void outputChunk(struct psl **pPslList, char *tempDir, int midIx, boolean noHead)
/* Sort and write out pslList and free it. */
{
char fileName[512];
FILE *f;
struct psl *psl;
if (*pPslList == NULL)
return; /* Empty. */
psl = *pPslList;
//slSort(pPslList, pslCmpTarget);
makeMidName(tempDir, midIx, fileName);
if (stripVer)
{
char *s = stringIn(".",psl->qName);
if (s != NULL)
*s = 0;
}
if (chunkSize ==1)
safef(fileName, sizeof(fileName), "%s/%s.psl",tempDir,psl->qName);
f = mustOpen(fileName, "w");
if (!noHead)
pslWriteHead(f);
for (psl = *pPslList; psl != NULL; psl = psl->next)
pslTabOut(psl, f);
fclose(f);
pslFreeList(pPslList);
}
void pslCut(char *cutList, char *inPsl, char *outPsl)
/* pslCut - Remove a list of clones from psl file.. */
{
struct hash *cutHash = newHash(0);
struct lineFile *lf = pslFileOpen(inPsl);
FILE *f = mustOpen(outPsl, "w");
struct psl *psl;
char cloneName[128];
int total = 0, cut = 0;
buildCutHash(cutList, cutHash);
pslWriteHead(f);
while ((psl = pslNext(lf)) != NULL)
{
fragToCloneName(psl->tName, cloneName);
if (!hashLookup(cutHash, cloneName))
{
pslTabOut(psl, f);
}
else
++cut;
++total;
pslFree(&psl);
}
printf("Cut %d of %d\n", cut, total);
}
void pslCopyInClones(char *listFile, char *partDir, char *outName)
/* Copy in the .psl files corresponding to the clones named in listFile. */
{
struct slName *inList, *inEl;
FILE *out = mustOpen(outName, "w");
struct psl *psl;
int pslCount = 0;
int fileCount = 0;
pslWriteHead(out);
inList = getFileList(listFile, partDir);
for (inEl = inList; inEl != NULL; inEl = inEl->next)
{
char *inName = inEl->name;
struct lineFile *lf = pslFileOpen(inName);
++fileCount;
while ((psl = pslNext(lf)) != NULL)
{
pslTabOut(psl, out);
pslFree(&psl);
++pslCount;
}
lineFileClose(&lf);
}
printf("%d psls in %d files written to %s\n", pslCount, fileCount, outName);
fclose(out);
}
void fastaToPsl(char *inName, char *outName)
/* fastaToPsl - Convert axt to psl format. */
{
struct lineFile *inLF;
FILE *outFh;
boolean read;
struct psl* pslAlign;
DNA *qSeq;
int qSize;
int qSeqLen;
char *qHeader;
DNA *tSeq;
int tSize;
int tSeqLen;
char *tHeader;
int queryCounter;
inLF = lineFileOpen(inName, TRUE);
outFh = mustOpen(outName, "w");
/* read the target sequence */
read = faMixedSpeedReadNext(inLF, &qSeq, &qSize, &qHeader);
if (!read)
errAbort("Could not read target FASTA entry.");
qSeq = cloneString(qSeq);
qSeqLen = countNonDash(qSeq, qSize);
qHeader = cloneString(qHeader);
verbose(2, "Query sequence header: %s\n", qHeader);
verbose(3, "Query sequence alignment length: %d\n", qSize);
verbose(3, "Query sequence length: %d\n", qSeqLen);
verbose(4, "Query sequence: %s\n", qSeq);
/* read the rest of the sequences */
queryCounter = 1;
pslWriteHead(outFh);
while (faMixedSpeedReadNext(inLF, &tSeq, &tSize, &tHeader))
{
tSeqLen = countNonDash(tSeq, tSize);
verbose(2, "Target sequence (%d) header: %s\n", queryCounter, tHeader);
verbose(3, "Target sequence (%d) length: %d\n", queryCounter, tSeqLen);
verbose(4, "Target sequence (%d): %s\n", queryCounter, tSeq);
pslAlign = pslFromAlign(qHeader, qSeqLen, 0, qSeqLen, qSeq,
tHeader, tSeqLen, 0, tSeqLen, tSeq,
"+", 0);
pslTabOut(pslAlign, outFh);
++queryCounter;
}
lineFileClose(&inLF);
}
void pslGlue(char *inNames[], int inCount, char *outName, char *glueName)
/* Reduce a psl file to only the gluing components. */
{
FILE *out;
FILE *glue;
struct psl *pslList = NULL, *psl, *nextPsl;
int i;
struct psl *localList = NULL;
int glueCount = 0;
int pslCount = 0;
printf("Reading");
for (i=0; i<inCount; ++i)
{
char *inName = inNames[i];
struct lineFile *lf = pslFileOpen(inName);
printf(" %s", inName);
fflush(stdout);
while ((psl = pslNext(lf)) != NULL)
{
slAddHead(&pslList, psl);
++pslCount;
}
lineFileClose(&lf);
}
printf("\n");
slSort(&pslList, pslCmpQuery);
out = mustOpen(outName, "w");
glue = mustOpen(glueName, "w");
pslWriteHead(out);
/* Chop this up into chunks that share the same query. */
for (psl = pslList; psl != NULL; psl = nextPsl)
{
nextPsl = psl->next;
if (localList != NULL)
{
if (!sameString(localList->qName, psl->qName))
{
glueCount += simpleOut(out, glue, &localList);
localList = NULL;
}
}
slAddHead(&localList, psl);
}
glueCount += simpleOut(out, glue, &localList);
printf("Got %d gluing mRNAs out of %d psls in %d bundles %d ltot %d mtot\n",
glueCount, pslCount, outCount, ltot, mtot);
fclose(out);
fclose(glue);
}
void pslReps(char *inName, char *bestAliName, char *repName)
/* Analyse inName and put best alignments for eacmRNA in estAliName.
* Put repeat info in repName. */
{
struct lineFile *in = pslFileOpen(inName);
FILE *bestFile = mustOpen(bestAliName, "w");
FILE *repFile = mustOpen(repName, "w");
int lineSize;
char *line;
char *words[32];
int wordCount;
struct psl *pslList = NULL, *psl = NULL;
char lastName[512];
int aliCount = 0;
quiet = sameString(bestAliName, "stdout") || sameString(repName, "stdout");
if (coverQSizeFile != NULL)
loadCoverQSizes(coverQSizeFile);
if (!quiet)
printf("Processing %s to %s and %s\n", inName, bestAliName, repName);
if (!noHead)
pslWriteHead(bestFile);
strcpy(lastName, "");
while (lineFileNext(in, &line, &lineSize))
{
if (((++aliCount & 0x1ffff) == 0) && !quiet)
{
printf(".");
fflush(stdout);
}
wordCount = chopTabs(line, words);
if (wordCount == 21)
psl = pslLoad(words);
else if (wordCount == 23)
psl = pslxLoad(words);
else
errAbort("Bad line %d of %s\n", in->lineIx, in->fileName);
if (!sameString(lastName, psl->qName))
{
doOneAcc(lastName, pslList, bestFile, repFile);
pslFreeList(&pslList);
safef(lastName, sizeof(lastName), "%s", psl->qName);
}
slAddHead(&pslList, psl);
}
doOneAcc(lastName, pslList, bestFile, repFile);
pslFreeList(&pslList);
lineFileClose(&in);
fclose(bestFile);
fclose(repFile);
if (!quiet)
printf("Processed %d alignments\n", aliCount);
}
void pslGlueRna(char *listFile, char *partDir, char *pslName, char *gluName)
/* Reduce a psl files for only the gluing mRNA/EST components. */
{
FILE *pslOut;
FILE *gluOut;
struct psl *pslList = NULL, *psl, *nextPsl;
struct psl *localList = NULL;
int glueCount = 0;
int pslCount = 0;
struct slName *inList, *inEl;
inList = getFileList(listFile, partDir);
for (inEl = inList; inEl != NULL; inEl = inEl->next)
{
char *inName = inEl->name;
struct lineFile *lf = pslFileOpen(inName);
while ((psl = pslNext(lf)) != NULL)
{
slAddHead(&pslList, psl);
++pslCount;
}
lineFileClose(&lf);
}
slSort(&pslList, pslCmpQuery);
pslOut = mustOpen(pslName, "w");
gluOut = mustOpen(gluName, "w");
pslWriteHead(pslOut);
/* Chop this up into chunks that share the same query. */
for (psl = pslList; psl != NULL; psl = nextPsl)
{
nextPsl = psl->next;
if (localList != NULL)
{
if (!sameString(localList->qName, psl->qName))
{
glueCount += output(pslOut, gluOut, &localList);
localList = NULL;
}
}
slAddHead(&localList, psl);
}
glueCount += output(pslOut, gluOut, &localList);
printf("Got %d gluing mRNAs out of %d psls in %d bundles %d ltot %d mtot to %s\n",
glueCount, pslCount, outCount, ltot, mtot, gluName);
fclose(pslOut);
fclose(gluOut);
}
void outputChunk(struct psl **pPslList, char *tempDir, int midIx)
/* Sort and write out pslList and free it. */
{
char fileName[512];
FILE *f;
struct psl *psl;
if (*pPslList == NULL)
return; /* Empty. */
slSort(pPslList, pslCmpTarget);
makeMidName(tempDir, midIx, fileName);
f = mustOpen(fileName, "w");
pslWriteHead(f);
for (psl = *pPslList; psl != NULL; psl = psl->next)
pslTabOut(psl, f);
fclose(f);
pslFreeList(pPslList);
}
int main(int argc, char *argv[])
{
char *genoListName;
char *otherListName;
char *oocFileName;
char *typeName;
char *outName;
struct patSpace *patSpace;
long startTime, endTime;
char **genoList;
int genoListSize;
char *genoListBuf;
char **otherList;
int otherListSize;
char *otherListBuf;
char *genoName;
int i;
int blockCount = 0;
struct dnaSeq **seqListList = NULL, *seq = NULL;
char *outRoot;
struct sqlConnection *conn;
enum ffStringency stringency = ffCdna;
int seedSize = 10;
FILE *out;
boolean noHead = FALSE;
struct repeatTracker *rt;
struct hash *repeatHash = newHash(10);
hostName = getenv("HOST");
pushWarnHandler(warnHandler);
startTime = clock1();
cgiSpoof(&argc, argv);
minMatch = cgiOptionalInt("minMatch", minMatch);
maxBad = cgiOptionalInt("maxBad", maxBad);
minBases = cgiOptionalInt("minBases", minBases);
dnaUtilOpen();
#ifdef DEBUG
/* Hard wire command line input so don't have to type it in each
* time run the stupid Gnu debugger. */
genoListName = "pFoo/geno.lst";
otherListName = "pFoo/bacend.lst";
typeName = "genomic";
oocFileName = "/d/biodata/human/10.ooc";
outName = "pFoo/pFoo.psl";
#else
if (argc != 6 && argc != 7)
usage();
genoListName = argv[1];
otherListName = argv[2];
typeName = argv[3];
oocFileName = argv[4];
if (sameWord(oocFileName, "none"))
oocFileName = NULL;
outName = argv[5];
if (argc == 7)
{
if (sameWord("noHead", argv[6]))
noHead = TRUE;
else
usage();
}
#endif
if (sameWord(typeName, "mRNA") || sameWord(typeName, "cDNA"))
{
stringency = ffCdna;
}
else if (sameWord(typeName, "genomic"))
{
stringency = ffTight;
}
else if (sameWord(typeName, "g2g"))
{
stringency = ffTight;
veryTight = TRUE;
seedSize = 11;
}
else if (sameString(typeName, "asm"))
{
stringency = ffTight;
avoidSelfSelf = TRUE;
}
else
{
warn("Unrecognized otherType %s\n", typeName);
usage();
}
readAllWordsOrFa(genoListName, &genoList, &genoListSize, &genoListBuf);
filterMissingFiles(genoList, &genoListSize);
if (genoListSize <= 0)
errAbort("There are no files that exist in %s\n", genoListName);
readAllWordsOrFa(otherListName, &otherList, &otherListSize, &otherListBuf);
if (otherListSize <= 0)
errAbort("There are no files that exist in %s\n", otherListName);
filterMissingFiles(otherList, &otherListSize);
out = mustOpen(outName, "w");
if (!noHead)
pslWriteHead(out);
AllocArray(seqListList, genoListSize);
for (i=0; i<genoListSize; ++i)
{
genoName = genoList[i];
if (!startsWith("#", genoName) )
seqListList[i] = seq = faReadAllDna(genoName);
for (;seq != NULL; seq = seq->next)
{
int size = seq->size;
char *name = seq->name;
struct hashEl *hel;
AllocVar(rt);
AllocArray(rt->repBytes, size);
rt->seq = seq;
if ((hel = hashLookup(repeatHash, name)) != NULL)
errAbort("Duplicate %s in %s\n", name, genoName);
hashAdd(repeatHash, name, rt);
}
storeMasked(repeatHash, genoName);
}
patSpace = makePatSpace(seqListList, genoListSize, seedSize, oocFileName, minMatch, 2000);
endTime = clock1();
printf("Made index in %ld seconds\n", (endTime-startTime));
startTime = endTime;
for (i=0; i<otherListSize; ++i)
{
FILE *f;
char *otherName;
int c;
int dotCount = 0;
struct dnaSeq otherSeq;
ZeroVar(&otherSeq);
otherName = otherList[i];
if (startsWith("#", otherName) )
continue;
f = mustOpen(otherName, "r");
while ((c = fgetc(f)) != EOF)
if (c == '>')
break;
printf("%s\n", otherName);
fflush(stdout);
while (faFastReadNext(f, &otherSeq.dna, &otherSeq.size, &otherSeq.name))
{
aliSeqName = otherSeq.name;
oneStrand(patSpace, repeatHash, &otherSeq, FALSE, stringency, out);
reverseComplement(otherSeq.dna, otherSeq.size);
oneStrand(patSpace, repeatHash, &otherSeq, TRUE, stringency, out);
aliSeqName = NULL;
}
fclose(f);
}
freePatSpace(&patSpace);
endTime = clock1();
printf("Alignment time is %ld sec\n", (endTime-startTime));
startTime = endTime;
fclose(out);
return 0;
}
void liftPsl(char *destFile, struct hash *liftHash, int sourceCount, char *sources[],
boolean querySide, boolean isExtended)
/* Lift up coordinates in .psl file. */
{
FILE *dest = mustOpen(destFile, "w");
char *source;
int i,j;
struct lineFile *lf;
struct psl *psl;
struct xAli *xa = NULL;
unsigned *starts;
unsigned *blockSizes;
struct liftSpec *spec;
int offset;
int blockCount;
char *seqName;
int dotMod = dots;
int seqSize;
int strandChar = (querySide ? 0 : 1);
if (!nohead)
pslWriteHead(dest);
for (i=0; i<sourceCount; ++i)
{
source = sources[i];
if (!fileExists(source))
{
warn("%s doesn't exist!", source);
continue;
}
verbose(1, "Lifting %s\n", source);
lf = pslFileOpenWithMeta(source, dest);
for (;;)
{
if (isExtended)
{
xa = xAliNext(lf);
psl = (struct psl *)xa;
}
else
psl = pslNext(lf);
if (psl == NULL)
break;
boolean isProt = pslIsProtein(psl);
doDots(&dotMod);
if (querySide)
seqName = psl->qName;
else
seqName = psl->tName;
spec = findLift(liftHash, seqName, lf);
if (spec == NULL)
{
if (how != carryMissing)
{
freePslOrXa(psl, isExtended);
continue;
}
}
else
{
offset = spec->offset;
blockSizes = psl->blockSizes;
if (querySide)
{
if (!isPtoG)
{
cantHandleSpecRevStrand(spec);
psl->qStart += offset;
psl->qEnd += offset;
}
else
{
psl->match *= 3;
psl->misMatch *= 3;
if (spec->strand == '-')
{
int tmp = psl->qEnd;
psl->qEnd = psl->qStart;
psl->qStart = tmp;
psl->qStart *= -3;
psl->qEnd *= -3;
psl->qStart += offset;
psl->qEnd += offset;
}
else if (spec->strand == '+')
{
psl->qStart *= 3;
psl->qStart += offset;
psl->qEnd *= 3;
psl->qEnd += offset;
}
}
starts = psl->qStarts;
seqSize = psl->qSize;
}
else
{
if (spec->strand == '-')
reverseIntRange(&psl->tStart, &psl->tEnd, psl->tSize);
psl->tStart += offset;
psl->tEnd += offset;
starts = psl->tStarts;
seqSize = psl->tSize;
}
blockCount = psl->blockCount;
if (isPtoG && (spec->strand == '-'))
{
psl->strand[strandChar] = spec->strand;
for (j=0; j<blockCount; ++j)
{
starts[j] *= -3;
starts[j] += offset;
starts[j] = spec->newSize - starts[j];
}
}
else if (isPtoG && (spec->strand == '+'))
{
psl->strand[strandChar] = spec->strand;
for (j=0; j<blockCount; ++j)
{
starts[j] *= 3;
starts[j] += offset;
}
}
else /* mRNA case. */
{
if (spec->strand == '+')
{
if (psl->strand[strandChar] == '-')
{
for (j=0; j<blockCount; ++j)
{
int tr = seqSize - starts[j];
tr += offset;
starts[j] = spec->newSize - tr;
}
}
else
{
for (j=0; j<blockCount; ++j)
starts[j] += offset;
}
}
else
{
if (isProt)
{
/* if it's protein, we can't reverse the query */
if (psl->strand[strandChar] == '-')
{
for (j=0; j<blockCount; ++j)
starts[j] += offset;
}
else
{
for (j=0; j<blockCount; ++j)
{
int tr = seqSize - starts[j];
tr += offset;
starts[j] = spec->newSize - tr;
}
}
psl->strand[strandChar] =
flipStrand(psl->strand[strandChar]);
}
else
{
if (psl->strand[strandChar] == '-')
errAbort("Can't handle all these minus strands! line %d",lf->lineIx);
else
{
for (j=0; j<blockCount; ++j)
{
psl->tStarts[j] = psl->tSize -
(psl->tStarts[j] + blockSizes[j]) + offset;
psl->qStarts[j] = psl->qSize -
(psl->qStarts[j] + blockSizes[j]); /* no offset. */
}
psl->strand[1-strandChar] =
flipStrand(psl->strand[1-strandChar]);
reverseUnsigned(blockSizes, blockCount);
reverseUnsigned(psl->qStarts, blockCount);
reverseUnsigned(psl->tStarts, blockCount);
}
}
}
}
if (isPtoG)
for (j=0; j<blockCount; ++j)
blockSizes[j] *= 3;
if (querySide)
{
psl->qSize = spec->newSize;
psl->qName = spec->newName;
}
else
{
psl->tSize = spec->newSize;
psl->tName = spec->newName;
}
}
if (isExtended)
{
xAliTabOut(xa, dest);
}
else
{
pslTabOut(psl, dest);
}
if (querySide)
psl->qName = seqName;
else
psl->tName = seqName;
freePslOrXa(psl, isExtended);
}
lineFileClose(&lf);
if (dots)
verbose(1, "\n");
}
if (ferror(dest))
errAbort("error writing %s", destFile);
fclose(dest);
}
static void pslHead(struct gfOutput *out, FILE *f)
/* Write out psl head */
{
pslWriteHead(f);
}
int main(int argc, char *argv[]) {
/*
* Arguments/options
*/
char outputFile[50];
char inputFile[50];
char query[100];
char target[100];
///////////////////////////////////////////////////////////////////////////
// (0) Parse the inputs handed by genomeCactus.py / setup stuff.
///////////////////////////////////////////////////////////////////////////
while(1) {
static struct option long_options[] = {
{ "query", required_argument, 0, 'q' },
{ "target", required_argument, 0, 't' },
{ "outputFile", required_argument, 0, 'o' },
{ "inputFile", required_argument, 0, 'i' },
{ "help", no_argument, 0, 'h' },
{ 0, 0, 0, 0 }
};
int option_index = 0;
int key = getopt_long(argc, argv, "i:o:q:t:h", long_options, &option_index);
if(key == -1) {
break;
}
switch(key) {
case 'i':
strcpy(inputFile, optarg);
break;
case 'o':
strcpy(outputFile, optarg);
break;
case 'q':
strcpy(query, optarg);
break;
case 't':
strcpy(target, optarg);
break;
case 'h':
usage();
return 0;
default:
usage();
return 1;
}
}
///////////////////////////////////////////////////////////////////////////
// (0) Check the inputs.
///////////////////////////////////////////////////////////////////////////
assert(outputFile != NULL);
assert(query != NULL);
assert(target != NULL);
FILE *fileHandle = fopen(outputFile, "w");
pslWriteHead(fileHandle);
struct psl *pslList = pslLoadAll(inputFile);
mapPSLs(pslList, fileHandle, query, target);
fclose(fileHandle);
return 0;
}
void pslSort2(char *outDir, char *tempDir, boolean noHead)
/* Do second step of sort - merge all sorted files in tempDir
* to final outdir. */
{
char fileName[512];
struct slName *tmpList, *tmp;
struct midFile *midList = NULL, *mid;
int aliCount = 0;
FILE *f = NULL;
char lastTargetAcc[256];
char targetAcc[256];
strcpy(lastTargetAcc, "");
tmpList = listDir(tempDir, "tmp*.psl");
if (tmpList == NULL)
errAbort("No tmp*.psl files in %s\n", tempDir);
for (tmp = tmpList; tmp != NULL; tmp = tmp->next)
{
sprintf(fileName, "%s/%s", tempDir, tmp->name);
AllocVar(mid);
mid->lf = pslFileOpen(fileName);
slAddHead(&midList, mid);
}
printf("writing %s", outDir);
fflush(stdout);
/* Write out the lowest sorting line from mid list until done. */
for (;;)
{
struct midFile *bestMid = NULL;
if ( (++aliCount & 0xffff) == 0)
{
printf(".");
fflush(stdout);
}
for (mid = midList; mid != NULL; mid = mid->next)
{
if (mid->lf != NULL && mid->psl == NULL)
{
if ((mid->psl = nextPsl(mid->lf)) == NULL)
lineFileClose(&mid->lf);
}
if (mid->psl != NULL)
{
if (bestMid == NULL || pslCmpTarget(&mid->psl, &bestMid->psl) < 0)
bestMid = mid;
}
}
if (bestMid == NULL)
break;
getTargetAcc(bestMid->psl->tName, targetAcc);
if (!sameString(targetAcc, lastTargetAcc))
{
strcpy(lastTargetAcc, targetAcc);
carefulClose(&f);
sprintf(fileName, "%s/%s.psl", outDir, targetAcc);
f = mustOpen(fileName, "w");
if (!noHead)
pslWriteHead(f);
}
pslTabOut(bestMid->psl, f);
pslFree(&bestMid->psl);
}
carefulClose(&f);
printf("\n");
printf("Cleaning up temp files\n");
for (tmp = tmpList; tmp != NULL; tmp = tmp->next)
{
sprintf(fileName, "%s/%s", tempDir, tmp->name);
remove(fileName);
}
}
void pslSort(char *command, char *outFile, char *tempDir, char *inDirs[], int inDirCount)
/* Do the two step sort. */
{
int i;
struct slName *fileList = NULL, *name;
char *inDir;
struct slName *dirDir, *dirFile;
char fileName[512];
int fileCount;
int totalFilesProcessed = 0;
int filesPerMidFile;
int midFileCount = 0;
FILE *f;
struct lineFile *lf;
boolean doReflect = FALSE;
boolean suppressSelf = FALSE;
boolean firstOnly = endsWith(command, "1");
boolean secondOnly = endsWith(command, "2");
if (startsWith("dirs", command))
;
else if (startsWith("g2g", command))
{
doReflect = TRUE;
suppressSelf = TRUE;
}
else
usage();
if (!secondOnly)
{
makeDir(tempDir);
/* Figure out how many files to process. */
for (i=0; i<inDirCount; ++i)
{
inDir = inDirs[i];
dirDir = listDir(inDir, "*.psl");
if (slCount(dirDir) == 0)
dirDir = listDir(inDir, "*.psl.gz");
if (slCount(dirDir) == 0)
errAbort("No psl files in %s\n", inDir);
verbose(1, "%s with %d files\n", inDir, slCount(dirDir));
for (dirFile = dirDir; dirFile != NULL; dirFile = dirFile->next)
{
sprintf(fileName, "%s/%s", inDir, dirFile->name);
name = newSlName(fileName);
slAddHead(&fileList, name);
}
slFreeList(&dirDir);
}
verbose(1, "%d files in %d dirs\n", slCount(fileList), inDirCount);
slReverse(&fileList);
fileCount = slCount(fileList);
filesPerMidFile = round(sqrt(fileCount));
// if (filesPerMidFile > 20)
// filesPerMidFile = 20; /* bandaide! Should keep track of mem usage. */
verbose(1, "Got %d files %d files per mid file\n", fileCount, filesPerMidFile);
/* Read in files a group at a time, sort, and write merged, sorted
* output of one group. */
name = fileList;
while (totalFilesProcessed < fileCount)
{
int filesInMidFile = 0;
struct psl *pslList = NULL, *psl;
int lfileCount = 0;
struct lm *lm = lmInit(256*1024);
for (filesInMidFile = 0; filesInMidFile < filesPerMidFile && name != NULL;
++filesInMidFile, ++totalFilesProcessed, name = name->next)
{
boolean reflectMe = FALSE;
if (doReflect)
{
reflectMe = !selfFile(name->name);
}
verbose(2, "Reading %s (%d of %d)\n", name->name, totalFilesProcessed+1, fileCount);
lf = pslFileOpen(name->name);
while ((psl = nextLmPsl(lf, lm)) != NULL)
{
if (psl->qStart == psl->tStart && psl->strand[0] == '+' &&
suppressSelf && sameString(psl->qName, psl->tName))
{
continue;
}
++lfileCount;
slAddHead(&pslList, psl);
if (reflectMe)
{
psl = mirrorLmPsl(psl, lm);
slAddHead(&pslList, psl);
}
}
lineFileClose(&lf);
}
slSort(&pslList, pslCmpQuery);
makeMidName(tempDir, midFileCount, fileName);
verbose(1, "Writing %s\n", fileName);
f = mustOpen(fileName, "w");
if (!nohead)
pslWriteHead(f);
for (psl = pslList; psl != NULL; psl = psl->next)
{
pslTabOut(psl, f);
}
fclose(f);
pslList = NULL;
lmCleanup(&lm);
verbose(2, "lfileCount %d\n", lfileCount);
++midFileCount;
}
}
if (!firstOnly)
pslSort2(outFile, tempDir);
}
void pslSort2(char *outFile, char *tempDir)
/* Do second step of sort - merge all sorted files in tempDir
* to final. */
{
char fileName[512];
struct slName *tmpList, *tmp;
struct midFile *midList = NULL, *mid;
int aliCount = 0;
FILE *f = mustOpen(outFile, "w");
if (!nohead)
pslWriteHead(f);
tmpList = listDir(tempDir, "tmp*.psl");
if (tmpList == NULL)
errAbort("No tmp*.psl files in %s\n", tempDir);
for (tmp = tmpList; tmp != NULL; tmp = tmp->next)
{
sprintf(fileName, "%s/%s", tempDir, tmp->name);
AllocVar(mid);
mid->lf = pslFileOpen(fileName);
slAddHead(&midList, mid);
}
verbose(1, "writing %s", outFile);
fflush(stdout);
/* Write out the lowest sorting line from mid list until done. */
for (;;)
{
struct midFile *bestMid = NULL;
if ( (++aliCount & 0xffff) == 0)
{
verboseDot();
fflush(stdout);
}
for (mid = midList; mid != NULL; mid = mid->next)
{
if (mid->lf != NULL && mid->psl == NULL)
{
if ((mid->psl = nextPsl(mid->lf)) == NULL)
lineFileClose(&mid->lf);
}
if (mid->psl != NULL)
{
if (bestMid == NULL || pslCmpQuery(&mid->psl, &bestMid->psl) < 0)
bestMid = mid;
}
}
if (bestMid == NULL)
break;
pslTabOut(bestMid->psl, f);
pslFree(&bestMid->psl);
}
printf("\n");
fclose(f);
/* The followint really shouldn't be necessary.... */
for (mid = midList; mid != NULL; mid = mid->next)
lineFileClose(&mid->lf);
printf("Cleaning up temp files\n");
for (tmp = tmpList; tmp != NULL; tmp = tmp->next)
{
sprintf(fileName, "%s/%s", tempDir, tmp->name);
remove(fileName);
}
}