static char *findType(struct hash *cvHash,char **requested,int requestCount,
char **queryBy, char **org,boolean silent)
/* returns the type that was requested or else the type associated with the term requested */
{
struct hashCookie hc = hashFirst(cvHash);
struct hashEl *hEl;
struct hash *ra;
char *type = typeOpt;
if (requested != NULL) // if no type, find it from requested terms. Will validate terms match type
{ // NOTE: Enter here even if there is a type, to confirm the type
while ((hEl = hashNext(&hc)) != NULL) // FIXME: This should be using mdbCv APIs to get hashes.
{ // One per "request[]"
ra = (struct hash *)hEl->val;
if (sameWord(hashMustFindVal(ra, CV_TYPE),CV_TOT)) // TOT = typeOfTerm
continue;
char *val = hashFindVal(ra, *queryBy);
if (val != NULL)
{
int ix = stringArrayIx(val,requested,requestCount);
if (ix != -1) // found
{
char *thisType = hashMustFindVal(ra, CV_TYPE);
char *thisOrg = hashFindVal(ra, ORGANISM);
if (type == NULL)
{
if (thisOrg != NULL)
{
*org = strLower(cloneString(thisOrg));
}
type = thisType;
}
else if (differentWord(type,thisType))
{
if (sameWord(CV_TERM_CONTROL,type))
type = thisType;
else if (differentWord(CV_TERM_CONTROL,thisType))
errAbort("Error: Requested %s of type '%s'. But '%s' has type '%s'\n",
*queryBy,type,requested[ix],thisType);
}
}
}
}
}
if (type == NULL && sameWord(*queryBy,CV_TERM)) // Special case of term becoming target
{
char *queryByTarget = CV_TARGET;
type = findType(cvHash,requested,requestCount,&queryByTarget,org,TRUE); // silent here
if (type != NULL)
*queryBy = queryByTarget;
}
if (type == NULL && !silent) // Still not type? abort
errAbort("Error: Required %s=%s ['%s', '%s', '%s', '%s' or '%s'] argument not found\n",
*queryBy,(requested != NULL) ? *requested : "?",
CV_TYPE, CV_TERM, CV_TAG, CV_TARGET, CV_LABEL);
return normalizeType(type);
}
enum wiggleAggregateFunctionEnum wigFetchAggregateFunctionWithCart(struct cart *theCart,
struct trackDb *tdb, char *name, char **optString)
/****** windowingFunction - Whiskers by default **************************/
{
char *Default = wiggleAggregateFunctionEnumToString(wiggleAggregateTransparent);
boolean parentLevel = isNameAtParentLevel(tdb,name);
char *aggregateFunction = NULL;
enum wiggleAggregateFunctionEnum ret;
aggregateFunction = cloneString(cartOptionalStringClosestToHome(theCart, tdb, parentLevel,
AGGREGATE));
/* If windowingFunction is a string, it came from the cart, otherwise
* see if it is specified in the trackDb option, finally
* return the default.
*/
if (!aggregateFunction)
{
char * tdbDefault =
trackDbSettingClosestToHomeOrDefault(tdb, AGGREGATE, Default);
freeMem(aggregateFunction);
if (differentWord(Default,tdbDefault))
aggregateFunction = cloneString(tdbDefault);
else
{
struct hashEl *hel;
/* no aggregateFunction from trackDb, maybe it is in tdb->settings
* (custom tracks keep settings here)
*/
aggregateFunction = cloneString(Default);
if ((tdb->settings != (char *)NULL) &&
(tdb->settingsHash != (struct hash *)NULL))
{
if ((hel =hashLookup(tdb->settingsHash, AGGREGATE)) !=NULL)
if (differentWord(Default,(char *)hel->val))
{
freeMem(aggregateFunction);
aggregateFunction = cloneString((char *)hel->val);
}
}
}
}
if (optString)
*optString = cloneString(aggregateFunction);
ret = wiggleAggregateFunctionStringToEnum(aggregateFunction);
freeMem(aggregateFunction);
return(ret);
} /* enum wiggleWindowingEnum wigFetchWindowingFunctionWithCart() */
static char *wigCheckBinaryOption(struct trackDb *tdb, char *Default,
char *notDefault, char *tdbString, char *secondTdbString)
/* A common operation for binary options (two values possible)
* check for trackDb.ra, then tdb->settings values
* return one of the two possibilities if found
* (the tdbString and secondTdbString are a result of
* early naming conventions changing over time resulting in
* two possible names for the same thing ...)
*/
{
char *tdbDefault = trackDbSettingClosestToHomeOrDefault(tdb, tdbString, "NONE");
char *ret;
ret = Default; /* the answer, unless found to be otherwise */
if (sameWord("NONE",tdbDefault) && (secondTdbString != (char *)NULL))
tdbDefault = trackDbSettingClosestToHomeOrDefault(tdb, secondTdbString, "NONE");
if (differentWord("NONE",tdbDefault))
{
if (differentWord(Default,tdbDefault))
ret = notDefault;
}
else
{
struct hashEl *hel;
/* no setting from trackDb, maybe it is in tdb->settings
* (custom tracks keep settings here)
*/
if ((tdb->settings != (char *)NULL) &&
(tdb->settingsHash != (struct hash *)NULL))
{
if ((hel = hashLookup(tdb->settingsHash, tdbString)) != NULL)
{
if (differentWord(Default,(char *)hel->val))
ret = notDefault;
}
else if (secondTdbString != (char *)NULL)
{
if ((hel = hashLookup(tdb->settingsHash, secondTdbString)) != NULL)
{
if (differentWord(Default,(char *)hel->val))
ret = notDefault;
}
}
}
}
return(cloneString(ret));
}
static void intersectDataVector(char *table, struct dataVector *dataVector1,
struct region *region, struct sqlConnection *conn)
/* Perform intersection (if specified) on dataVector. */
{
/* If table is type wig (not bedGraph), then intersection has already been
* performed on each input (other selected subtracks must be the same type
* as table).
* Otherwise, handle intersection here. */
if (anyIntersection() && !isWiggle(database, table) && !isBigWigTable(table))
{
char *track2 = cartString(cart, hgtaIntersectTrack);
char *table2 = cartString(cart, hgtaIntersectTable);
if (table2 && differentWord(table2, table))
{
struct trackDb *tdb2 = findTrack(track2, fullTrackList);
struct trackTable *tt2 = trackTableNew(tdb2, table2, conn);
struct dataVector *dataVector2 = dataVectorFetchOneRegion(tt2, region,
conn);
char *op = cartString(cart, hgtaIntersectOp);
boolean dv2IsWiggle = (isWiggle(database, table2) || isBigWigTable(table2) ||
isBedGraph(table2));
dataVectorIntersect(dataVector1, dataVector2,
dv2IsWiggle, sameString(op, "none"));
dataVectorFree(&dataVector2);
}
}
}
static char *getDbForGenome(char *genome, struct cart *cart)
/*
Function to find the default database for the given Genome.
It looks in the cart first and then, if that database's Genome matches the
passed-in Genome, returns it. If the Genome does not match, it returns the default
database that does match that Genome.
param Genome - The Genome for which to find a database
param cart - The cart to use to first search for a suitable database name
return - The database matching this Genome type
*/
{
char *retDb = cartUsualString(cart, dbCgiName, NULL);
if ((retDb == NULL) || !hDbExists(retDb))
{
retDb = hDefaultDb();
}
/* If genomes don't match, then get the default db for that genome */
if (differentWord(genome, hGenome(retDb)))
{
retDb = hDefaultDbForGenome(genome);
}
return retDb;
}
enum wiggleSmoothingEnum wigFetchSmoothingWindowWithCart(struct cart *theCart, struct trackDb *tdb,
char *name, char **optString)
/****** smoothingWindow - OFF by default **************************/
{
char * Default = wiggleSmoothingEnumToString(wiggleSmoothingOff);
boolean parentLevel = isNameAtParentLevel(tdb,name);
char * smoothingWindow = NULL;
enum wiggleSmoothingEnum ret;
smoothingWindow = cloneString(cartOptionalStringClosestToHome(theCart, tdb, parentLevel,
SMOOTHINGWINDOW));
if (!smoothingWindow) /* if nothing from the Cart, check trackDb/settings */
{
char * tdbDefault =
trackDbSettingClosestToHomeOrDefault(tdb, SMOOTHINGWINDOW, Default);
if (differentWord(Default,tdbDefault))
smoothingWindow = cloneString(tdbDefault);
else
{
struct hashEl *hel;
/* no smoothingWindow from trackDb, maybe it is in tdb->settings
* (custom tracks keep settings here)
*/
smoothingWindow = cloneString(Default);
if ((tdb->settings != (char *)NULL) &&
(tdb->settingsHash != (struct hash *)NULL))
{
if ((hel = hashLookup(tdb->settingsHash, SMOOTHINGWINDOW)) != NULL)
if (differentWord(Default,(char *)hel->val))
{
freeMem(smoothingWindow);
smoothingWindow = cloneString((char *)hel->val);
}
}
}
}
if (optString)
*optString = cloneString(smoothingWindow);
ret = wiggleSmoothingStringToEnum(smoothingWindow);
freeMem(smoothingWindow);
return(ret);
} /* enum wiggleSmoothingEnum wigFetchSmoothingWindowWithCart() */
char *showTableField(struct trackDb *track, char *varName, boolean useJoiner)
/* Show table control and label. */
{
struct slName *name, *nameList = NULL;
char *selTable;
if (track == NULL)
nameList = tablesForDb(findSelDb());
else
nameList = cartTrackDbTablesForTrack(database, track, useJoiner);
/* Get currently selected table. If it isn't in our list
* then revert to first in list. */
selTable = cartUsualString(cart, varName, nameList->name);
if (!slNameInListUseCase(nameList, selTable))
selTable = nameList->name;
/* Print out label and drop-down list. */
hPrintf("<B>table: </B>");
hPrintf("<SELECT NAME=\"%s\" %s>\n", varName, onChangeTable());
struct trackDb *selTdb = NULL;
for (name = nameList; name != NULL; name = name->next)
{
struct trackDb *tdb = NULL;
if (track != NULL)
tdb = findTdbForTable(database,track,name->name, ctLookupName);
hPrintf("<OPTION VALUE=\"%s\"", name->name);
// Disable options for related tables that are noGenome -- if a non-positional table
// is selected then we output its entire contents.
if (cartTrackDbIsNoGenome(database, name->name) &&
(track == NULL || differentString(track->table, name->name)))
hPrintf(" DISABLED"NO_GENOME_CLASS);
else if (sameString(selTable, name->name))
{
hPrintf(" SELECTED");
selTdb = tdb;
}
if (tdb != NULL)
if ((curTrack == NULL) || differentWord(tdb->shortLabel, curTrack->shortLabel))
hPrintf(">%s (%s)\n", tdb->shortLabel, name->name);
else
hPrintf(">%s\n", name->name);
else
hPrintf(">%s\n", name->name);
}
hPrintf("</SELECT>\n");
if (!trackHubDatabase(database))
{
char *restrictDate = encodeRestrictionDateDisplay(database,selTdb);
if (restrictDate)
{
hPrintf("<A HREF=\'%s\' TARGET=BLANK>restricted until:</A> %s",
ENCODE_DATA_RELEASE_POLICY, restrictDate);
freeMem(restrictDate);
}
}
return selTable;
}
void chainMethods(struct track *tg, struct trackDb *tdb,
int wordCount, char *words[])
/* Fill in custom parts of alignment chains. */
{
struct cartOptions *chainCart;
AllocVar(chainCart);
boolean normScoreAvailable = chainDbNormScoreAvailable(tdb);
/* what does the cart say about coloring option */
chainCart->chainColor = chainFetchColorOption(cart, tdb, FALSE);
chainCart->scoreFilter = cartUsualIntClosestToHome(cart, tdb,
FALSE, SCORE_FILTER, 0);
linkedFeaturesMethods(tg);
tg->itemColor = lfChromColor; /* default coloring option */
/* if normScore column is available, then allow coloring */
if (normScoreAvailable)
{
switch (chainCart->chainColor)
{
case (chainColorScoreColors):
tg->itemColor = chainScoreColor;
tg->colorShades = shadesOfGray;
break;
case (chainColorNoColors):
setNoColor(tg);
break;
default:
case (chainColorChromColors):
break;
}
}
else
{
char *optionStr; /* this old option was broken before */
optionStr = cartUsualStringClosestToHome(cart, tdb, FALSE, "color", "on");
if (differentWord("on",optionStr))
{
setNoColor(tg);
chainCart->chainColor = chainColorNoColors;
}
else
chainCart->chainColor = chainColorChromColors;
}
tg->loadItems = chainLoadItems;
tg->drawItems = chainDraw;
tg->mapItemName = lfMapNameFromExtra;
tg->subType = lfSubChain;
tg->extraUiData = (void *) chainCart;
}
int findChromIx(char *name)
{
int i;
for (i=0; i<chromCount; ++i)
{
if (!differentWord(chromNames[i], name))
return i;
}
errAbort("Unknown chromosome %s", name);
return -1;
}
static void gapSanityCheck(struct agpGap *gapList)
{
int prevEnd = 0;
int prevStart = 0;
char *prevChr = NULL;
char *prevType = NULL;
struct agpGap *gap;
for (gap = gapList; gap; gap = gap->next)
{
int chrSize = hashIntVal(cInfoHash, gap->chrom);
if (gap->chromStart < 0)
verbose(1, "WARNING: gap chromStart < 0 at %s:%d-%d\n",
gap->chrom, gap->chromStart, gap->chromEnd);
if (gap->chromEnd > chrSize)
verbose(1, "WARNING: gap chromEnd > chromSize(%d) "
"at %s:%d-%d\n", chrSize, gap->chrom,
gap->chromStart, gap->chromEnd);
if (gap->chromEnd == chrSize && differentString(gap->type, "telomere"))
verbose(1, "WARNING: gap at end of chromosome not telomere "
"at %s:%d-%d, type: %s\n", gap->chrom,
gap->chromStart, gap->chromEnd, gap->type);
if (gap->chromStart >= gap->chromEnd)
verbose(1, "WARNING: gap chromStart >= chromEnd at %s:%d-%d\n",
gap->chrom, gap->chromStart, gap->chromEnd);
if (prevEnd > 0)
{
if (sameWord(prevChr, gap->chrom) &&
(prevEnd >= gap->chromStart))
verbose(1,"WARNING: overlapping gap at "
"%s:%d-%d(%s) and %s:%d-%d(%s)\n",
gap->chrom, prevStart, prevEnd, prevType, gap->chrom,
gap->chromStart, gap->chromEnd, gap->type);
}
else
{
prevStart = gap->chromStart;
prevEnd = gap->chromEnd;
prevType = gap->type;
}
if (isNotEmpty(prevChr))
{
if (differentWord(prevChr, gap->chrom))
{
freeMem(prevChr);
prevChr = cloneString(gap->chrom);
}
}
else
prevChr = cloneString(gap->chrom);
prevStart = gap->chromStart;
prevEnd = gap->chromEnd;
}
}
struct gffGene *gffFindGeneIgnoreCase(struct gff *gff, char *geneName)
/* Find gene with given name. Not case sensitive. */
{
struct gffGene *g;
for (g=gff->genes; g!=NULL; g=g->next)
{
if (differentWord(geneName, g->name) == 0)
return g;
}
return NULL;
}
struct hash *hashPsls(char *pslFileName)
{
struct psl *pslList = NULL, *psl = NULL, *pslSubList = NULL, *pslNext = NULL;
struct hash *pslHash = newHash(15);
char *last = NULL;
char key[128];
char *tmp = NULL;
pslList = pslLoadAll(pslFileName);
/* Fix psl names */
for(psl = pslList; psl != NULL; psl = psl->next)
{
tmp = strrchr(psl->qName, ';');
*tmp = '\0';
tmp = strstr(psl->qName,prefix);
assert(tmp);
/* checks if there are 2 occurrences of ":" in probe name as in full name */
/* if probe name is shortened to fit in the seq table, there is only 1 ":"*/
/* e.g. full: consensus:HG-U133A:212933_x_at; short:HG-U133A:212933_x_at;*/
if (countChars(psl->qName, *prefix) == 2)
{
tmp = strstr(tmp+1,prefix);
assert(tmp);
}
tmp = tmp + strlen(prefix);
safef(psl->qName, strlen(psl->qName), "%s", tmp);
}
/* Sort based on query name. */
slSort(&pslList, pslCmpQuery);
/* For each psl, if it is has the same query name add it to the
sublist. Otherwise store the sublist in the hash and start
another. */
for(psl = pslList; psl != NULL; psl = pslNext)
{
pslNext = psl->next;
if(last != NULL && differentWord(last, psl->qName))
{
hashAddUnique(pslHash, last, pslSubList);
pslSubList = NULL;
}
slAddTail(&pslSubList, psl);
last = psl->qName;
}
/* Add the last sublist */
hashAddUnique(pslHash, last, pslSubList);
return pslHash;
}
static void wikiTrackMapItem(struct track *tg, struct hvGfx *hvg, void *item,
char *itemName, char *mapItemName, int start, int end, int x, int y, int width, int height)
/* create a special map box item with different i=hgcClickName and
* pop-up statusLine with the item name
*/
{
char *userName;
/* already been determined to be enabled by getting here, need to verify
* userName vs editors and vs owner
*/
(void) wikiTrackEnabled(database, &userName);
char *hgcClickName = tg->mapItemName(tg, item);
char *statusLine = tg->itemName(tg, item);
boolean editor = isWikiEditor(userName);
struct wikiTrack *wikiItem = NULL;
boolean enableHgcClick = FALSE;
/* allow hgc click (i.e. delete privs) if the following are true
1. this is the item 0 "add new item"
2. logged into the wiki
3. user is an editor or user is the owner
*/
if (differentWord("0", hgcClickName))
wikiItem = findWikiItemId(hgcClickName);
else
enableHgcClick = TRUE; /* item 0 "add new item" must go to hgc */
if (wikiItem)
{
if (isNotEmpty(userName) && sameWord(userName, wikiItem->owner))
enableHgcClick = TRUE; /* owner has delete privls */
if (editor)
enableHgcClick = TRUE; /* editors have delete privls */
}
if (enableHgcClick)
{
mapBoxHgcOrHgGene(hvg, start, end, x, y, width, height, tg->track,
hgcClickName, statusLine, NULL, FALSE, NULL );
}
else
{ /* go directly to the wiki description */
char *directUrl = wikiUrl(wikiItem);
mapBoxHgcOrHgGene(hvg, start, end, x, y, width, height, tg->track,
hgcClickName, statusLine, directUrl, FALSE, NULL);
freeMem(directUrl);
}
}
void wigFetchYLineMarkValueWithCart(struct cart *theCart,struct trackDb *tdb, char *name, double *tDbYMark )
/* y= marker line value
* User requested value is defined in the cart
* A Default value can be defined as
* yLineMark declaration from trackDb
*****************************************************************************/
{
boolean parentLevel = isNameAtParentLevel(tdb,name);
char *yLineMarkValue = NULL; /* string from cart */
double yLineValue; /* value from cart or trackDb */
char * tdbDefault = cloneString(
trackDbSettingClosestToHomeOrDefault(tdb, YLINEMARK, "NONE") );
if (sameWord("NONE",tdbDefault))
{
struct hashEl *hel;
/* no yLineMark from trackDb, maybe it is in tdb->settings
* (custom tracks keep settings here)
*/
if ((tdb->settings != (char *)NULL) &&
(tdb->settingsHash != (struct hash *)NULL))
{
if ((hel = hashLookup(tdb->settingsHash, YLINEMARK)) != NULL)
{
freeMem(tdbDefault);
tdbDefault = cloneString((char *)hel->val);
}
}
}
/* If nothing else, it is zero */
yLineValue = 0.0;
/* Let's see if a value is available in the cart */
yLineMarkValue = cartOptionalStringClosestToHome(theCart, tdb, parentLevel, YLINEMARK);
/* if yLineMarkValue is non-Null, it is the requested value */
if (yLineMarkValue)
yLineValue = atof(yLineMarkValue);
else /* See if a default line is specified in the trackDb.ra file */
if (differentWord("NONE",tdbDefault))
yLineValue = atof(tdbDefault);
/* If possible to return */
if (tDbYMark)
*tDbYMark = yLineValue;
freeMem(tdbDefault);
} /* void wigFetchYLineMarkValueWithCart() */
static int linkCmpPriority(const void *va, const void *vb)
/* Compare to sort links based on priority. */
{
const struct link *a = *((struct link **)va);
const struct link *b = *((struct link **)vb);
float dif = a->priority - b->priority;
if (dif < 0)
return -1;
else if (dif > 0)
return 1;
else
{
return differentWord(b->shortLabel, a->shortLabel);
}
}
/****** itemRgb - not on by default **************************/
boolean bedItemRgb(struct trackDb *tdb)
{
char *Default="Off"; /* anything different than this will turn it on */
char *tdbDefault = (char *)NULL;
if (tdb)
tdbDefault = trackDbSettingClosestToHome(tdb, OPT_ITEM_RGB);
if (tdbDefault)
{
if (differentWord(Default,tdbDefault))
return TRUE;
}
return FALSE;
} /* boolean bedItemRgb(struct trackDb *tdb) */
static void initialPlacement(struct chrGapList *bounding, struct bed *placed)
/* simply place the items from the 'placed' list into the gaps
* defined by the 'bounding' list, the locations of the 'placed'
* items are what they say they are in their bed structure.
*/
{
struct bed *bedEl;
int unplacedCount = 0;
for (bedEl = placed; bedEl != NULL; bedEl = bedEl->next)
{
struct chrGapList *cl;
boolean placedOK = FALSE;
for (cl = bounding; cl != NULL; cl = cl->next)
{
struct gap *gl;
if (differentWord(cl->chrom, bedEl->chrom))
continue;
for (gl = cl->gList; gl != NULL; gl = gl->next)
{
if ( ((gl->downstream != NULL) && (gl->upstream != NULL)) &&
((bedEl->chromEnd <= gl->downstream->chromStart) &&
(bedEl->chromStart >= gl->upstream->chromEnd)) )
{
placeItem(bedEl, gl);
placedOK = TRUE;
verbose(5,"item: %s:%d-%d, gap: %s:%d-%d\n",
bedEl->chrom, bedEl->chromStart, bedEl->chromEnd,
bedEl->chrom, gl->upstream->chromEnd,
gl->downstream->chromStart);
break;
}
}
}
if (! placedOK)
{
verbose(3,"can not place item: %s:%d-%d\n", bedEl->chrom,
bedEl->chromStart, bedEl->chromEnd);
++unplacedCount;
}
}
if (unplacedCount)
verbose(2,"Could not place %d items\n", unplacedCount);
}
static void linksPrint(struct section *section, struct sqlConnection *conn,
char *geneId)
/* Print the links section. */
{
int maxPerRow = 6, itemPos = 0;
int rowIx = 0;
struct link *link, *linkList = section->items;
webPrintLinkTableStart();
printGenomicSeqLink(conn, geneId, curGeneChrom, curGeneStart, curGeneEnd);
printMrnaSeqLink(conn,geneId);
printProteinSeqLink(conn,geneId);
hPrintf("</TR>\n<TR>");
++rowIx;
for (link = linkList; link != NULL; link = link->next)
{
char *url = linkGetUrl(link, conn, geneId);
if (url != NULL)
{
boolean fakeOut = link->useHgsid &&
differentWord(link->name,"wikiTrack");
char *target = (fakeOut ? "" : " TARGET=_blank");
if (++itemPos > maxPerRow)
{
hPrintf("</TR>\n<TR>");
itemPos = 1;
++rowIx;
}
if (fakeOut)
webPrintLinkCellStart();
else
webPrintLinkOutCellStart();
hPrintf("<A HREF=\"%s\"%s class=\"toc\">", url, target);
hPrintf("%s", link->shortLabel);
hPrintf("</A>");
webPrintLinkCellEnd();
freez(&url);
}
}
webFinishPartialLinkOutTable(rowIx, itemPos, maxPerRow);
webPrintLinkTableEnd();
}
void readPrimerInfo(struct lineFile *sf)
/* Read in primer info from all.primers file */
{
int wordCount;
char *words[5];
char *dist1, *dist[2];
struct sts *sts;
stsHash = newHash(16);
while (lineFileChopCharNext(sf, '\t', words, 5))
{
verbose(2, "# line %d words1-4: '%s' '%s' '%s' '%s'\n", sf->lineIx, words[1], words[2], words[3], words[4]);
if (words[1] && words[2] && words[3] && words[4])
{
AllocVar(sts);
sts->dbstsId = cloneString(words[0]);
sts->leftPrimer = cloneString(words[1]);
sts->rightPrimer = cloneString(words[2]);
sts->size = cloneString(words[3]);
sts->ucscId = cloneString(words[4]);
sts->found = FALSE;
dist1 = cloneString(words[3]);
if (sts->leftPrimer && dist1 && differentWord("-", dist1))
{
wordCount = chopByChar(dist1, '-', dist, ArraySize(dist));
sts->minSize = sqlUnsigned(dist[0]);
if (wordCount == 1)
sts->maxSize = sqlUnsigned(dist[0]);
else
sts->maxSize = sqlUnsigned(dist[1]);
if (sts->maxSize == 0)
sts->maxSize = 1000;
sts->next = NULL;
sts->place = NULL;
sts->epcr = NULL;
hashAdd(stsHash, sts->dbstsId, sts);
}
slAddHead(&stsList, sts);
}
}
}
static void synonymPrint(struct section *section,
struct sqlConnection *conn, char *id)
/* Print out SwissProt comments - looking up typeId/commentVal. */
{
char *protAcc = getSwissProtAcc(conn, spConn, id);
char *spDisplayId;
char *refSeqAcc = "";
char *mrnaAcc = "";
char *oldDisplayId;
char condStr[255];
char *kgProteinID;
char *parAcc; /* parent accession of a variant splice protein */
char *chp;
if (isRgdGene(conn))
{
rgdGene2SynonymPrint(section,conn, id);
return;
}
if (sqlTablesExist(conn, "kgAlias"))
printAlias(id, conn);
if (sameWord(genome, "Zebrafish"))
{
char *xrefTable = "ensXRefZfish";
char *geneIdCol = "ensGeneId";
/* get Gene Symbol and RefSeq accession from Zebrafish-specific */
/* cross-reference table */
printGeneSymbol(id, xrefTable, geneIdCol, conn);
refSeqAcc = getRefSeqAcc(id, xrefTable, geneIdCol, conn);
hPrintf("<B>ENSEMBL ID:</B> %s", id);
}
else
{
char query[256];
char *toRefTable = genomeOptionalSetting("knownToRef");
if (toRefTable != NULL && sqlTableExists(conn, toRefTable))
{
safef(query, sizeof(query), "select value from %s where name='%s'", toRefTable,
id);
refSeqAcc = emptyForNull(sqlQuickString(conn, query));
}
if (sqlTableExists(conn, "kgXref"))
{
safef(query, sizeof(query), "select mRNA from kgXref where kgID='%s'", id);
mrnaAcc = emptyForNull(sqlQuickString(conn, query));
}
if (sameWord(genome, "C. elegans"))
hPrintf("<B>WormBase ID:</B> %s<BR>", id);
else
hPrintf("<B>UCSC ID:</B> %s<BR>", id);
}
if (refSeqAcc[0] != 0)
{
hPrintf("<B>RefSeq Accession: </B> <A HREF=\"");
printOurRefseqUrl(stdout, refSeqAcc);
hPrintf("\">%s</A><BR>\n", refSeqAcc);
}
else if (mrnaAcc[0] != 0)
{
safef(condStr, sizeof(condStr), "acc = '%s'", mrnaAcc);
if (sqlGetField(database, "gbCdnaInfo", "acc", condStr) != NULL)
{
hPrintf("<B>Representative RNA: </B> <A HREF=\"");
printOurMrnaUrl(stdout, mrnaAcc);
hPrintf("\">%s</A><BR>\n", mrnaAcc);
}
else
/* do not show URL link if it is not found in gbCdnaInfo */
{
hPrintf("<B>Representative RNA: %s </B>", mrnaAcc);
}
}
if (protAcc != NULL)
{
kgProteinID = cloneString("");
if (hTableExists(sqlGetDatabase(conn), "knownGene")
&& (isNotEmpty(cartOptionalString(cart, hggChrom)) &&
differentWord(cartOptionalString(cart, hggChrom),"none")))
{
safef(condStr, sizeof(condStr), "name = '%s' and chrom = '%s' and txStart=%s and txEnd=%s",
id, cartOptionalString(cart, hggChrom),
cartOptionalString(cart, hggStart),
cartOptionalString(cart, hggEnd));
kgProteinID = sqlGetField(database, "knownGene", "proteinID", condStr);
}
hPrintf("<B>Protein: ");
if (strstr(kgProteinID, "-") != NULL)
{
parAcc = cloneString(kgProteinID);
chp = strstr(parAcc, "-");
*chp = '\0';
/* show variant splice protein and the UniProt link here */
hPrintf("<A HREF=\"http://www.uniprot.org/uniprot%s\" "
"TARGET=_blank>%s</A></B>, splice isoform of ",
kgProteinID, kgProteinID);
hPrintf("<A HREF=\"http://www.uniprot.org/uniprot/%s\" "
"TARGET=_blank>%s</A></B>\n",
parAcc, parAcc);
}
else
{
hPrintf("<A HREF=\"http://www.uniprot.org/uniprot/%s\" "
"TARGET=_blank>%s</A></B>\n",
protAcc, protAcc);
}
/* show SWISS-PROT display ID if it is different than the accession ID */
/* but, if display name is like: Q03399 | Q03399_HUMAN, then don't show display name */
spDisplayId = spAnyAccToId(spConn, protAcc);
if (spDisplayId == NULL)
{
errAbort("<br>%s seems to no longer be a valid protein ID in our latest UniProtKB DB.", protAcc);
}
if (strstr(spDisplayId, protAcc) == NULL)
{
hPrintf(" (aka %s", spDisplayId);
/* show once if the new and old displayId are the same */
oldDisplayId = oldSpDisplayId(spDisplayId);
if (oldDisplayId != NULL)
{
if (!sameWord(spDisplayId, oldDisplayId)
&& !sameWord(protAcc, oldDisplayId))
{
hPrintf(" or %s", oldDisplayId);
}
}
hPrintf(")<BR>\n");
}
}
printCcds(id, conn);
}
void analyse(int start, int stop)
{
struct hash *hash;
char line[512];
int lineCount = 0;
char *words[32];
int wordCount;
struct cdnaInfo *cdnaList = NULL;
struct cdnaInfo *ci = NULL;
int cdnaCount;
int maxCdnaCount = stop - start;
cdnaCount = 1;
if (start > 1)
{
for (;;)
{
if (!fgets(line, sizeof(line), inFile))
errAbort("Not %d cDNAs in file, only %d\n", start, cdnaCount);
++lineCount;
if (line[0] == '#') /* Skip comments. */
continue;
wordCount = chopString(line, whiteSpaceChopper, words, ArraySize(words));
if (wordCount <= 0) /* Skip empty lines. */
continue;
if (!differentWord(words[1], "alignments"))
{
++cdnaCount;
if (cdnaCount >= start)
break;
}
}
}
cdnaCount = 0;
hash = newHash(14); /* Hash table with 16k entries. */
for (;;)
{
if (!fgets(line, sizeof(line), inFile))
break;
++lineCount;
if (line[0] == '#') /* Skip comments. */
continue;
wordCount = chopString(line, whiteSpaceChopper, words, ArraySize(words));
if (wordCount <= 0) /* Skip empty lines. */
continue;
if (wordCount < 4) /* Everyone else has at least four words. */
{
errAbort("Short line %d:\n", lineCount);
}
if (sameWord(words[1], "Blasting"))
{
char *cdnaName = words[2];
if ((ci = lookupInfo(hash, cdnaName)) == NULL)
{
struct hashEl *hel;
ci = needMem(sizeof(*ci));
hel = hashAdd(hash, cdnaName, ci);
ci->next = cdnaList;
cdnaList = ci;
ci->ix = atoi(words[0]);
ci->name = hel->name;
}
}
else if (sameWord(words[2], "hits"))
{
/* Newer style - includes cDNA matching range. */
if (ci == NULL)
continue;
hitLine(ci, lineCount, words[0], words[1], words[3], words[4], words[5], words[9]);
}
else if (sameWord(words[1], "hits"))
/* Older style - no cDNA matching range. */
{
if (ci == NULL)
continue;
hitLine(ci, lineCount, words[0], NULL, words[2], words[3], words[4], words[8]);
}
else if (sameWord(words[1], "alignments"))
{
struct dnaSeq *cdnaSeq;
struct wormCdnaInfo info;
if (ci == NULL)
continue;
if (differentWord(ci->name, words[3]))
errAbort("Line %d - %s is not %s", lineCount, words[3], ci->name);
if (!ci->finished)
{
if (!anyCdnaSeq(ci->name, &cdnaSeq, &info))
{
warn("Can't find cDNA %s", ci->name);
ci->isDupe = TRUE;
}
else
{
ci->baseCount = cdnaSeq->size;
ci->baseCrc = dnaCrc(cdnaSeq->dna, cdnaSeq->size);
slReverse(&ci->roughAli);
ci->roughScore = bestRoughScore(ci->roughAli);
filterDupeCdna(ci, cdnaSeq);
ci->isBackwards = (info.orientation == '-');
refineAlis(ci, cdnaSeq);
ci->fineScore = bestFineScore(ci->fineAli);
ci->isEmbryonic = info.isEmbryonic;
ci->finished = TRUE;
freeDnaSeq(&cdnaSeq);
++cdnaCount;
if (cdnaCount >= maxCdnaCount)
break;
}
}
}
else
{
errAbort("Can't deal with line %d\n", lineCount);
}
}
slReverse(&cdnaList);
doGoodBad(cdnaList);
doUnusual(cdnaList);
//makeCdnaToGene(cdnaList);
/* Clean up. */
/* These two are slow and not really necessary. */
#ifdef FASTIDIOUS
slFreeList(&cdnaList);
freeHash(&hash);
#endif
uglyf("Done analyse\n");
}
void bedItemOverlapCount(struct hash *chromHash, char *infile, char *outfile){
unsigned maxChromSize = 0;
unitSize *counts = (unitSize *)NULL;
FILE *f = mustOpen(outfile, "w");
struct hashCookie hc = hashFirst(chromHash);
struct hashEl *hel;
while( (hel = hashNext(&hc)) != NULL) {
unsigned num = (unsigned) ptToInt(hel->val);
maxChromSize = max(num, maxChromSize);
}
verbose(2,"#\tmaxChromSize: %u\n", maxChromSize);
if (maxChromSize < 1)
errAbort("maxChromSize is zero ?");
/* Allocate just once for the largest chrom and reuse this array */
counts = needHugeMem(sizeof(unitSize) * maxChromSize);
/* Reset the array to be zero to be reused */
memset((void *)counts, 0, sizeof(unitSize)*(size_t)maxChromSize);
unsigned chromSize = 0;
char *prevChrom = (char *)NULL;
boolean outputToDo = FALSE;
struct hash *seenHash = newHash(5);
struct lineFile *bf = lineFileOpen(infile , TRUE);
struct bed *bed = (struct bed *)NULL;
char *row[12];
int numFields = doBed12 ? 12 : 3;
while (lineFileNextRow(bf,row, numFields))
{
int i;
bed = bedLoadN(row, numFields);
verbose(3,"#\t%s\t%d\t%d\n",bed->chrom,bed->chromStart, bed->chromEnd);
if (prevChrom && differentWord(bed->chrom,prevChrom)) // End a chr
{
verbose(2,"#\tchrom %s done, size %d\n", prevChrom, chromSize);
if (outputToDo)
outputCounts(counts, prevChrom, chromSize, f);
outputToDo = FALSE;
memset((void *)counts, 0,
sizeof(unitSize)*(size_t)maxChromSize); /* zero counts */
freez(&prevChrom);
// prevChrom is now NULL so it will be caught by next if!
}
if ((char *)NULL == prevChrom) // begin a chr
{
if (hashLookup(seenHash, bed->chrom))
errAbort("ERROR:input file not sorted. %s seen before on line %d\n",
bed->chrom, bf->lineIx);
hashAdd(seenHash, bed->chrom, NULL);
prevChrom = cloneString(bed->chrom);
chromSize = hashIntVal(chromHash, prevChrom);
verbose(2,"#\tchrom %s starting, size %d\n", prevChrom,chromSize);
}
if (bed->chromEnd > chromSize)
{
// check for circular chrM
if (doBed12 || bed->chromStart>=chromSize
|| differentWord(bed->chrom,"chrM"))
{
warn("ERROR: %s\t%d\t%d", bed->chrom, bed->chromStart,
bed->chromEnd);
errAbort("chromEnd > chromSize ? %d > %d",
bed->chromEnd,chromSize);
}
for (i = bed->chromStart; i < chromSize; ++i)
INCWOVERFLOW(counts,i);
for (i = 0; i < (bed->chromEnd - chromSize); ++i)
INCWOVERFLOW(counts,i);
}
else if (doBed12)
{
int *starts = bed->chromStarts;
int *sizes = bed->blockSizes;
int *endStarts = &bed->chromStarts[bed->blockCount];
for(; starts < endStarts; starts++, sizes++)
{
unsigned int end = *starts + *sizes + bed->chromStart;
for (i = *starts + bed->chromStart; i < end; ++i)
INCWOVERFLOW(counts,i);
}
}
else
{
for (i = bed->chromStart; i < bed->chromEnd; ++i)
INCWOVERFLOW(counts, i);
}
outputToDo = TRUE;
bedFree(&bed); // plug the memory leak
}
lineFileClose(&bf);
// Note, next file could be on same chr!
if (outputToDo)
outputCounts(counts, prevChrom, chromSize, f);
if (doOutBounds)
fprintf(stderr, "min %lu max %lu\n", (unsigned long)overMin, (unsigned long)overMax);
verbose(2,"#\tchrom %s done, size %d\n", prevChrom, chromSize);
carefulClose(&f);
freeMem(counts);
freez(&prevChrom);
// hashFreeWithVals(&chromHash, freez);
freeHash(&seenHash);
}
void hitLine(struct cdnaInfo *ci, int lineCount, char *cdnaName, char *cdnaRange,
char *mitoChrom, char *chromName, char *genomeRange, char *scoreString)
/* Process line with hit info. */
{
char *s;
struct roughAli *ra;
if (ci == NULL || differentWord(cdnaName, ci->name))
{
errAbort("Got hit line %d without corresponding blasting for cDNA %s",
lineCount, cdnaName);
}
/* Allocate structure to hold data for line. */
ra = needMem(sizeof(*ra));
/* Grab range of bases covered in cDNA */
if (cdnaRange != NULL)
{
s = cdnaRange;
ra->cStart = atoi(s);
s = strchr(s, '-')+1;
ra->cEnd = atoi(s)+1;
if (ra->cStart >= ra->cEnd)
{
int temp;
warn("%s line %d cdna %s: cStart %d > cEnd %d. Swapping.",
cdnaName, lineCount, cdnaName, ra->cStart, ra->cEnd);
addRedoHash(ci, "swappedCdna");
temp = ra->cStart;
ra->cStart = ra->cEnd;
ra->cEnd = ra->cStart;
}
}
/* Grab range of bases covered in genome. */
s = genomeRange;
ra->gStart = atoi(s);
s = strchr(s, '-')+1;
ra->gEnd = atoi(s)+1;
if (ra->gStart >= ra->gEnd)
{
int temp;
warn("%s line %d cdna %s: gStart %d > gEnd %d. Swapping.",
cdnaName, lineCount, cdnaName, ra->gStart, ra->gEnd);
addRedoHash(ci, "swappedGenome");
temp = ra->gStart;
ra->gStart = ra->gEnd;
ra->gEnd = ra->gStart;
}
/* Figure out chromosome it's on. */
if (!differentWord(mitoChrom, "Mitochondrial"))
ra->chromIx = 6;
else
{
int i;
for (i=0; i<chromCount; ++i)
{
if (!differentWord(chromName, chromNames[i]))
{
ra->chromIx = i;
break;
}
}
}
/* Get alignment score. */
ra->score = atoi(scoreString);
/* Hang it on list. */
ra->next = ci->roughAli;
ci->roughAli = ra;
}
boolean viewableOutput(char *output)
/* Returns TRUE if this is a viewable output type */
{
return differentWord(output, "reads");
}
boolean gffReadGenes(struct gff *gff)
/* Read all the gene (as opposed to base) info in file. */
{
int wordCount;
struct gffSegLine seg;
char curGroup[128];
struct gffGene *gene = NULL;
GffIntron *intron = NULL;
GffExon *exon = NULL;
boolean warnedUnknown = FALSE;
boolean isNewGene;
curGroup[0] = 0; /* Start off with no group */
/* Line scanning loop. */
for (;;)
{
/* Get next line and parse it into segLine data structure. */
if (!_gffGetLine(gff))
break; /* End of file. */
if (gff->buf[0] == '#')
continue; /* Ignore sharp containing lines. */
wordCount = gffSegLineScan(gff, &seg);
if (wordCount < 9)
continue; /* Ignore blank lines and short ones. */
/* Make sure that start is less than or equal end. */
if (seg.start > seg.end)
{
warn("start greater than end line %d of %s.\n",
gff->lineNumber, gff->fileName);
return FALSE;
}
/* Get the gene we're working on. First see if
* it's the same as last time around. */
isNewGene = FALSE;
if (strcmp(seg.group, curGroup) != 0)
{
strcpy(curGroup, seg.group);
if ((gene = gffFindGene(gff, seg.group)) == NULL)
{
/* It's a new gene! */
if (!checkWordCount(gff, wordCount)) return FALSE;
isNewGene = TRUE;
gene = gffNeedMem(gff, sizeof(*gene));
strcpy(gene->name, seg.group);
slAddTail(&gff->genes, gene);
gene->strand = seg.strand[0];
gene->frame = atoi(seg.frame);
if (differentWord(seg.feature, "CDS") == 0)
{
gene->start = seg.start-1;
gene->end = seg.end-1;
}
}
}
/* Look at what sort of feature it is, and decide what to do. */
if (differentWord(seg.feature, "CDS")==0)
{
/* CDS (coding segments) have been processed already
* for the most part. Here just make sure they aren't
* duplicated. */
if (!checkWordCount(gff, wordCount)) return FALSE;
if (!isNewGene)
{
if (gene->start != 0 || gene->end != 0)
{
warn("Warning duplicate CDS for %s\n",
seg.group);
warn("Line %d of %s\n",
gff->lineNumber, gff->fileName);
}
}
}
else if (differentWord(seg.feature, "SE") == 0
|| differentWord(seg.feature, "IE") == 0
|| differentWord(seg.feature, "FE") == 0
|| differentWord(seg.feature, "E") == 0
|| differentWord(seg.feature, "exon") == 0)
{
/* It's some sort of exon. We'll deal with the complications
* of it being possibly on the minus strand later, so can
* tread initial, final, single, and regular exons the same
* here. */
if (!checkWordCount(gff, wordCount)) return FALSE;
exon = gffNeedMem(gff, sizeof(*exon));
exon->start = seg.start-1;
exon->end = seg.end-1;
exon->frame = atoi(seg.frame);
gffSegmentInsertSort(&gene->exons, exon);
}
else if (differentWord(seg.feature, "I") == 0
|| differentWord(seg.feature, "intron") == 0)
{
/* It's an intron. */
if (!checkWordCount(gff, wordCount)) return FALSE;
intron = gffNeedMem(gff, sizeof(*intron));
intron->start = seg.start-1;
intron->end = seg.end-1;
intron->frame = atoi(seg.frame);
gffSegmentInsertSort(&gene->introns, intron);
}
else if (strcmp(seg.feature, "IG") == 0)
{
/* I don't know what it is, but we can ignore it. */
}
else
{
if (!warnedUnknown)
{
warn("Unknown feature %s line %d of %s, ignoring\n",
seg.feature, gff->lineNumber, gff->fileName);
warnedUnknown = TRUE;
}
}
}
/* Fix up gene length from exons if needed. */
for (gene = gff->genes; gene != NULL; gene = gene->next)
{
if (gene->start >= gene->end)
{
offsetsFromExons(gene);
}
}
return TRUE;
}
static void gapToLift(char *db, char *outFile)
/* gapToLift - create lift file from gap table(s). */
{
FILE *out = mustOpen(outFile, "w");
struct sqlConnection *conn = sqlConnect(db);
struct chromInfo *cInfoList = loadChromInfo(conn);
struct agpGap *gapList = loadAllGaps(conn, db, cInfoList);
struct agpGap *gap;
int start = 0;
int end = 0;
char *prevChr = NULL;
int liftCount = 0;
int chrSize = 0;
static struct hash *chrDone = NULL;
chrDone = newHash(0);
if (isNotEmpty(bedFileName))
{
bedFile = mustOpen(bedFileName, "w");
verbose(2,"#\tbed output requested to %s\n", bedFileName);
}
for (gap = gapList; gap; gap = gap->next)
{
verbose(3,"#\t%s\t%d\t%d\t%s\n", gap->chrom, gap->chromStart,
gap->chromEnd, gap->bridge);
if (prevChr && sameWord(prevChr, gap->chrom))
{ /* continuing same segment, check for gap break,
* or gap at end of chrom */
if (sameWord("no",gap->bridge) || (gap->chromEnd == chrSize))
{
end = gap->chromStart;
liftCount = liftOutLine(out, gap->chrom, start, end, liftCount, chrSize);
start = gap->chromEnd;
end = start;
}
else
end = gap->chromEnd;
}
else /* new chrom encountered */
{ /* output last segment of previous chrom when necessary */
if (prevChr && differentWord(prevChr, gap->chrom))
{
if (end < chrSize)
liftCount = liftOutLine(out, prevChr, start, chrSize, liftCount, chrSize);
}
liftCount = 0;
chrSize = hashIntVal(cInfoHash, gap->chrom);
hashAddInt(chrDone, gap->chrom, 1);
if (gap->chromStart > 0)
{ /* starting first segment at position 0 */
start = 0;
end = gap->chromStart;
/* does the first gap break it ? Or gap goes to end of chrom. */
if (sameWord("no",gap->bridge) || (gap->chromEnd == chrSize))
{
liftCount = liftOutLine(out, gap->chrom, start, end, liftCount, chrSize);
start = gap->chromEnd;
end = start;
}
}
else /* first gap is actually the beginning of the chrom */
{ /* thus, first segment starts after this first gap */
start = gap->chromEnd;
end = start;
}
}
prevChr = gap->chrom; /* remember prev chrom to detect next chrom */
}
/* potentially a last one */
if (end < chrSize)
liftCount = liftOutLine(out, prevChr, start, chrSize, liftCount, chrSize);
/* check that all chroms have been used */
struct hashCookie cookie = hashFirst(cInfoHash);
struct hashEl *hel;
while ((hel = hashNext(&cookie)) != NULL)
{
if (NULL == hashLookup(chrDone, hel->name))
{
chrSize = hashIntVal(cInfoHash, hel->name);
verbose(2, "#\tno gaps on chrom: %s, size: %d\n", hel->name, chrSize);
liftCount = liftOutLine(out, hel->name, 0, chrSize, 0, chrSize);
}
}
carefulClose(&out);
sqlDisconnect(&conn);
}
void getDbGenomeClade(struct cart *cart, char **retDb, char **retGenome,
char **retClade, struct hash *oldVars)
/* Examine CGI and cart variables to determine which db, genome, or clade
* has been selected, and then adjust as necessary so that all three are
* consistent. Detect changes and reset db-specific cart variables.
* Save db, genome and clade in the cart so it will be consistent hereafter.
* The order of preference here is as follows:
* If we got a request that explicitly names the db, that takes
* highest priority, and we synch the organism to that db.
* If we get a cgi request for a specific organism then we use that
* organism to choose the DB. If just clade, go from there.
* In the cart only, we use the same order of preference.
* If someone requests an Genome we try to give them the same db as
* was in their cart, unless the Genome doesn't match.
*/
{
boolean gotClade = hGotClade();
*retDb = cgiOptionalString(dbCgiName);
*retGenome = cgiOptionalString(orgCgiName);
*retClade = cgiOptionalString(cladeCgiName);
/* phoneHome business */
phoneHome();
/* Was the database passed in as a cgi param?
* If so, it takes precedence and determines the genome. */
if (*retDb && hDbExists(*retDb))
{
*retGenome = hGenome(*retDb);
}
/* If no db was passed in as a cgi param then was the organism (a.k.a. genome)
* passed in as a cgi param?
* If so, the we use the proper database for that genome. */
else if (*retGenome && !sameWord(*retGenome, "0"))
{
*retDb = getDbForGenome(*retGenome, cart);
*retGenome = hGenome(*retDb);
}
else if (*retClade && gotClade)
{
*retGenome = hDefaultGenomeForClade(*retClade);
*retDb = getDbForGenome(*retGenome, cart);
}
/* If no cgi params passed in then we need to inspect the session */
else
{
*retDb = cartOptionalString(cart, dbCgiName);
*retGenome = cartOptionalString(cart, orgCgiName);
*retClade = cartOptionalString(cart, cladeCgiName);
/* If there was a db found in the session that determines everything. */
if (*retDb && hDbExists(*retDb))
{
*retGenome = hGenome(*retDb);
}
else if (*retGenome && !sameWord(*retGenome, "0"))
{
*retDb = hDefaultDbForGenome(*retGenome);
}
else if (*retClade && gotClade)
{
*retGenome = hDefaultGenomeForClade(*retClade);
*retDb = getDbForGenome(*retGenome, cart);
}
/* If no organism in the session then get the default db and organism. */
else
{
*retDb = hDefaultDb();
*retGenome = hGenome(*retDb);
}
}
*retDb = cloneString(*retDb);
*retGenome = cloneString(*retGenome);
*retClade = hClade(*retGenome);
/* Detect change of database and reset db-specific cart variables: */
if (oldVars)
{
char *oldDb = hashFindVal(oldVars, "db");
char *oldOrg = hashFindVal(oldVars, "org");
char *oldClade = hashFindVal(oldVars, "clade");
if ((!IS_CART_VAR_EMPTY(oldDb) && differentWord(oldDb, *retDb)) ||
(!IS_CART_VAR_EMPTY(oldOrg) && differentWord(oldOrg, *retGenome)) ||
(!IS_CART_VAR_EMPTY(oldClade) && differentWord(oldClade, *retClade)))
{
/* Change position to default -- unless it was passed in via CGI: */
if (cgiOptionalString("position") == NULL)
cartSetString(cart, "position", hDefaultPos(*retDb));
/* hgNear search term -- unless it was passed in via CGI: */
if (cgiOptionalString("near_search") == NULL)
cartRemove(cart, "near_search");
/* hgBlat results (hgUserPsl track): */
cartRemove(cart, "ss");
/* hgTables correlate: */
cartRemove(cart, "hgta_correlateTrack");
cartRemove(cart, "hgta_correlateTable");
cartRemove(cart, "hgta_correlateGroup");
cartRemove(cart, "hgta_correlateOp");
cartRemove(cart, "hgta_nextCorrelateTrack");
cartRemove(cart, "hgta_nextCorrelateTable");
cartRemove(cart, "hgta_nextCorrelateGroup");
cartRemove(cart, "hgta_nextCorrelateOp");
cartRemove(cart, "hgta_corrWinSize");
cartRemove(cart, "hgta_corrMaxLimitCount");
}
}
/* Save db, genome (as org) and clade in cart. */
cartSetString(cart, "db", *retDb);
cartSetString(cart, "org", *retGenome);
if (gotClade)
cartSetString(cart, "clade", *retClade);
}
struct trackHubSettingSpec *trackHubSettingsForVersion(char *specHost, char *version)
/* Return list of settings with support level. Version can be version string or spec url */
{
struct htmlPage *page = NULL;
if (version == NULL)
{
version = trackHubVersionDefault(specHost, &page);
if (version == NULL)
errAbort("Can't get default spec from host %s", specHost);
}
/* Retrieve specs from file url.
* Settings are the first text word within any <code> tag having class="level-" attribute.
* The level represents the level of support for the setting (e.g. base, full, deprecated)
* The support level ('level-*') is the class value of the * <code> tag.
* E.g. <code class="level-full">boxedConfig on</code> produces:
* setting=boxedConfig, class=full */
if (page == NULL)
page = trackHubVersionSpecMustGet(specHost, version);
if (page == NULL)
errAbort("Can't get settings spec for version %s from host %s", version, specHost);
verbose(5, "Found %d tags\n", slCount(page->tags));
struct trackHubSettingSpec *spec, *savedSpec;
struct hash *specHash = hashNew(0);
struct htmlTag *tag;
struct htmlAttribute *attr;
char buf[256];
for (tag = page->tags; tag != NULL; tag = tag->next)
{
if (differentWord(tag->name, "code"))
continue;
attr = tag->attributes;
if (attr == NULL || differentString(attr->name, "class") || !startsWith("level-", attr->val))
continue;
AllocVar(spec);
int len = min(tag->next->start - tag->end, sizeof buf - 1);
memcpy(buf, tag->end, len);
buf[len] = 0;
verbose(6, "Found spec: %s\n", buf);
spec->name = cloneString(firstWordInLine(buf));
if (spec->name == NULL || strlen(spec->name) == 0)
{
warn("ERROR: format problem with trackDbHub.html -- contact UCSC.");
continue;
}
spec->level = cloneString(chopPrefixAt(attr->val, '-'));
verbose(6, "spec: name=%s, level=%s\n", spec->name, spec->level);
savedSpec = (struct trackHubSettingSpec *)hashFindVal(specHash, spec->name);
if (savedSpec != NULL)
verbose(6, "found spec %s level %s in hash\n", savedSpec->name, savedSpec->level);
if (savedSpec == NULL)
{
hashAdd(specHash, spec->name, spec);
verbose(6, "added spec %s at level %s\n", spec->name, spec->level);
}
else if (trackHubSettingLevelCmp(spec, savedSpec) > 0)
{
hashReplace(specHash, spec->name, spec);
verbose(6, "replaced spec %s at level %s, was %s\n",
spec->name, spec->level, savedSpec->level);
}
}
struct hashEl *el, *list = hashElListHash(specHash);
int settingsCt = slCount(list);
verbose(5, "Found %d settings's\n", slCount(list));
if (settingsCt == 0)
errAbort("Can't find hub setting info for version %s (host %s)."
" Use -version to indicate a different version number or url.", version, specHost);
slSort(&list, hashElCmp);
struct trackHubSettingSpec *specs = NULL;
int baseCt = 0;
int requiredCt = 0;
int deprecatedCt = 0;
for (el = list; el != NULL; el = el->next)
{
if (sameString(((struct trackHubSettingSpec *)el->val)->level, "base"))
baseCt++;
else if (sameString(((struct trackHubSettingSpec *)el->val)->level, "required"))
requiredCt++;
else if (sameString(((struct trackHubSettingSpec *)el->val)->level, "deprecated"))
deprecatedCt++;
slAddHead(&specs, el->val);
}
slReverse(&specs);
verbose(3,
"Found %d supported settings for this version (%d required, %d base, %d deprecated)\n",
slCount(specs), requiredCt, baseCt, deprecatedCt);
return specs;
}
int rbTreeCmpWord(void *a, void *b)
/* Set up rbTree so as to work on case-insensitive strings. */
{
return differentWord(a,b);
}
/* The single externally visible routine.
* Future improvements will need to add a couple more arguments to
* satisify the needs of the command line version and its options.
* Currently, this is used only in customTrack input parsing.
*/
void wigAsciiToBinary( char *wigAscii, char *wigFile, char *wibFile,
double *upperLimit, double *lowerLimit, struct wigEncodeOptions *options)
/* given the three file names, read the ascii wigAscii file and produce
* the wigFile and wibFile outputs
*/
{
struct lineFile *lf; /* for line file utilities */
char *line = (char *) NULL; /* to receive data input line */
char *words[10]; /* to split data input line */
int wordCount = 0; /* result of split */
int validLines = 0; /* counting only lines with data */
double dataValue = 0.0; /* from data input */
boolean bedData = FALSE; /* in bed format data */
boolean variableStep = FALSE; /* in variableStep data */
boolean fixedStep = FALSE; /* in fixedStep data */
char *prevChromName = (char *)NULL; /* to watch for chrom name changes */
int trackCount = 0; /* We abort if we see more than one track. */
if ((wigAscii == (char *)NULL) || (wigFile == (char *)NULL) ||
(wibFile == (char *)NULL))
errAbort("wigAsciiToBinary: missing data file names, ascii: %s, wig: %s, wib: %s", wigAscii, wigFile, wibFile);
/* need to be careful here and initialize all the global variables */
freez(&wibFileName); /* send this name to the global */
wibFileName = cloneString(wibFile); /* variable for use in output_row() */
lineCount = 0; /* to count all lines */
add_offset = 0; /* to allow "lifting" of the data */
validLines = 0; /* to count only lines with data */
rowCount = 0; /* to count rows output */
bincount = 0; /* to count up to binsize */
binsize = 1024; /* # of data points per table row */
dataSpan = 1; /* default bases spanned per data point */
chromStart = 0; /* for table row data */
previousOffset = 0; /* for data missing detection */
fileOffset = 0; /* current location within binary data file */
fileOffsetBegin = 0;/* location in binary data file where this bin starts*/
freez(&data_values);
freez(&validData);
data_values = (double *) needMem( (size_t) (binsize * sizeof(double)));
validData = (unsigned char *)
needMem( (size_t) (binsize * sizeof(unsigned char)));
if (options != NULL)
{
if (options->lift != 0)
add_offset = options->lift;
if (options->noOverlap)
noOverlap = TRUE;
if (options->flagOverlapSpanData)
flagOverlapSpanData = TRUE;
if (options->wibSizeLimit > 0)
wibSizeLimit = options->wibSizeLimit;
}
/* limits for the complete set of data, they must change from these initial
defaults during processing */
overallLowerLimit = wigEncodeStartingLowerLimit;
overallUpperLimit = wigEncodeStartingUpperLimit;
binout = mustOpen(wibFile,"w"); /* binary data file */
wigout = mustOpen(wigFile,"w"); /* table row definition file */
#if defined(DEBUG) /* dbg */
chmod(wibFile, 0666);
chmod(wigFile, 0666);
#endif
lf = lineFileOpen(wigAscii, TRUE); /* input file */
while (lineFileNext(lf, &line, NULL))
{
boolean readingFrameSlipped;
++lineCount;
if ((wibSizeLimit > 0) && (wibSize >= wibSizeLimit))
errAbort("data size limit of %lld data values has been exceeded. This data can be efficiently displayed with the <A HREF='/goldenPath/help/bigWig.html' TARGET=_blank>bigWig file format</A> in a custom track, or in a <A HREF='/goldenPath/help/hgTrackHubHelp.html' TARGET=_blank>Track Hub</A> for multiple large datasets.", wibSizeLimit);
line = skipLeadingSpaces(line);
/* ignore blank or comment lines */
if ((line == (char *)NULL) || (line[0] == '\0') || (line[0] == '#'))
continue; /* !!! go to next line of input */
wordCount = chopByWhite(line, words, ArraySize(words));
if (sameWord("track",words[0]))
{
/* Allow (and ignore) one track line, but no more. */
++trackCount;
if (trackCount > 1)
errAbort("Multiple tracks seen, second at line %d of %s, can only handle one.",
lf->lineIx, lf->fileName);
continue;
}
else if (sameWord("browser", words[0]))
{
continue; /* ignore browser lines if present */
}
else if (sameWord("variableStep",words[0]))
{
int i;
boolean foundChrom = FALSE;
/* safest thing to do if we were processing anything is to
* output that previous block and start anew
* Future improvement could get fancy here and decide if it
* is really necessary to start over, although the concept
* of a line between data points on one item may use this
* block behavior later to define line segments, so don't
* get too quick to be fancy here. This line behavior
* implies that feature names will need to be specified to
* identify the line segments that belong together.
*/
if (variableStep || bedData || fixedStep)
{
output_row();
validLines = 0; /* to cause reset for first offset */
}
dataSpan = 1; /* default bases spanned per data point */
for(i = 1; i < wordCount; ++i)
{
if (startsWith("chrom",words[i]))
{
setChromName(words[i]);
foundChrom = TRUE;
}
else if (startsWith("span",words[i]))
setDataSpan(words[i]);
else
errAbort("illegal specification on variableStep at line %lu: %s",
lineCount, words[i]);
}
if (!foundChrom)
errAbort("missing chrom=<name> specification on variableStep declaration at line %lu", lineCount);
variableStep = TRUE;
bedData = FALSE;
fixedStep = FALSE;
freez(&prevChromName);
prevChromName = cloneString(chromName);
continue; /* !!! go to next input line */
}
else if (sameWord("fixedStep",words[0]))
{
boolean foundChrom = FALSE;
boolean foundStart = FALSE;
int i;
/* same comment as above */
if (variableStep || bedData || fixedStep)
{
output_row();
validLines = 0; /* to cause reset for first offset */
}
stepSize = 1; /* default step size */
dataSpan = 0; /* this will match step size if not set*/
for(i = 1; i < wordCount; ++i)
{
if (startsWith("chrom",words[i]))
{
setChromName(words[i]);
foundChrom = TRUE;
}
else if (startsWith("start",words[i]))
{
setFixedStart(words[i]);
foundStart = TRUE;
}
else if (startsWith("step",words[i]))
setStepSize(words[i]);
else if (startsWith("span",words[i]))
setDataSpan(words[i]);
else
errAbort("illegal specification on variableStep at line %lu: %s",
lineCount, words[i]);
}
if (dataSpan == 0)
dataSpan = stepSize;
if (!foundChrom)
errAbort("missing chrom=<name> specification on fixedStep declaration at line %lu", lineCount);
if (!foundStart)
errAbort("missing start=<position> specification on fixedStep declaration at line %lu", lineCount);
if (noOverlap && validLines && prevChromName)
{
if (sameWord(prevChromName,chromName) && (fixedStart < chromStart))
errAbort("specified fixedStep chromStart %llu is less than expected next chromStart %llu", fixedStart, chromStart);
}
variableStep = FALSE;
bedData = FALSE;
fixedStep = TRUE;
freez(&prevChromName);
prevChromName = cloneString(chromName);
continue; /* !!! go to next input line */
}
else if (wordCount == 4)
{
/* while in bedData, we do not necessarily need to start a new
* batch unless the chrom name is changing, since dataSpan
* is always 1 for bedData. As above, this may change in
* the future if each bed line specification is talking
* about a different feature.
*/
if (variableStep || fixedStep ||
(bedData && ((prevChromName != (char *)NULL) &&
differentWord(prevChromName,words[0]))))
{
output_row();
validLines = 0; /* to cause reset for first offset */
}
dataSpan = 1; /* default bases spanned per data point */
variableStep = FALSE;
bedData = TRUE;
fixedStep = FALSE;
freez(&chromName);
chromName=cloneString(words[0]);
freez(&featureName);
featureName=cloneString(words[0]);
bedChromStart = sqlLongLong(words[1]);
bedChromEnd = sqlLongLong(words[2]);
bedDataValue = sqlDouble(words[3]);
/* the bed format coordinate system is zero relative, half-open,
* hence, no adjustment of bedChromStart is needed here, unlike the
* fixed and variable step formats which will subtract one from the
* incoming coordinate.
*/
if (bedChromStart >= bedChromEnd)
errAbort("Found chromStart >= chromEnd at line %lu (%llu > %llu)",
lineCount, bedChromStart, bedChromEnd);
if (bedChromEnd > (bedChromStart + 10000000))
errAbort("Limit of 10,000,000 length specification for bed format at line %lu, found: %llu)",
lineCount, bedChromEnd-bedChromStart);
if ((validLines > 0) && (bedChromStart < previousOffset))
errAbort("chrom positions not in numerical order at line %lu. previous: %llu > %llu <-current (bed)", lineCount, previousOffset, bedChromStart);
freez(&prevChromName);
prevChromName = cloneString(chromName);
}
/* We must be in one of these data formats at this point */
if (!(variableStep || fixedStep || bedData))
errAbort("at the line beginning: %s, variableStep or fixedStep data declaration not found or BED data 4 column format not recognized.", words[0]);
if (variableStep && (wordCount != 2))
errAbort("Expecting two words for variableStep data at line %lu, found %d",
lineCount, wordCount);
if (fixedStep && (wordCount != 1))
errAbort("Expecting one word for fixedStep data at line %lu, found %d",
lineCount, wordCount);
if (bedData && (wordCount != 4))
errAbort("Expecting four words for bed format data at line %lu, found %d",
lineCount, wordCount);
++validLines; /* counting good lines of data input */
/* Offset is the incoming specified position for this value,
* fixedStart has already been converted to zero
* relative half open
*/
if (variableStep)
{
Offset = sqlLongLong(words[0]);
Offset = BASE_0(Offset); /* zero relative half open */
dataValue = sqlDouble(words[1]);
}
else if (fixedStep)
{
Offset = fixedStart + (stepSize * (validLines - 1));
dataValue = sqlDouble(words[0]);
}
else if (bedData)
{
Offset = bedChromStart;
dataValue = bedDataValue;
}
if (dataValue > overallUpperLimit) overallUpperLimit = dataValue;
if (dataValue < overallLowerLimit) overallLowerLimit = dataValue;
/* see if this is the first time through, establish chromStart */
if (validLines == 1)
{
chromStart = Offset;
verbose(2, "first offset: %llu\n", chromStart);
}
else if ((validLines > 1) && (Offset <= previousOffset))
errAbort("chrom positions not in numerical order at line %lu. previous: %llu > %llu "
"<-current (offset)", lineCount, BASE_1(previousOffset), BASE_1(Offset));
/* if we are working on a zoom level and the data is not exactly
* spaced according to the span, then we need to put each value
* in its own row in order to keep positioning correct for these
* data values. The number of skipped bases has to be an even
* multiple of dataSpan
*/
readingFrameSlipped = FALSE;
if ((validLines > 1) && (dataSpan > 1))
{
unsigned long long prevEnd = previousOffset + dataSpan;
int skippedBases;
int spansSkipped;
skippedBases = Offset - previousOffset;
if (flagOverlapSpanData && (prevEnd > Offset))
errAbort("ERROR: data points overlapping at input line %lu.\n"
"previous data point position: %s:%llu-%llu overlaps current: %s:%llu-%llu",
lineCount, chromName, BASE_1(previousOffset), prevEnd,
chromName, BASE_1(Offset),Offset+dataSpan);
spansSkipped = skippedBases / dataSpan;
if ((spansSkipped * dataSpan) != skippedBases)
readingFrameSlipped = TRUE;
}
if (readingFrameSlipped)
{
verbose(2, "data not spanning %llu bases, prev: %llu, this: %llu, at line: %lu\n", dataSpan, previousOffset, Offset, lineCount);
output_row();
chromStart = Offset; /* a full reset here */
}
/* Check to see if data is being skipped */
else if ( (validLines > 1) && (Offset > (previousOffset + dataSpan)) )
{
unsigned long long off;
unsigned long long fillSize; /* number of bytes */
verbose(2, "missing data offsets: %llu - %llu\n",
BASE_1(previousOffset),BASE_0(Offset));
/* If we are just going to fill the rest of this bin with
* no data, then may as well stop here. No need to fill
* it with nothing.
*/
fillSize = (Offset - (previousOffset + dataSpan)) / dataSpan;
verbose(2, "filling NO_DATA for %llu bytes\n", fillSize);
if (fillSize + bincount >= binsize)
{
verbose(2, "completing a bin due to NO_DATA for %llu bytes, only %llu - %llu = %llu to go\n", fillSize, binsize, bincount, binsize - bincount);
verbose(2, "Offset: %llu, previousOffset: %llu\n",
Offset, previousOffset);
output_row();
chromStart = Offset; /* a full reset here */
} else {
fillSize = 0;
/* fill missing data with NO_DATA indication */
for (off = previousOffset + dataSpan; off < Offset;
off += dataSpan)
{
++fillSize;
++fileOffset;
++bincount; /* count scores in this bin */
if (bincount >= binsize) break;
}
verbose(2, "filled NO_DATA for %llu bytes\n", fillSize);
/* If that finished off this bin, output it
* This most likely should not happen here. The
* check above: if (fillSize + bincount >= binsize)
* should have caught this case already.
*/
if (bincount >= binsize)
{
output_row();
chromStart = Offset; /* a full reset here */
}
}
}
/* With perhaps the missing data taken care of, back to the
* real data.
*/
if (bedData)
{
unsigned long long bedSize = bedChromEnd - bedChromStart;
for ( ; bedSize > 0; --bedSize )
{
setDataValue(bedDataValue);
Offset += 1;
}
Offset -= 1; /* loop above increments this one too much.
* This Offset is supposed to be the last
* valid chrom position written, not the
* next to be written */
}
else
{
setDataValue(dataValue);
}
previousOffset = Offset; /* remember position for gap calculations */
} /* reading file input loop end */
/* Done with input file, any data points left in this bin ? */
if (bincount)
output_row();
lineFileClose(&lf);
fclose(binout);
fclose(wigout);
freez(&chromName);
freez(&featureName);
freez(&data_values);
freez(&validData);
freez(&wibFileName);
/* return limits if pointers are given */
if (upperLimit)
*upperLimit = overallUpperLimit;
if (lowerLimit)
*lowerLimit = overallLowerLimit;
if (wibSizeLimit > 0)
options->wibSizeLimit = wibSize;
}