void gbAlignInstall(struct gbSelect* select, struct gbSelect* prevSelect)
/* Install alignments, optionally migrating unchanged ones from a previous
* release. This does one update, accPrefix and either native or xeno */
{
char nativeAlignIdx[PATH_LEN], xenoAlignIdx[PATH_LEN];
struct gbAlignInfo alignInfo;
gbVerbEnter(1, "gbAlignInstall: %s", gbSelectDesc(select));
/* load required entry date */
gbReleaseLoadProcessed(select);
if (prevSelect != NULL)
{
gbReleaseLoadProcessed(prevSelect);
gbReleaseLoadAligned(prevSelect);
}
/* mark entries and updates to migrate or align */
alignInfo = gbAlignFindNeedAligned(select, prevSelect);
/* Process each category */
if (select->orgCats & GB_NATIVE)
installOrgCatAligned(select, GB_NATIVE, prevSelect, &alignInfo,
nativeAlignIdx);
if (select->orgCats & GB_XENO)
installOrgCatAligned(select, GB_XENO, prevSelect, &alignInfo,
xenoAlignIdx);
/* now indices can be renamed, not completely atomic, but good enough */
if (select->orgCats & GB_NATIVE)
gbOutputRename(nativeAlignIdx, NULL);
if (select->orgCats & GB_XENO)
gbOutputRename(xenoAlignIdx, NULL);
/* print message before memory is freed */
gbVerbLeave(1, "gbAlignInstall: %s", gbSelectDesc(select));
/* unload entries to free memory */
gbReleaseUnload(select->release);
if (prevSelect != NULL)
gbReleaseUnload(prevSelect->release);
}
void copyRefSeqPepFa(struct gbUpdate* update,
char* outDir, char *gbFile)
/* copy a subset of the RefSeq peptide file for the select genes */
{
struct gbRelease* release = update->release;
char faInPath[PATH_LEN];
char faOutPath[PATH_LEN];
struct lineFile* inLf;
boolean copying = FALSE;
FILE* outFh;
char* line;
/* change the .gbff.Z suffix to .fsa.Z */
if (!endsWith(gbFile, ".gbff.Z"))
errAbort("expected a file ending in .gbff.Z, got: %s", gbFile);
strcpy(faInPath, gbFile);
faInPath[strlen(faInPath)-7] = '\0';
strcat(faInPath, ".fsa.Z");
strcpy(faOutPath, outDir);
strcat(faOutPath, "/");
strcat(faOutPath, faInPath);
verbose(1, "copying from %s\n", faInPath);
/* copy selected, don't bother with fa readers */
inLf = gzLineFileOpen(faInPath);
outFh = gbMustOpenOutput(faOutPath);
while (lineFileNext(inLf, &line, NULL))
{
if (line[0] == '>')
{
char *geneAcc = parsePepGeneAcc(line);
struct gbEntry* entry = NULL;
if (geneAcc != NULL)
entry = gbReleaseFindEntry(release, geneAcc);
copying = ((entry != NULL) && (entry->selectVer > 0));
verbose(2, "acc for pep: %s: %s\n", geneAcc,
(copying ? "yes" : "no"));
}
if (copying)
{
fputs(line, outFh);
fputc('\n', outFh);
if (ferror(outFh))
errnoAbort("write failed: %s: ", faOutPath);
}
}
gbOutputRename(faOutPath, &outFh);
gzLineFileClose(&inLf);
}
void markAligns(struct gbSelect* select, unsigned orgCat)
/* create a file indicating that sequences either needs aligned or migated for
* this for this partation. This is used to determine what needs to be
* installed after the alignment. This is needed because they might be all
* be migrate, so that fasta can't be the indicator. */
{
char path[PATH_LEN];
FILE* fh;
unsigned orgCatsHold = select->orgCats;
select->orgCats = orgCat;
gbAlignedGetPath(select, "aligns", workDir, path);
fh = gbMustOpenOutput(path);
gbOutputRename(path, &fh);
select->orgCats = orgCatsHold;
}
void gbFaClose(struct gbFa **faPtr)
/* close a fasta file, check for any undetected I/O errors. */
{
struct gbFa *fa = *faPtr;
if (fa != NULL)
{
if (ferror(fa->fh))
errAbort("%s error on %s", ((fa->mode[0] != 'r') ? "write" : "read"),
fa->fileName);
if (fa->mode[0] == 'w')
gbOutputRename(fa->fileName, &fa->fh);
else
gzClose(&fa->fh);
freez(&fa->headerBuf);
freez(&fa->seqBuf);
freez(&fa->fhBuf);
freez(faPtr); /* NULLs var */
}
}
void installOrgCatAligned(struct gbSelect* select, unsigned orgCat,
struct gbSelect* prevSelect,
struct gbAlignInfo* alignInfo,
char* alignIdx)
/* Install alignments for either native or xeno. The alignment index is
* created and named returned, but not renamed until both native and xeno are
* processed. */
{
unsigned holdOrgCats = select->orgCats;
struct outputFiles out;
struct recCounts recCounts;
ZeroVar(&out);
ZeroVar(&recCounts);
select->orgCats = orgCat;
if (prevSelect != NULL)
prevSelect->orgCats = orgCat;
/* setup out PSL and orientInfo files */
gbAlignedGetPath(select, "psl.gz", NULL, out.psl.path);
out.psl.fh = openSortOutput(out.psl.path, PSL_SORT_SPEC);
if (select->orgCats == GB_NATIVE)
{
gbAlignedGetPath(select, "oi.gz", NULL, out.oi.path);
out.oi.fh = openSortOutput(out.oi.path, OI_SORT_SPEC);
if (select->type == GB_EST)
{
gbAlignedGetPath(select, "intronPsl.gz", NULL, out.intronPsl.path);
out.intronPsl.fh = openSortOutput(out.intronPsl.path, PSL_SORT_SPEC);
}
}
if (select->type == GB_MRNA)
{
/* we don't bother sorting raw psl */
gbAlignedGetPath(select, "rawPsl.gz", NULL, out.rawPsl.path);
out.rawPsl.fh = gbMustOpenOutput(out.rawPsl.path);
}
/* previous aligned if this is a full update */
if (prevSelect != NULL)
migrateAligned(select, prevSelect, alignInfo, &out, &recCounts);
/* copy currently aligned, if they exist */
copyPsls(select, MAIN_PSL_FILE, out.psl.fh, &recCounts.pslCnts);
if (select->type == GB_MRNA)
copyPsls(select, RAW_PSL_FILE, out.rawPsl.fh, &recCounts.rawPslCnts);
if ((select->orgCats == GB_NATIVE) && (recCounts.pslCnts.recTotalCnt > 0))
{
/* copy new OI and intronPsls */
copyOrientInfos(select, out.oi.fh, &recCounts);
if (select->type == GB_EST)
copyPsls(select, INTRON_PSL_FILE, out.intronPsl.fh,
&recCounts.intronPslCnts);
}
/* Install or remove files. Done seperate from copy due to posibility of
* all being migrated*/
if (recCounts.intronPslCnts.recTotalCnt > 0)
gbOutputRename(out.intronPsl.path, &out.intronPsl.fh);
else
gbOutputRemove(out.intronPsl.path, &out.intronPsl.fh);
if (recCounts.oiCnts.recTotalCnt > 0)
gbOutputRename(out.oi.path, &out.oi.fh);
else
gbOutputRemove(out.oi.path, &out.oi.fh);
if (recCounts.rawPslCnts.recTotalCnt > 0)
gbOutputRename(out.rawPsl.path, &out.rawPsl.fh);
else
gbOutputRemove(out.rawPsl.path, &out.rawPsl.fh);
if (recCounts.pslCnts.recTotalCnt > 0)
gbOutputRename(out.psl.path, &out.psl.fh);
else
gbOutputRemove(out.psl.path, &out.psl.fh);
createAlignedIndex(select, alignIdx);
select->orgCats = holdOrgCats;
if (prevSelect != NULL)
prevSelect->orgCats = holdOrgCats;
}
void extractAccFromGb(char *inName, char* outName, struct hash *accTbl)
/* Parse records of genBank file and print ones that match accession names.
* (yanked from gbOneAcc, changed to use stdio so we can access compressed). */
{
enum {maxHeadLines=20, headLineSize=256 };
char *headLines[maxHeadLines]; /* Store stuff between locus and accession. */
char line[headLineSize];
FILE *inFh;
FILE *outFh = NULL;
int lineNum = 0;
int i;
char* acc;
verbose(1, "copying from %s\n", inName);
inFh = gzMustOpen(inName, "r");
for (i=0; i<maxHeadLines; ++i)
headLines[i] = needMem(headLineSize);
while (TRUE)
{
boolean gotAcc = FALSE;
boolean gotMyAcc = FALSE;
int headLineCount = 0;
/* Seek to LOCUS */
for (;;)
{
if (!readData(inFh, inName, line, headLineSize, FALSE))
break;
lineNum++;
if (startsWith("LOCUS", line))
break;
}
if (feof(inFh))
break;
for (i=0; i<maxHeadLines; ++i)
{
++headLineCount;
strcpy(headLines[i], line);
readData(inFh, inName, line, headLineSize, TRUE);
lineNum++;
if (startsWith("ACCESSION", line))
{
gotAcc = TRUE;
break;
}
}
if (!gotAcc)
errAbort("LOCUS without ACCESSION in %d lines at line %d of %s",
maxHeadLines, lineNum, inName);
acc = lastWordInLine(line);
gotMyAcc = (hashLookup(accTbl, acc) != NULL);
if (gotMyAcc)
{
if (outFh == NULL)
outFh = gbMustOpenOutput(outName);
for (i=0; i<headLineCount; ++i)
{
fputs(headLines[i], outFh);
fputc('\n', outFh);
}
fputs(line, outFh);
fputc('\n', outFh);
}
for (;;)
{
readData(inFh, inName, line, headLineSize, TRUE);
lineNum++;
if (gotMyAcc)
{
fputs(line, outFh);
fputc('\n', outFh);
}
if (startsWith("//", line))
break;
}
if ((outFh != NULL) && ferror(outFh))
break; /* write error */
}
if (outFh != NULL)
gbOutputRename(outName, &outFh);
gzClose(&inFh);
}
