struct bed *bamGetFilteredBedsOnRegions(struct sqlConnection *conn,
char *db, char *table, struct region *regionList, struct lm *lm,
int *retFieldCount)
/* Get list of beds from BAM, in all regions, that pass filtering. */
{
int maxOut = bigFileMaxOutput();
/* Figure out bam file name get column info and filter. */
struct asObject *as = bamAsObj();
struct asFilter *filter = asFilterFromCart(cart, db, table, as);
struct hash *idHash = identifierHash(db, table);
/* Get beds a region at a time. */
struct bed *bedList = NULL;
struct region *region;
for (region = regionList; region != NULL; region = region->next)
{
char *fileName = bamFileName(table, conn, region->chrom);
addFilteredBedsOnRegion(fileName, region, table, filter, lm, &bedList, idHash, &maxOut);
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());
break;
}
}
slReverse(&bedList);
return bedList;
}
struct wigAsciiData *getWiggleAsData(struct sqlConnection *conn, char *table,
struct region *region)
/* return the wigAsciiData list */
{
int maxOut = 0;
struct wigAsciiData *data = NULL;
maxOut = bigFileMaxOutput();
// ignore return outCount from wigOutRegion:
wigOutRegion(table, conn, region, maxOut, wigDataNoPrint, &data, 0);
return data;
}
struct bed *getWiggleAsBed(
char *db, char *table, /* Database and table. */
struct region *region, /* Region to get data for. */
char *filter, /* Filter to add to SQL where clause if any. */
struct hash *idHash, /* Restrict to id's in this hash if non-NULL. */
struct lm *lm, /* Where to allocate memory. */
struct sqlConnection *conn) /* SQL connection to work with */
/* Return a bed list of all items in the given range in table.
* Cleanup result via lmCleanup(&lm) rather than bedFreeList. */
/* filter, idHash and lm are currently unused, perhaps future use */
{
struct bed *bedList=NULL;
char splitTableOrFileName[HDB_MAX_TABLE_STRING];
struct customTrack *ct = NULL;
boolean isCustom = FALSE;
boolean hasConstraint = FALSE;
struct wiggleDataStream *wds = NULL;
unsigned long long valuesMatched = 0;
int operations = wigFetchBed;
char *dataConstraint;
double ll = 0.0;
double ul = 0.0;
char *table2 = NULL;
struct bed *intersectBedList = NULL;
int maxOut;
WIG_INIT; /* ct, isCustom, hasConstraint, wds and table2 are set here */
if (hasConstraint)
freeMem(dataConstraint); /* been cloned into wds */
maxOut = bigFileMaxOutput();
wds->setMaxOutput(wds, maxOut);
wds->setChromConstraint(wds, region->chrom);
wds->setPositionConstraint(wds, region->start, region->end);
if (table2)
intersectBedList = bedTable2(conn, region, table2);
if (isCustom)
{
if (ct->dbTrack)
{
unsigned span = 0;
struct sqlConnection *trashConn = hAllocConn(CUSTOM_TRASH);
struct trackDb *tdb = findTdbForTable(database, curTrack, table, ctLookupName);
valuesMatched = getWigglePossibleIntersection(wds, region,
CUSTOM_TRASH, table2, &intersectBedList,
splitTableOrFileName, operations);
span = minSpan(trashConn, splitTableOrFileName, region->chrom,
region->start, region->end, cart, tdb);
wds->setSpanConstraint(wds, span);
hFreeConn(&trashConn);
}
else
valuesMatched = getWigglePossibleIntersection(wds, region, NULL, table2,
&intersectBedList, splitTableOrFileName, operations);
}
else
{
if (conn == NULL)
errAbort( "getWiggleAsBed: NULL conn given for database table");
if (hFindSplitTable(database, region->chrom, table, splitTableOrFileName, sizeof splitTableOrFileName, NULL))
{
struct trackDb *tdb = findTdbForTable(database, curTrack, table, ctLookupName);
unsigned span = 0;
/* XXX TBD, watch for a span limit coming in as an SQL filter */
span = minSpan(conn, splitTableOrFileName, region->chrom,
region->start, region->end, cart, tdb);
wds->setSpanConstraint(wds, span);
valuesMatched = getWigglePossibleIntersection(wds, region, database,
table2, &intersectBedList, splitTableOrFileName, operations);
}
}
if (valuesMatched > 0)
{
struct bed *bed;
wds->sortResults(wds);
for (bed = wds->bed; bed != NULL; bed = bed->next)
{
struct bed *copy = lmCloneBed(bed, lm);
slAddHead(&bedList, copy);
}
slReverse(&bedList);
}
wiggleDataStreamFree(&wds);
return bedList;
} /* struct bed *getWiggleAsBed() */
static void doOutWig(struct trackDb *track, char *table, struct sqlConnection *conn,
enum wigOutputType wigOutType)
{
struct region *regionList = getRegions(), *region;
int maxOut = 0, outCount, curOut = 0;
char *shortLabel = table, *longLabel = table;
if (track == NULL)
errAbort("Sorry, can't find necessary track information for %s. "
"If you reached this page by selecting \"All tables\" as the "
"group, please go back and select the same table via a regular "
"track group if possible.",
table);
maxOut = bigFileMaxOutput();
if (cartUsualBoolean(cart, hgtaDoGreatOutput, FALSE))
fputs("#", stdout);
else
textOpen();
if (track != NULL)
{
if (!sameString(track->table, table) && track->subtracks != NULL)
{
struct slRef *tdbRefList = trackDbListGetRefsToDescendantLeaves(track->subtracks);
struct slRef *tdbRef;
for (tdbRef = tdbRefList; tdbRef != NULL; tdbRef = tdbRef->next)
{
struct trackDb *tdb = tdbRef->val;
if (sameString(tdb->table, table))
{
track = tdb;
break;
}
}
slFreeList(&tdbRefList);
}
shortLabel = track->shortLabel;
longLabel = track->longLabel;
}
wigDataHeader(shortLabel, longLabel, NULL, wigOutType);
for (region = regionList; region != NULL; region = region->next)
{
int curMaxOut = maxOut - curOut;
if (anySubtrackMerge(database, table))
outCount = mergedWigOutRegion(table, conn, region, curMaxOut,
wigOutType);
else if (startsWithWord("bedGraph", track->type))
outCount = bedGraphOutRegion(table, conn, region, curMaxOut,
wigOutType);
else if (startsWithWord("mathWig", track->type))
outCount = mathWigOutRegion(track, table, conn, region, curMaxOut, wigOutType);
else if (startsWithWord("bigWig", track->type))
outCount = bigWigOutRegion(table, conn, region, curMaxOut, wigOutType);
else
outCount = wigOutRegion(table, conn, region, curMaxOut,
wigOutType, NULL, 0);
curOut += outCount;
if (curOut >= maxOut)
break;
}
if (curOut >= maxOut)
errAbort("Reached output limit of %d data values, please make region smaller,\n\tor set a higher output line limit with the filter settings.", curOut);
}
void bamTabOut(char *db, char *table, struct sqlConnection *conn, char *fields, FILE *f)
/* Print out selected fields from BAM. 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 = bamGetFields();
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]);
}
/* Output row of labels */
fprintf(f, "#%s", fieldArray[0]);
for (i=1; i<fieldCount; ++i)
fprintf(f, "\t%s", fieldArray[i]);
fprintf(f, "\n");
struct asObject *as = bamAsObj();
struct asFilter *filter = NULL;
if (anyFilter())
{
filter = asFilterFromCart(cart, db, table, as);
if (filter)
{
fprintf(f, "# Filtering on %d columns\n", slCount(filter->columnList));
}
}
/* Loop through outputting each region */
struct region *region, *regionList = getRegions();
int maxOut = bigFileMaxOutput();
for (region = regionList; region != NULL && (maxOut > 0); region = region->next)
{
struct lm *lm = lmInit(0);
char *fileName = bamFileName(table, conn, region->chrom);
struct samAlignment *sam, *samList = bamFetchSamAlignment(fileName, region->chrom,
region->start, region->end, lm);
char *row[SAMALIGNMENT_NUM_COLS];
char numBuf[BAM_NUM_BUF_SIZE];
for (sam = samList; sam != NULL && (maxOut > 0); sam = sam->next)
{
samAlignmentToRow(sam, numBuf, row);
if (asFilterOnRow(filter, row))
{
/* if we're looking for identifiers, check if this matches */
if ((idHash != NULL)&&(hashLookup(idHash, row[idFieldNum]) == NULL))
continue;
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 --;
}
}
freeMem(fileName);
lmCleanup(&lm);
}
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. */
hashFree(&fieldHash);
freeMem(fieldArray);
freeMem(columnArray);
}
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);
}
