void erase_if_dispatch(Sequence& c, Predicate p,
sequence_tag, IteratorStability)
{
#if 0
c.erase(std::remove_if(c.begin(), c.end(), p), c.end());
#else
if (! c.empty())
c.erase(std::remove_if(c.begin(), c.end(), p), c.end());
#endif
}
void
sequence_iteration(
Sequence &_sequence,
UpdateAction const &_update_action
)
{
for(
auto it(
_sequence.begin()
);
it != _sequence.end();
)
{
switch(
_update_action(
*it
)
)
{
case fcppt::algorithm::update_action::remove:
it =
_sequence.erase(
it
);
break;
case fcppt::algorithm::update_action::keep:
++it;
break;
}
}
}
/** Append the sequence of contig v to seq. */
static void mergeContigs(const Graph& g, const Contigs& contigs,
vertex_descriptor u, vertex_descriptor v,
Sequence& seq, const ContigPath& path)
{
int d = get(edge_bundle, g, u, v).distance;
assert(d < 0);
unsigned overlap = -d;
const Sequence& s = sequence(contigs, v);
assert(s.length() > overlap);
Sequence ao;
Sequence bo(s, 0, overlap);
Sequence o;
do {
assert(seq.length() > overlap);
ao = seq.substr(seq.length() - overlap);
o = createConsensus(ao, bo);
if (!o.empty()) {
seq.resize(seq.length() - overlap);
seq += o;
seq += Sequence(s, overlap);
return;
}
} while (chomp(seq, 'n'));
// Try an overlap alignment.
if (opt::verbose > 2)
cerr << '\n';
vector<overlap_align> overlaps;
alignOverlap(ao, bo, 0, overlaps, false, opt::verbose > 2);
bool good = false;
if (!overlaps.empty()) {
assert(overlaps.size() == 1);
const overlap_align& o = overlaps.front();
unsigned matches = o.overlap_match;
const string& consensus = o.overlap_str;
float identity = (float)matches / consensus.size();
good = matches >= opt::minOverlap
&& identity >= opt::minIdentity;
if (opt::verbose > 2)
cerr << matches << " / " << consensus.size()
<< " = " << identity
<< (matches < opt::minOverlap ? " (too few)"
: identity < opt::minIdentity ? " (too low)"
: " (good)") << '\n';
}
if (good) {
assert(overlaps.size() == 1);
const overlap_align& o = overlaps.front();
seq.erase(seq.length() - overlap + o.overlap_t_pos);
seq += o.overlap_str;
seq += Sequence(s, o.overlap_h_pos + 1);
} else {
cerr << "warning: the head of " << get(vertex_name, g, v)
<< " does not match the tail of the previous contig\n"
<< ao << '\n' << bo << '\n' << path << endl;
seq += 'n';
seq += s;
}
}
void erase_dispatch(Sequence& c, const T& x,
sequence_tag)
{
c.erase(std::remove(c.begin(), c.end(), x), c.end());
}
/** Read a single record. */
Sequence FastaReader::read(string& id, string& comment,
char& anchor, string& q)
{
next_record:
id.clear();
comment.clear();
anchor = 0;
q.clear();
// Discard comments.
while (peek() == '#')
ignoreLines(1);
signed char recordType = peek();
Sequence s;
unsigned qualityOffset = 0;
if (eof() || recordType == EOF || ftell(m_in) >= m_end) {
string header;
getline(header);
return s;
} else if (recordType == '>' || recordType == '@') {
// Read the header.
string header;
getline(header);
istringstream headerStream(header);
headerStream >> recordType >> id >> ws;
std::getline(headerStream, comment);
// Ignore SAM headers.
if (id.length() == 2 && isupper(id[0]) && isupper(id[1])
&& comment.length() > 2 && comment[2] == ':')
goto next_record;
// Casava FASTQ format
if (comment.size() > 3
&& comment[1] == ':' && comment[3] == ':') {
// read, chastity, flags, index: 1:Y:0:AAAAAA
if (opt::chastityFilter && comment[2] == 'Y') {
m_unchaste++;
if (recordType == '@') {
ignoreLines(3);
} else {
while (peek() != '>' && peek() != '#'
&& ignoreLines(1))
;
}
goto next_record;
}
if (id.size() > 2 && id.rbegin()[1] != '/') {
// Add the read number to the ID.
id += '/';
id += comment[0];
}
}
getline(s);
if (recordType == '>') {
// Read a multi-line FASTA record.
string line;
while (peek() != '>' && peek() != '#'
&& getline(line))
s += line;
if (eof())
clear();
}
if (recordType == '@') {
char c = peek();
if (c != '+') {
die() << s << '\n' << header << '\n';
string line;
getline(line);
die() << "expected `+' and saw ";
if (eof())
cerr << "end-of-file\n";
else
cerr << "`" << c << "' near\n"
<< c << line << "\n";
exit(EXIT_FAILURE);
}
ignoreLines(1);
getline(q);
} else
q.clear();
if (s.empty()) {
die() << "sequence with ID `" << id << "' is empty\n";
exit(EXIT_FAILURE);
}
if (s.length() < opt::minLength) {
goto next_record;
}
bool colourSpace = isColourSpace(s);
if (colourSpace && !isdigit(s[0])) {
// The first character is the primer base. The second
// character is the dibase read of the primer and the
// first base of the sample, which is not part of the
// assembly.
assert(s.length() > 2);
anchor = colourToNucleotideSpace(s[0], s[1]);
s.erase(0, 2);
q.erase(0, 1);
}
if (!q.empty())
checkSeqQual(s, q);
if (opt::trimMasked && !colourSpace) {
// Removed masked (lower case) sequence at the beginning
// and end of the read.
size_t trimFront = 0;
while (trimFront <= s.length() && islower(s[trimFront]))
trimFront++;
size_t trimBack = s.length();
while (trimBack > 0 && islower(s[trimBack - 1]))
trimBack--;
s.erase(trimBack);
s.erase(0, trimFront);
if (!q.empty()) {
q.erase(trimBack);
q.erase(0, trimFront);
}
}
if (flagFoldCase())
transform(s.begin(), s.end(), s.begin(), ::toupper);
qualityOffset = 33;
} else {
