void doPrintSelectedFields()
/* Actually produce selected field output as text stream. */
{
char *db = cartString(cart, hgtaDatabase);
char *table = cartString(cart, hgtaTable);
char *varPrefix = checkVarPrefix();
int varPrefixSize = strlen(varPrefix);
struct hashEl *varList = NULL, *var;
struct slName *fieldList = NULL, *field;
textOpen();
/* Gather together field list for primary and linked tables from cart. */
varList = cartFindPrefix(cart, varPrefix);
for (var = varList; var != NULL; var = var->next)
{
if (!sameString(var->val, "0"))
{
field = slNameNew(var->name + varPrefixSize);
if (primaryOrLinked(field->name))
slAddHead(&fieldList, field);
}
}
if (fieldList == NULL)
errAbort("Please go back and select at least one field");
slReverse(&fieldList);
/* Do output. */
tabOutSelectedFields(db, table, NULL, fieldList);
/* Clean up. */
slFreeList(&fieldList);
hashElFreeList(&varList);
}
void doGenePredPal(struct sqlConnection *conn)
/* Output genePred protein alignment. */
{
if (doGalaxy() && !cgiOptionalString(hgtaDoGalaxyQuery))
{
sendParamsToGalaxy(hgtaDoPalOut, "submit");
return;
}
/* get rid of pesky cookies that would bring us back here */
cartRemove(cart, hgtaDoPal);
cartRemove(cart, hgtaDoPalOut);
if (anyIntersection() && intersectionIsBpWise())
errAbort("Can't do CDS FASTA output when bit-wise intersection is on. "
"Please go back and select another output type, or clear the intersection.");
checkNoGenomeDisabled(database, curTable);
struct lm *lm = lmInit(64*1024);
int fieldCount;
struct bed *bedList = cookedBedsOnRegions(conn, curTable, getRegions(),
lm, &fieldCount);
//lmCleanup(&lm);
textOpen();
int outCount = palOutPredsInBeds(conn, cart, bedList, curTable);
/* Do some error diagnostics for user. */
if (outCount == 0)
explainWhyNoResults(NULL);
}
void doTest(struct sqlConnection *conn)
/* Put up a page to see what happens. */
{
char *s = NULL;
textOpen();
hPrintf("Doing test!\n");
s = filterClause("hg18", "knownGene", "chrX", NULL);
if (s != NULL)
hPrintf("%s\n", s);
else
hPrintf("%p\n", s);
}
void doOutGff(char *table, struct sqlConnection *conn, boolean outputGtf)
/* Save as GFF/GTF. */
{
struct hTableInfo *hti = getHti(database, table, conn);
struct bed *bedList;
struct hash *chromHash = NULL;
struct slName *exonFramesList = NULL;
char source[HDB_MAX_TABLE_STRING];
int itemCount;
struct region *region, *regionList = getRegions();
textOpen();
int efIdx = -1;
safef(source, sizeof(source), "%s_%s", database, table);
if (conn)
{
boolean simpleTableExists = sqlTableExists(conn, table);
// simpleTable means not split table, not custom track
// However it still can include bbi table with bam fileName path
if (simpleTableExists) // no tables having exonFrames are split tables anyway
efIdx = sqlFieldIndex(conn, table, "exonFrames");
itemCount = 0;
int regionCount = slCount(regionList);
// regionList can have many thousands of items e.g. rheMac3 has 34000 chroms!
// This regionCount threshold should be just above the # chroms in the latest human assembly
if (simpleTableExists && (regionCount > 500))
{
chromHash = makeChromHashForTable(conn, table);
}
}
// Note: code could be added here to extract exonFrames from bigGenePred
// Process each region
for (region = regionList; region != NULL; region = region->next)
{
if (chromHash && (!hashFindVal(chromHash, region->chrom)))
continue;
struct lm *lm = lmInit(64*1024);
int fieldCount;
bedList = cookedBedList(conn, table, region, lm, &fieldCount);
// Use exonFrames field if available for better accuracy instead of calculating from coordinates
if (efIdx != -1)
exonFramesList = getExonFrames(table, conn, bedList);
itemCount += bedToGffLines(bedList, exonFramesList, hti, fieldCount, source, outputGtf);
lmCleanup(&lm);
}
if (itemCount == 0)
hPrintf(NO_RESULTS);
}
void doGenomicDna(struct sqlConnection *conn)
/* Get genomic sequence (UI has already told us how). */
{
struct region *region, *regionList = getRegions();
struct hTableInfo *hti = getHti(database, curTable, conn);
int fieldCount;
textOpen();
int resultCount = 0;
for (region = regionList; region != NULL; region = region->next)
{
struct lm *lm = lmInit(64*1024);
struct bed *bedList = cookedBedList(conn, curTable, region, lm, &fieldCount);
if (bedList != NULL)
resultCount += hgSeqBed(database, hti, bedList);
lmCleanup(&lm);
}
if (!resultCount)
hPrintf(NO_RESULTS);
}
void doHalMaf(struct trackDb *parentTrack, char *table, struct sqlConnection *conn)
/* Output regions as MAF. maf tables look bed-like enough for
* cookedBedsOnRegions to handle intersections. */
{
#ifdef USE_HAL
struct region *region = NULL, *regionList = getRegions();
struct trackDb *tdb;
if ((tdb = findTrackDb(parentTrack, table)) == NULL)
errAbort("cannot find track named %s under %s\n", table, parentTrack->table);
char *fileName = trackDbSetting(tdb, "bigDataUrl");
char *otherSpecies = trackDbSetting(tdb, "otherSpecies");
int handle = halOpenLOD(fileName);
struct hal_species_t *speciesList = halGetSpecies(handle);
for(; speciesList; speciesList = speciesList->next)
{
if (sameString(speciesList->name, otherSpecies))
break;
}
if (speciesList == NULL)
errAbort("cannot find species %s in hal file %s\n",
otherSpecies, fileName);
speciesList->next = NULL;
textOpen();
for (region = regionList; region != NULL; region = region->next)
{
halGetMAF(stdout, handle, speciesList,
trackHubSkipHubName(database), region->chrom,
region->start, region->end, FALSE);
}
#else // USE_HAL
errAbort("hgTables not compiled with HAL support.");
#endif // USE_HAL
}
boolean doGetBedOrCt(struct sqlConnection *conn, boolean doCt,
boolean doCtFile, boolean redirectToGb)
/* Actually output bed or custom track. Return TRUE unless no results. */
{
char *db = cloneString(database);
char *table = curTable;
struct hTableInfo *hti = getHti(db, table, conn);
struct featureBits *fbList = NULL, *fbPtr;
struct customTrack *ctNew = NULL;
boolean doCtHdr = (cartUsualBoolean(cart, hgtaPrintCustomTrackHeaders, FALSE)
|| doCt || doCtFile);
char *ctWigOutType = cartCgiUsualString(cart, hgtaCtWigOutType, outWigData);
char *fbQual = fbOptionsToQualifier();
char fbTQ[128];
int fields = hTableInfoBedFieldCount(hti);
boolean gotResults = FALSE;
struct region *region, *regionList = getRegions();
boolean isBedGr = isBedGraph(curTable);
boolean isBgWg = isBigWigTable(curTable);
boolean needSubtrackMerge = anySubtrackMerge(database, curTable);
boolean doDataPoints = FALSE;
boolean isWig = isWiggle(database, table);
struct wigAsciiData *wigDataList = NULL;
struct dataVector *dataVectorList = NULL;
boolean doRgb = bedItemRgb(hTrackDbForTrack(db, curTable));
if (!cartUsualBoolean(cart, hgtaDoGreatOutput, FALSE) && !doCt)
{
textOpen();
}
if (cartUsualBoolean(cart, hgtaDoGreatOutput, FALSE))
fputs("#", stdout);
if ((isWig || isBedGr || isBgWg) && sameString(outWigData, ctWigOutType))
doDataPoints = TRUE;
for (region = regionList; region != NULL; region = region->next)
{
struct bed *bedList = NULL, *bed;
struct lm *lm = lmInit(64*1024);
struct dataVector *dv = NULL;
if (isWig && doDataPoints)
{
if (needSubtrackMerge)
{
dv = wiggleDataVector(curTrack, curTable, conn, region);
if (dv != NULL)
slAddHead(&dataVectorList, dv);
}
else
{
int count = 0;
struct wigAsciiData *wigData = NULL;
struct wigAsciiData *asciiData;
struct wigAsciiData *next;
wigData = getWiggleAsData(conn, curTable, region);
for (asciiData = wigData; asciiData; asciiData = next)
{
next = asciiData->next;
if (asciiData->count)
{
slAddHead(&wigDataList, asciiData);
++count;
}
}
slReverse(&wigDataList);
}
}
else if (isBedGr && doDataPoints)
{
dv = bedGraphDataVector(curTable, conn, region);
if (dv != NULL)
slAddHead(&dataVectorList, dv);
}
else if (isBgWg && doDataPoints)
{
dv = bigWigDataVector(curTable, conn, region);
if (dv != NULL)
slAddHead(&dataVectorList, dv);
}
else if (isWig || isBgWg)
{
dv = wiggleDataVector(curTrack, curTable, conn, region);
bedList = dataVectorToBedList(dv);
dataVectorFree(&dv);
}
else if (isBedGr)
{
bedList = getBedGraphAsBed(conn, curTable, region);
}
else
{
bedList = cookedBedList(conn, curTable, region, lm, &fields);
}
/* this is a one-time only initial creation of the custom track
* structure to receive the results. gotResults turns it off after
* the first time.
*/
if (doCtHdr && !gotResults &&
((bedList != NULL) || (wigDataList != NULL) ||
(dataVectorList != NULL)))
{
ctNew = beginCustomTrack(table, fields,
doCt, (isWig || isBedGr || isBgWg), doDataPoints);
}
if (doDataPoints && (wigDataList || dataVectorList))
gotResults = TRUE;
else
{
if ((fbQual == NULL) || (fbQual[0] == 0))
{
for (bed = bedList; bed != NULL; bed = bed->next)
{
if (bed->name != NULL)
{
subChar(bed->name, ' ', '_');
}
if (doCt)
{
struct bed *dupe = cloneBed(bed); /* Out of local memory. */
slAddHead(&ctNew->bedList, dupe);
}
else
{
if (doRgb)
bedTabOutNitemRgb(bed, fields, stdout);
else
bedTabOutN(bed, fields, stdout);
}
gotResults = TRUE;
}
}
else
{
safef(fbTQ, sizeof(fbTQ), "%s:%s", hti->rootName, fbQual);
fbList = fbFromBed(db, fbTQ, hti, bedList, 0, 0, FALSE, FALSE);
if (fields >= 6)
fields = 6;
else if (fields >= 4)
fields = 4;
else
fields = 3;
if (doCt && ctNew)
{
ctNew->fieldCount = fields;
safef(ctNew->tdb->type, strlen(ctNew->tdb->type)+1,
"bed %d", fields);
}
for (fbPtr=fbList; fbPtr != NULL; fbPtr=fbPtr->next)
{
if (fbPtr->name != NULL)
{
char *ptr = strchr(fbPtr->name, ' ');
if (ptr != NULL)
*ptr = 0;
}
if (doCt)
{
struct bed *fbBed = fbToBedOne(fbPtr);
slAddHead(&ctNew->bedList, fbBed );
}
else
{
if (fields >= 6)
hPrintf("%s\t%d\t%d\t%s\t%d\t%c\n",
fbPtr->chrom, fbPtr->start, fbPtr->end, fbPtr->name,
0, fbPtr->strand);
else if (fields >= 4)
hPrintf("%s\t%d\t%d\t%s\n",
fbPtr->chrom, fbPtr->start, fbPtr->end, fbPtr->name);
else
hPrintf("%s\t%d\t%d\n",
fbPtr->chrom, fbPtr->start, fbPtr->end);
}
gotResults = TRUE;
}
featureBitsFreeList(&fbList);
}
}
bedList = NULL;
lmCleanup(&lm);
}
if (!gotResults)
{
hPrintf(NO_RESULTS);
}
else if (doCt)
{
int wigDataSize = 0;
/* Load existing custom tracks and add this new one: */
struct customTrack *ctList = getCustomTracks();
removeNamedCustom(&ctList, ctNew->tdb->table);
if (doDataPoints)
{
if (needSubtrackMerge || isBedGr || isBgWg)
{
slReverse(&dataVectorList);
wigDataSize = dataVectorWriteWigAscii(dataVectorList, ctNew->wigAscii, 0, NULL);
// TODO: see if can make prettier wig output here that
// doesn't necessarily have one value per base
}
else
{
struct wiggleDataStream *wds = NULL;
/* create an otherwise empty wds so we can print out the list */
wds = wiggleDataStreamNew();
wds->ascii = wigDataList;
wigDataSize = wds->asciiOut(wds, db, ctNew->wigAscii, TRUE, FALSE);
#if defined(DEBUG) /* dbg */
/* allow file readability for debug */
chmod(ctNew->wigAscii, 0666);
#endif
wiggleDataStreamFree(&wds);
}
}
else
slReverse(&ctNew->bedList);
slAddHead(&ctList, ctNew);
/* Save the custom tracks out to file (overwrite the old file): */
customTracksSaveCart(db, cart, ctList);
/* Put up redirect-to-browser page. */
if (redirectToGb)
{
char browserUrl[256];
char headerText[512];
int redirDelay = 3;
safef(browserUrl, sizeof(browserUrl),
"%s?%s&db=%s", hgTracksName(), cartSidUrlString(cart), database);
safef(headerText, sizeof(headerText),
"<META HTTP-EQUIV=\"REFRESH\" CONTENT=\"%d;URL=%s\">",
redirDelay, browserUrl);
webStartHeader(cart, database, headerText,
"Table Browser: %s %s: %s", hOrganism(database),
freezeName, "get custom track");
if (doDataPoints)
{
hPrintf("There are %d data points in custom track. ", wigDataSize);
}
else
{
hPrintf("There are %d items in custom track. ",
slCount(ctNew->bedList));
}
hPrintf("You will be automatically redirected to the genome browser in\n"
"%d seconds, or you can \n"
"<A HREF=\"%s\">click here to continue</A>.\n",
redirDelay, browserUrl);
}
}
else if (doDataPoints)
{
if (needSubtrackMerge || isBedGr || isBgWg)
{
slReverse(&dataVectorList);
dataVectorWriteWigAscii(dataVectorList, "stdout", 0, NULL);
}
else
{
/* create an otherwise empty wds so we can print out the list */
struct wiggleDataStream *wds = NULL;
wds = wiggleDataStreamNew();
wds->ascii = wigDataList;
wds->asciiOut(wds, db, "stdout", TRUE, FALSE);
wiggleDataStreamFree(&wds);
}
}
return gotResults;
}
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 doGenePredNongenomic(struct sqlConnection *conn, int typeIx)
/* Get mrna or protein associated with selected genes. */
{
/* Note this does do the whole genome at once rather than one
* chromosome at a time, but that's ok because the gene prediction
* tracks this serves are on the small side. */
char *typeWords[3];
char *table;
struct lm *lm = lmInit(64*1024);
int fieldCount;
struct bed *bed, *bedList = cookedBedsOnRegions(conn, curTable, getRegions(),
lm, &fieldCount);
int typeWordCount;
textOpen();
/* Figure out which table to use. */
if (isRefGeneTrack(curTable))
{
if (typeIx == 1) /* Protein */
doRefGeneProteinSequence(conn, bedList);
else
doRefGeneMrnaSequence(conn, bedList);
}
else
{
char *dupType = cloneString(findTypeForTable(database, curTrack, curTable, ctLookupName));
typeWordCount = chopLine(dupType, typeWords);
if (typeIx >= typeWordCount)
internalErr();
table = typeWords[typeIx];
if (sqlTableExists(conn, table))
{
struct sqlResult *sr;
char **row;
char query[256];
struct hash *hash = newHash(18);
boolean gotResults = FALSE;
/* Make hash of all id's passing filters. */
for (bed = bedList; bed != NULL; bed = bed->next)
hashAdd(hash, bed->name, NULL);
/* Scan through table, outputting ones that match. */
sqlSafef(query, sizeof(query), "select name, seq from %s", table);
sr = sqlGetResult(conn, query);
while ((row = sqlNextRow(sr)) != NULL)
{
if (hashLookup(hash, row[0]))
{
hPrintf(">%s\n", row[0]);
writeSeqWithBreaks(stdout, row[1], strlen(row[1]), 60);
gotResults = TRUE;
}
}
sqlFreeResult(&sr);
hashFree(&hash);
if (!gotResults)
hPrintf(NO_RESULTS);
}
else
{
internalErr();
}
freez(&dupType);
}
lmCleanup(&lm);
}
