reinit the db size counter. */
dbSize = 0;
}
dbSize += itSeq->getDataSize();
/** We insert the sequence into the current output bank. */
currentBank->insert (*itSeq);
}
if (currentBank != 0) { currentBank->flush(); }
}
catch (OptionFailure& e)
{
return e.displayErrors (cout);
}
catch (Exception& e)
{
cerr << "EXCEPTION: " << e.getMessage() << endl;
}
}
