static void asvClose(struct annoStreamer **pVSelf)
/* Close VCF file and free self. */
{
if (pVSelf == NULL)
return;
struct annoStreamVcf *self = *(struct annoStreamVcf **)pVSelf;
vcfFileFree(&(self->vcff));
// Don't free self->record -- currently it belongs to vcff's localMem
dyStringFree(&(self->dyGt));
annoStreamerFree(pVSelf);
}
static void addFilteredBedsOnRegion(char *fileName, struct region *region, char *table,
struct asFilter *filter, struct lm *bedLm,
struct bed **pBedList, struct hash *idHash, int *pMaxOut,
boolean isTabix)
/* Add relevant beds in reverse order to pBedList */
{
struct vcfFile *vcff;
if (isTabix)
vcff = vcfTabixFileMayOpen(fileName, region->chrom, region->start, region->end,
100, *pMaxOut);
else
vcff = vcfFileMayOpen(fileName, region->chrom, region->start, region->end,
100, *pMaxOut, TRUE);
if (vcff == NULL)
noWarnAbort();
struct lm *lm = lmInit(0);
char *row[VCFDATALINE_NUM_COLS];
char numBuf[VCF_NUM_BUF_SIZE];
// Temporary storage for row-ification:
struct dyString *dyAlt = newDyString(1024);
struct dyString *dyFilter = newDyString(1024);
struct dyString *dyInfo = newDyString(1024);
struct dyString *dyGt = newDyString(1024);
struct vcfRecord *rec;
for (rec = vcff->records; rec != NULL; rec = rec->next)
{
vcfRecordToRow(rec, region->chrom, numBuf, dyAlt, dyFilter, dyInfo, dyGt, row);
if (asFilterOnRow(filter, row))
{
if ((idHash != NULL) && (hashLookup(idHash, rec->name) == NULL))
continue;
struct bed *bed;
lmAllocVar(bedLm, bed);
bed->chrom = lmCloneString(bedLm, region->chrom);
bed->chromStart = rec->chromStart;
bed->chromEnd = rec->chromEnd;
bed->name = lmCloneString(bedLm, rec->name);
slAddHead(pBedList, bed);
}
(*pMaxOut)--;
if (*pMaxOut <= 0)
break;
}
dyStringFree(&dyAlt); dyStringFree(&dyFilter); dyStringFree(&dyInfo); dyStringFree(&dyGt);
lmCleanup(&lm);
vcfFileFree(&vcff);
}
void vcfTabOut(char *db, char *table, struct sqlConnection *conn, char *fields, FILE *f,
boolean isTabix)
/* Print out selected fields from VCF. If fields is NULL, then print out all fields. */
{
struct hTableInfo *hti = NULL;
hti = getHti(db, table, conn);
struct hash *idHash = NULL;
char *idField = getIdField(db, curTrack, table, hti);
int idFieldNum = 0;
/* if we know what field to use for the identifiers, get the hash of names */
if (idField != NULL)
idHash = identifierHash(db, table);
if (f == NULL)
f = stdout;
/* Convert comma separated list of fields to array. */
int fieldCount = chopByChar(fields, ',', NULL, 0);
char **fieldArray;
AllocArray(fieldArray, fieldCount);
chopByChar(fields, ',', fieldArray, fieldCount);
/* Get list of all fields in big bed and turn it into a hash of column indexes keyed by
* column name. */
struct hash *fieldHash = hashNew(0);
struct slName *bb, *bbList = vcfGetFields();
int i;
for (bb = bbList, i=0; bb != NULL; bb = bb->next, ++i)
{
/* if we know the field for identifiers, save it away */
if ((idField != NULL) && sameString(idField, bb->name))
idFieldNum = i;
hashAddInt(fieldHash, bb->name, i);
}
/* Create an array of column indexes corresponding to the selected field list. */
int *columnArray;
AllocArray(columnArray, fieldCount);
for (i=0; i<fieldCount; ++i)
{
columnArray[i] = hashIntVal(fieldHash, fieldArray[i]);
}
// If we are outputting a subset of fields, invalidate the VCF header.
boolean allFields = (fieldCount == VCFDATALINE_NUM_COLS);
if (!allFields)
fprintf(f, "# Only selected columns are included below; output is not valid VCF.\n");
struct asObject *as = vcfAsObj();
struct asFilter *filter = NULL;
if (anyFilter())
filter = asFilterFromCart(cart, db, table, as);
/* Loop through outputting each region */
struct region *region, *regionList = getRegions();
int maxOut = bigFileMaxOutput();
struct trackDb *tdb = hashFindVal(fullTableToTdbHash, table);
// Include the header, absolutely necessary for VCF parsing.
boolean printedHeader = FALSE;
// Temporary storage for row-ification:
struct dyString *dyAlt = newDyString(1024);
struct dyString *dyFilter = newDyString(1024);
struct dyString *dyInfo = newDyString(1024);
struct dyString *dyGt = newDyString(1024);
struct vcfRecord *rec;
for (region = regionList; region != NULL && (maxOut > 0); region = region->next)
{
char *fileName = vcfFileName(tdb, conn, table, region->chrom);
struct vcfFile *vcff;
if (isTabix)
vcff = vcfTabixFileMayOpen(fileName, region->chrom, region->start, region->end,
100, maxOut);
else
vcff = vcfFileMayOpen(fileName, region->chrom, region->start, region->end,
100, maxOut, TRUE);
if (vcff == NULL)
noWarnAbort();
// If we are outputting all fields, but this VCF has no genotype info, omit the
// genotype columns from output:
if (allFields && vcff->genotypeCount == 0)
fieldCount = VCFDATALINE_NUM_COLS - 2;
if (!printedHeader)
{
fprintf(f, "%s", vcff->headerString);
if (filter)
fprintf(f, "# Filtering on %d columns\n", slCount(filter->columnList));
if (!allFields)
{
fprintf(f, "#%s", fieldArray[0]);
for (i=1; i<fieldCount; ++i)
fprintf(f, "\t%s", fieldArray[i]);
fprintf(f, "\n");
}
printedHeader = TRUE;
}
char *row[VCFDATALINE_NUM_COLS];
char numBuf[VCF_NUM_BUF_SIZE];
for (rec = vcff->records; rec != NULL && (maxOut > 0); rec = rec->next)
{
vcfRecordToRow(rec, region->chrom, numBuf, dyAlt, dyFilter, dyInfo, dyGt, row);
if (asFilterOnRow(filter, row))
{
/* if we're looking for identifiers, check if this matches */
if ((idHash != NULL) && (hashLookup(idHash, row[idFieldNum]) == NULL))
continue;
// All fields output: after asFilter'ing, preserve original VCF chrom
if (allFields && !sameString(rec->chrom, region->chrom))
row[0] = rec->chrom;
int i;
fprintf(f, "%s", row[columnArray[0]]);
for (i=1; i<fieldCount; ++i)
{
fprintf(f, "\t%s", row[columnArray[i]]);
}
fprintf(f, "\n");
maxOut --;
}
}
vcfFileFree(&vcff);
freeMem(fileName);
}
if (maxOut == 0)
warn("Reached output limit of %d data values, please make region smaller,\n\tor set a higher output line limit with the filter settings.", bigFileMaxOutput());
/* Clean up and exit. */
dyStringFree(&dyAlt); dyStringFree(&dyFilter); dyStringFree(&dyInfo); dyStringFree(&dyGt);
hashFree(&fieldHash);
freeMem(fieldArray);
freeMem(columnArray);
}
