/**
* Parse a single quality string from fb and store qualities in r.
* Assume the next character obtained via fb.get() is the first
* character of the quality string. When returning, the next
* character returned by fb.peek() or fb.get() should be the first
* character of the following line.
*/
int parseQuals(
Read& r,
FileBuf& fb,
int firstc,
int readLen,
int trim3,
int trim5,
bool intQuals,
bool phred64,
bool solexa64)
{
int c = firstc;
assert(c != '\n' && c != '\r');
r.qual.clear();
if (intQuals) {
while (c != '\r' && c != '\n' && c != -1) {
bool neg = false;
int num = 0;
while(!isspace(c) && !fb.eof()) {
if(c == '-') {
neg = true;
assert_eq(num, 0);
} else {
if(!isdigit(c)) {
char buf[2048];
cerr << "Warning: could not parse quality line:" << endl;
fb.getPastNewline();
cerr << fb.copyLastN(buf);
buf[2047] = '\0';
cerr << buf;
throw 1;
}
assert(isdigit(c));
num *= 10;
num += (c - '0');
}
c = fb.get();
}
if(neg) num = 0;
// Phred-33 ASCII encode it and add it to the back of the
// quality string
r.qual.append('!' + num);
// Skip over next stretch of whitespace
while(c != '\r' && c != '\n' && isspace(c) && !fb.eof()) {
c = fb.get();
}
}
} else {
while (c != '\r' && c != '\n' && c != -1) {
r.qual.append(charToPhred33(c, solexa64, phred64));
c = fb.get();
while(c != '\r' && c != '\n' && isspace(c) && !fb.eof()) {
c = fb.get();
}
}
}
if ((int)r.qual.length() < readLen-1 ||
((int)r.qual.length() < readLen && !r.color))
{
tooFewQualities(r.name);
}
r.qual.trimEnd(trim3);
if(r.qual.length()-trim5 < r.patFw.length()) {
assert(gColor && r.primer != -1);
assert_gt(trim5, 0);
trim5--;
}
r.qual.trimBegin(trim5);
if(r.qual.length() <= 0) return 0;
assert_eq(r.qual.length(), r.patFw.length());
while(fb.peek() == '\n' || fb.peek() == '\r') fb.get();
return (int)r.qual.length();
}
static void driver(
const string& infile,
EList<string>& infiles,
const string& snpfile,
const string& htfile,
const string& ssfile,
const string& exonfile,
const string& svfile,
const string& outfile,
bool packed,
int reverse)
{
initializeCntLut();
initializeCntBit();
EList<FileBuf*> is(MISC_CAT);
bool bisulfite = false;
RefReadInParams refparams(false, reverse, nsToAs, bisulfite);
assert_gt(infiles.size(), 0);
if(format == CMDLINE) {
// Adapt sequence strings to stringstreams open for input
stringstream *ss = new stringstream();
for(size_t i = 0; i < infiles.size(); i++) {
(*ss) << ">" << i << endl << infiles[i].c_str() << endl;
}
FileBuf *fb = new FileBuf(ss);
assert(fb != NULL);
assert(!fb->eof());
assert(fb->get() == '>');
ASSERT_ONLY(fb->reset());
assert(!fb->eof());
is.push_back(fb);
} else {
// Adapt sequence files to ifstreams
for(size_t i = 0; i < infiles.size(); i++) {
FILE *f = fopen(infiles[i].c_str(), "r");
if (f == NULL) {
cerr << "Error: could not open "<< infiles[i].c_str() << endl;
throw 1;
}
FileBuf *fb = new FileBuf(f);
assert(fb != NULL);
if(fb->peek() == -1 || fb->eof()) {
cerr << "Warning: Empty fasta file: '" << infile.c_str() << "'" << endl;
continue;
}
assert(!fb->eof());
assert(fb->get() == '>');
ASSERT_ONLY(fb->reset());
assert(!fb->eof());
is.push_back(fb);
}
}
if(is.empty()) {
cerr << "Warning: All fasta inputs were empty" << endl;
throw 1;
}
filesWritten.push_back(outfile + ".1." + gfm_ext);
filesWritten.push_back(outfile + ".2." + gfm_ext);
// Vector for the ordered list of "records" comprising the input
// sequences. A record represents a stretch of unambiguous
// characters in one of the input sequences.
EList<RefRecord> szs(MISC_CAT);
std::pair<size_t, size_t> sztot;
{
if(verbose) cerr << "Reading reference sizes" << endl;
Timer _t(cerr, " Time reading reference sizes: ", verbose);
if(!reverse && (writeRef || justRef)) {
filesWritten.push_back(outfile + ".3." + gfm_ext);
filesWritten.push_back(outfile + ".4." + gfm_ext);
sztot = BitPairReference::szsFromFasta(is, outfile, bigEndian, refparams, szs, sanityCheck);
} else {
sztot = BitPairReference::szsFromFasta(is, string(), bigEndian, refparams, szs, sanityCheck);
}
}
if(justRef) return;
assert_gt(sztot.first, 0);
assert_gt(sztot.second, 0);
assert_gt(szs.size(), 0);
// Construct index from input strings and parameters
filesWritten.push_back(outfile + ".5." + gfm_ext);
filesWritten.push_back(outfile + ".6." + gfm_ext);
filesWritten.push_back(outfile + ".7." + gfm_ext);
filesWritten.push_back(outfile + ".8." + gfm_ext);
TStr s;
HGFM<TIndexOffU> hGFM(
s,
packed,
1, // TODO: maybe not?
lineRate,
offRate, // suffix-array sampling rate
ftabChars, // number of chars in initial arrow-pair calc
localOffRate,
localFtabChars,
nthreads,
snpfile,
htfile,
ssfile,
exonfile,
svfile,
outfile, // basename for .?.ht2 files
reverse == 0, // fw
!entireSA, // useBlockwise
bmax, // block size for blockwise SA builder
bmaxMultSqrt, // block size as multiplier of sqrt(len)
bmaxDivN, // block size as divisor of len
noDc? 0 : dcv,// difference-cover period
is, // list of input streams
szs, // list of reference sizes
(TIndexOffU)sztot.first, // total size of all unambiguous ref chars
refparams, // reference read-in parameters
seed, // pseudo-random number generator seed
-1, // override offRate
verbose, // be talkative
autoMem, // pass exceptions up to the toplevel so that we can adjust memory settings automatically
sanityCheck); // verify results and internal consistency
// Note that the Ebwt is *not* resident in memory at this time. To
// load it into memory, call ebwt.loadIntoMemory()
if(verbose) {
// Print Ebwt's vital stats
hGFM.gh().print(cerr);
}
if(sanityCheck) {
// Try restoring the original string (if there were
// multiple texts, what we'll get back is the joined,
// padded string, not a list)
hGFM.loadIntoMemory(
reverse ? (refparams.reverse == REF_READ_REVERSE) : 0,
true, // load SA sample?
true, // load ftab?
true, // load rstarts?
false,
false);
SString<char> s2;
hGFM.restore(s2);
hGFM.evictFromMemory();
{
SString<char> joinedss = GFM<>::join<SString<char> >(
is, // list of input streams
szs, // list of reference sizes
(TIndexOffU)sztot.first, // total size of all unambiguous ref chars
refparams, // reference read-in parameters
seed); // pseudo-random number generator seed
if(refparams.reverse == REF_READ_REVERSE) {
joinedss.reverse();
}
assert_eq(joinedss.length(), s2.length());
assert(sstr_eq(joinedss, s2));
}
if(verbose) {
if(s2.length() < 1000) {
cout << "Passed restore check: " << s2.toZBuf() << endl;
} else {
cout << "Passed restore check: (" << s2.length() << " chars)" << endl;
}
}
}
}