unsigned int SubsetPartition::output_partitioned_file(const std::string infilename,
const std::string outputfile,
bool output_unassigned,
CallbackFn callback,
void * callback_data)
{
IParser* parser = IParser::get_parser(infilename);
ofstream outfile(outputfile.c_str());
unsigned int total_reads = 0;
unsigned int reads_kept = 0;
unsigned int n_singletons = 0;
PartitionSet partitions;
Read read;
string seq;
std::string first_kmer;
HashIntoType kmer = 0;
const unsigned int ksize = _ht->ksize();
//
// go through all the reads, and take those with assigned partitions
// and output them.
//
while(!parser->is_complete()) {
read = parser->get_next_read();
seq = read.seq;
if (_ht->check_read(seq)) {
const char * kmer_s = seq.c_str();
bool found_tag = false;
for (unsigned int i = 0; i < seq.length() - ksize + 1; i++) {
kmer = _hash(kmer_s + i, ksize);
// is this a known tag?
if (partition_map.find(kmer) != partition_map.end()) {
found_tag = true;
break;
}
}
// all sequences should have at least one tag in them.
// assert(found_tag); @CTB currently breaks tests. give fn flag to
// disable.
PartitionID partition_id = 0;
if (found_tag) {
PartitionID * partition_p = partition_map[kmer];
if (partition_p == NULL ){
partition_id = 0;
n_singletons++;
} else {
partition_id = *partition_p;
partitions.insert(partition_id);
}
}
if (partition_id > 0 || output_unassigned) {
outfile << ">" << read.name << "\t" << partition_id;
outfile << "\n" << seq << "\n";
}
#ifdef VALIDATE_PARTITIONS
std::cout << "checking: " << read.name << "\n";
assert(is_single_partition(seq));
#endif // VALIDATE_PARTITIONS
total_reads++;
// run callback, if specified
if (total_reads % CALLBACK_PERIOD == 0 && callback) {
try {
callback("output_partitions", callback_data,
total_reads, reads_kept);
} catch (...) {
delete parser; parser = NULL;
outfile.close();
throw;
}
}
}
}
delete parser; parser = NULL;
return partitions.size() + n_singletons;
}
virtual unsigned int output_partitioned_file(const std::string infilename,
const std::string outputfilename,
CallbackFn callback=0,
void * callback_data=0)
{
IParser* parser = IParser::get_parser(infilename);
ofstream outfile(outputfilename.c_str());
unsigned int total_reads = 0;
unsigned int reads_kept = 0;
Read read;
string seq;
std::string first_kmer;
HashIntoType forward_hash = 0, reverse_hash = 0;
map<SetID, unsigned int> lReadCounts;
while(!parser->is_complete()) {
read = parser->get_next_read();
seq = read.seq;
if (check_read(seq)) {
first_kmer = seq.substr(0, _ksize);
// generate the hash for the first kmer in the read (fair amount of work)
HashIntoType bin = _hash(first_kmer.c_str(), _ksize, forward_hash, reverse_hash);
SetID lActualFinalSetID = _sets[ _set_IDs[ bin ] ]->getCurrentPrimarySetID();
lReadCounts[lActualFinalSetID]++;
outfile << ">" << read.name << "\t" << lActualFinalSetID
<< " " << "\n"
<< seq << "\n";
// reset the sequence info, increment read number
total_reads++;
// run callback, if specified
if (total_reads % CALLBACK_PERIOD == 0 && callback) {
try {
callback("do_truncated_partition/output", callback_data,
total_reads, reads_kept);
} catch (...) {
delete parser; parser = NULL;
outfile.close();
throw;
}
}
}
}
for ( map<SetID, unsigned int>::iterator lIt = lReadCounts.begin(); lIt != lReadCounts.end(); ++lIt )
{
cout << setw(10) << lIt->first;
cout << setw(10) << lIt->second << endl;
}
cout << setw(6) << "unique set count: "<< lReadCounts.size() << endl;
cout << endl;
delete parser; parser = NULL;
return lReadCounts.size();
}
void Hashtable::consume_fasta(const std::string &filename,
unsigned int &total_reads,
unsigned long long &n_consumed,
HashIntoType lower_bound,
HashIntoType upper_bound,
ReadMaskTable ** orig_readmask,
bool update_readmask,
CallbackFn callback,
void * callback_data)
{
total_reads = 0;
n_consumed = 0;
IParser* parser = IParser::get_parser(filename.c_str());
Read read;
string currName = "";
string currSeq = "";
//
// readmask stuff: were we given one? do we want to update it?
//
ReadMaskTable * readmask = NULL;
std::list<unsigned int> masklist;
if (orig_readmask && *orig_readmask) {
readmask = *orig_readmask;
}
//
// iterate through the FASTA file & consume the reads.
//
while(!parser->is_complete()) {
read = parser->get_next_read();
currSeq = read.seq;
currName = read.name;
// do we want to process it?
if (!readmask || readmask->get(total_reads)) {
// yep! process.
unsigned int this_n_consumed;
bool is_valid;
this_n_consumed = check_and_process_read(currSeq,
is_valid,
lower_bound,
upper_bound);
// was this an invalid sequence -> mark as bad?
if (!is_valid && update_readmask) {
if (readmask) {
readmask->set(total_reads, false);
} else {
masklist.push_back(total_reads);
}
} else { // nope -- count it!
n_consumed += this_n_consumed;
}
}
// reset the sequence info, increment read number
total_reads++;
// run callback, if specified
if (total_reads % CALLBACK_PERIOD == 0 && callback) {
try {
callback("consume_fasta", callback_data, total_reads, n_consumed);
} catch (...) {
throw;
}
}
}
//
// We've either updated the readmask in place, OR we need to create a
// new one.
//
if (orig_readmask && update_readmask && readmask == NULL) {
// allocate, fill in from masklist
readmask = new ReadMaskTable(total_reads);
list<unsigned int>::const_iterator it;
for(it = masklist.begin(); it != masklist.end(); ++it) {
readmask->set(*it, false);
}
*orig_readmask = readmask;
}
}
