void readPairFile(struct lineFile *prf)
/* Read in pairs and initialize clone list */
{
int lineSize, i;
char *line;
char *words[4];
char *names[16];
int wordCount, nameCount;
struct clone *clone;
struct cloneName *cloneName;
while (lineFileNext(prf, &line, &lineSize))
{
wordCount = chopTabs(line,words);
if (wordCount != 3)
errAbort("Bad line %d of %s\n", prf->lineIx, prf->fileName);
if (!hashLookup(clones, words[2]))
{
clone = createClone(words[2],NULL,NULL);
hashAdd(clones, words[2], clone);
slAddHead(&cloneList,clone);
}
AllocVar(cloneName);
sprintf(cloneName->name, "%s", words[2]);
nameCount = chopCommas(words[0],names);
for (i = 0; i < nameCount; i++)
hashAdd(leftNames, names[i], cloneName);
nameCount = chopCommas(words[1],names);
for (i = 0; i < nameCount; i++)
hashAdd(rightNames, names[i], cloneName);
}
}
int findBedSize(char *fileName, struct lineFile **retLf)
/* Read first line of file and figure out how many words in it. */
/* Input file could be stdin, in which case we really don't want to open,
* read, and close it here. So if retLf is non-NULL, return the open
* linefile (having told it to reuse the line we just read). */
{
struct lineFile *lf = lineFileOpen(fileName, TRUE);
char *words[64], *line;
int wordCount;
if (!lineFileNextReal(lf, &line))
if (ignoreEmpty)
return(0);
line = cloneString(line);
if (strictTab)
wordCount = chopTabs(line, words);
else
wordCount = chopLine(line, words);
if (wordCount == 0)
errAbort("%s appears to be empty", fileName);
if (retLf != NULL)
{
lineFileReuse(lf);
*retLf = lf;
}
else
lineFileClose(&lf);
freeMem(line);
return wordCount;
}
void readHugoMultiTable(char *fileName, struct hugoMulti **retList,
struct hash **retIdHash, struct hash **retSymbolHash)
/* Read in file into list and hashes. Make hash keyed on omim ID
* and on OMIM symbol. */
{
struct hash *idHash = newHash(0);
struct hash *symbolHash = newHash(0);
struct hugoMulti *list = NULL, *el;
struct lineFile *lf = lineFileOpen(fileName, TRUE);
char *words[16];
char *line;
int lineSize, wordCount;
char *name;
while (lineFileNext(lf, &line, &lineSize))
{
if (line[0] == 0 || line[0] == '#')
continue;
wordCount = chopTabs(line, words);
lineFileExpectWords(lf, 11, wordCount);
el = hugoMultiLoad(words);
slAddHead(&list, el);
name = el->omimId;
if (name[0] != 0)
hashAdd(idHash, name, el);
name = el->symbol;
if (name[0] != 0)
hashAdd(symbolHash, name, el);
}
lineFileClose(&lf);
slReverse(&list);
*retList = list;
*retIdHash = idHash;
*retSymbolHash = symbolHash;
}
struct psl *nextLmPsl(struct lineFile *lf, struct lm *lm)
/* Read next line from file and convert it to psl. Return
* NULL at eof. */
{
char *line;
int lineSize;
char *words[32];
int wordCount;
if (!lineFileNext(lf, &line, &lineSize))
return NULL;
wordCount = chopTabs(line, words);
if (wordCount == 21)
{
return pslLoadLm(words, lm);
}
else if (wordCount == 23)
{
return pslxLoadLm(words, lm);
}
else
{
warn("Bad line %d of %s", lf->lineIx, lf->fileName);
return NULL;
}
}
struct psl *nextPsl(struct lineFile *lf)
/* Read next line from file and convert it to psl. Return
* NULL at eof. */
{
char *line;
int lineSize;
char *words[32];
int wordCount;
if (!lineFileNext(lf, &line, &lineSize))
{
//warn("File %s appears to be incomplete\n", lf->fileName);
return NULL;
}
wordCount = chopTabs(line, words);
if (wordCount == 21)
{
return pslLoad(words);
}
else if (wordCount == 23)
{
return pslxLoad(words);
}
else
{
warn("Bad line %d of %s", lf->lineIx, lf->fileName);
return NULL;
}
}
boolean mgcStatusTblCopyRow(struct lineFile *inLf, FILE *outFh)
/* read a copy one row of a status table tab file without
* fully parsing. Expand if optional fields are missing */
{
char *line;
int numCols, i;
char *row[MGCSTATUS_NUM_COLS];
if (!lineFileNextReal(inLf, &line))
return FALSE;
numCols = chopTabs(line, row);
numCols = min(numCols, MGCSTATUS_NUM_COLS);
lineFileExpectAtLeast(inLf, MGCSTATUS_MIN_NUM_COLS, numCols);
for (i = 0; i < numCols; i++)
{
if (i > 0)
fputc('\t', outFh);
fputs(row[i], outFh);
}
/* pad */
for (; i < MGCSTATUS_NUM_COLS; i++)
fputc('\t', outFh);
fputc('\n', outFh);
return TRUE;
}
void readPslFile(struct lineFile *pf)
/* Process all records in a psl file of mRNA alignments */
{
int lineSize;
char *line;
char *words[32];
int wordCount;
struct psl *psl;
struct clone *clone;
struct pslAli *pa = NULL;
struct cloneName *cloneName;
while (lineFileNext(pf, &line, &lineSize))
{
wordCount = chopTabs(line, words);
if (wordCount != 21)
errAbort("Bad line %d of %s\n", pf->lineIx, pf->fileName);
psl = pslLoad(words);
if (hashLookup(leftNames, psl->qName))
cloneName = hashMustFindVal(leftNames, psl->qName);
else if (hashLookup(rightNames, psl->qName))
cloneName = hashMustFindVal(rightNames, psl->qName);
else
continue;
clone = hashMustFindVal(clones, cloneName->name);
if ((psl->tBaseInsert < TINSERT) && ((!NORANDOM) || (strlen(psl->tName) < 7)))
{
pa = createPslAli(psl);
if (hashLookup(leftNames, psl->qName))
slAddHead(&(clone->end1), pa);
else
slAddHead(&(clone->end2), pa);
}
}
}
void fixGdup(char *inName, char *outName)
/* fixGdup - Reformat genomic dups table a little.. */
{
struct lineFile *lf = lineFileOpen(inName, TRUE);
FILE *f = mustOpen(outName, "w");
int wordCount, lineSize;
char *words[32], *line;
int i;
while (lineFileNext(lf, &line, &lineSize))
{
if (line[0] == '#')
continue;
wordCount = chopTabs(line, words);
if (wordCount == 0)
continue;
lineFileExpectWords(lf, 15, wordCount);
for (i=0; i<3; ++i)
fprintf(f, "%s\t", words[i]);
fprintf(f, "%s:%s\t", words[6], words[7]);
for (i=4; i<9; ++i)
fprintf(f, "%s\t", words[i]);
for (i=10; i<wordCount; ++i)
{
fprintf(f, "%s", words[i]);
if (i == wordCount-1)
fprintf(f, "\n");
else
fprintf(f, "\t");
}
}
}
void gffFileAdd(struct gffFile *gff, char *fileName, int baseOffset)
/* Create a gffFile structure from a GFF file. */
{
/* Open file and do basic allocations. */
struct lineFile *lf = lineFileOpen(fileName, TRUE);
char *line, *words[9];
int lineSize, wordCount;
while (lineFileNext(lf, &line, &lineSize))
{
if (line[0] != '#')
{
wordCount = chopTabs(line, words);
if (wordCount > 0)
gffFileAddRow(gff, baseOffset, words, wordCount, lf->fileName, lf->lineIx);
}
}
slReverse(&gff->lineList);
slReverse(&gff->seqList);
slReverse(&gff->sourceList);
slReverse(&gff->featureList);
slReverse(&gff->groupList);
slReverse(&gff->geneIdList);
lineFileClose(&lf);
}
void writeBedTab(char *fileName, struct bedStub *bedList, int bedSize)
/* Write out bed list to tab-separated file. */
{
struct bedStub *bed;
FILE *f = mustOpen(fileName, "w");
char *words[64];
int i, wordCount;
for (bed = bedList; bed != NULL; bed = bed->next)
{
if (!noBin)
if (fprintf(f, "%u\t", hFindBin(bed->chromStart, bed->chromEnd)) <= 0)
writeFailed(fileName);
if (strictTab)
wordCount = chopTabs(bed->line, words);
else
wordCount = chopLine(bed->line, words);
for (i=0; i<wordCount; ++i)
{
/* new definition for old "reserved" field, now itemRgb */
/* and when itemRgb, it is a comma separated string r,g,b */
if (itemRgb && (i == 8))
{
char *comma;
/* Allow comma separated list of rgb values here */
comma = strchr(words[8], ',');
if (comma)
{
int itemRgb = 0;
if (-1 == (itemRgb = bedParseRgb(words[8])))
errAbort("ERROR: expecting r,g,b specification, "
"found: '%s'", words[8]);
else
if (fprintf(f, "%d", itemRgb) <= 0)
writeFailed(fileName);
verbose(2, "itemRgb: %s, rgb: %#x\n", words[8], itemRgb);
}
else
if (fputs(words[i], f) == EOF)
writeFailed(fileName);
}
else
if (fputs(words[i], f) == EOF)
writeFailed(fileName);
if (i == wordCount-1)
{
if (fputc('\n', f) == EOF)
writeFailed(fileName);
}
else
if (fputc('\t', f) == EOF)
writeFailed(fileName);
}
}
fclose(f);
}
void pslReps(char *inName, char *bestAliName, char *repName)
/* Analyse inName and put best alignments for eacmRNA in estAliName.
* Put repeat info in repName. */
{
struct lineFile *in = pslFileOpen(inName);
FILE *bestFile = mustOpen(bestAliName, "w");
FILE *repFile = mustOpen(repName, "w");
int lineSize;
char *line;
char *words[32];
int wordCount;
struct psl *pslList = NULL, *psl = NULL;
char lastName[512];
int aliCount = 0;
quiet = sameString(bestAliName, "stdout") || sameString(repName, "stdout");
if (coverQSizeFile != NULL)
loadCoverQSizes(coverQSizeFile);
if (!quiet)
printf("Processing %s to %s and %s\n", inName, bestAliName, repName);
if (!noHead)
pslWriteHead(bestFile);
strcpy(lastName, "");
while (lineFileNext(in, &line, &lineSize))
{
if (((++aliCount & 0x1ffff) == 0) && !quiet)
{
printf(".");
fflush(stdout);
}
wordCount = chopTabs(line, words);
if (wordCount == 21)
psl = pslLoad(words);
else if (wordCount == 23)
psl = pslxLoad(words);
else
errAbort("Bad line %d of %s\n", in->lineIx, in->fileName);
if (!sameString(lastName, psl->qName))
{
doOneAcc(lastName, pslList, bestFile, repFile);
pslFreeList(&pslList);
safef(lastName, sizeof(lastName), "%s", psl->qName);
}
slAddHead(&pslList, psl);
}
doOneAcc(lastName, pslList, bestFile, repFile);
pslFreeList(&pslList);
lineFileClose(&in);
fclose(bestFile);
fclose(repFile);
if (!quiet)
printf("Processed %d alignments\n", aliCount);
}
void processPrimers(struct lineFile *pf, FILE *of)
/* Read and process isPCR file and sts locations */
{
int lineSize, wordCount;
char *line;
char *words[21];
char *dbsts_name, *dbsts[4], *currDbsts;
struct sts *sts=NULL;
struct psl *psl;
struct place *place;
currDbsts = "\0";
while (lineFileNext(pf, &line, &lineSize))
{
wordCount = chopTabs(line, words);
if (wordCount != 21)
errAbort("Bad line %d of %s\n", pf->lineIx, pf->fileName);
psl = pslLoad(words);
dbsts_name = cloneString(psl->qName);
wordCount = chopByChar(dbsts_name, '_', dbsts, ArraySize(dbsts));
if (differentString(dbsts[1], currDbsts))
{
if (sts != NULL)
{
filterPrimersAndWrite(of, sts);
/* stsFree(&sts); */
freez(&currDbsts);
}
currDbsts = cloneString(dbsts[1]);
sts = NULL;
if (hashLookup(stsHash, dbsts[1]))
sts = hashMustFindVal(stsHash, dbsts[1]);
}
if (sts)
{
AllocVar(place);
/* Check if this psl record is already present */
if (!pslInList(place->psl, psl))
{
slAddHead(&place->psl, psl);
place->unali = calcUnali(sts, psl);
place->sizeDiff = calcSizeDiff(sts, psl);
place->badBits = calcBadBits(place);
if (place->sizeDiff < (200 - (place->badBits * 50)))
slAddHead(&sts->place, place);
else
placeFree(&place);
}
}
}
if (sts != NULL)
filterPrimersAndWrite(of, sts);
}
boolean gffHasGtfGroup(char *line)
/* Return TRUE if line has a GTF group field */
{
char *words[10];
char *dupe = cloneString(line);
int wordCt = chopTabs(dupe, words);
boolean isGtf = FALSE;
if (wordCt >= 9)
if (isGtfGroup(words[8]))
isGtf = TRUE;
freeMem(dupe);
return isGtf;
}
char* restField(struct bigBedInterval *bb, int fieldIdx)
/* return a given field from the bb->rest field, NULL on error */
{
if (fieldIdx==0) // we don't return the first(=name) field of bigBed
return NULL;
char *rest = cloneString(bb->rest);
char *restFields[256];
int restCount = chopTabs(rest, restFields);
char *field = NULL;
if (fieldIdx < restCount)
field = cloneString(restFields[fieldIdx]);
freeMem(rest);
return field;
}
struct mgcStatusTbl *mgcStatusTblLoad(char *mgcStatusTab, unsigned opts)
/* Load a mgcStatusTbl object from a tab file */
{
struct mgcStatusTbl *mst = mgcStatusTblNew(opts);
struct lineFile *lf = lineFileOpen(mgcStatusTab, TRUE);
char *line;
char *row[MGCSTATUS_NUM_COLS];
while (lineFileNextReal(lf, &line))
{
int numCols = chopTabs(line, row);
lineFileExpectAtLeast(lf, MGCSTATUS_MIN_NUM_COLS, numCols);
loadRow(mst, lf, row, numCols);
}
lineFileClose(&lf);
return mst;
}
void loadOneBed(struct lineFile *lf, int bedSize, struct bedStub **pList)
/* Load one bed file. Make sure all lines have bedSize fields.
* Put results in *pList. */
{
char *words[64], *line, *dupe;
int wordCount;
struct bedStub *bed;
verbose(1, "Reading %s\n", lf->fileName);
while (lineFileNextReal(lf, &line))
{
if (hasBin)
nextWord(&line);
dupe = cloneString(line);
if (strictTab)
wordCount = chopTabs(line, words);
else
wordCount = chopLine(line, words);
/* ignore empty lines */
if (0 == wordCount)
continue;
lineFileExpectWords(lf, bedSize, wordCount);
AllocVar(bed);
bed->chrom = cloneString(words[0]);
bed->chromStart = lineFileNeedNum(lf, words, 1);
bed->chromEnd = lineFileNeedNum(lf, words, 2);
if (! noStrict)
{
if (bed->chromEnd < 1)
errAbort("ERROR: line %d:'%s'\nchromEnd is less than 1\n",
lf->lineIx, dupe);
if (bed->chromStart == bed->chromEnd && !allowStartEqualEnd)
errAbort("ERROR: line %d:'%s'\nchromStart == chromEnd (%d) (zero-length item)\n"
"Use -allowStartEqualEnd if that is legit (e.g. for insertion point).\n",
lf->lineIx, dupe, bed->chromStart);
if (bed->chromStart > bed->chromEnd)
errAbort("ERROR: line %d:'%s'\nchromStart after chromEnd (%d > %d)\n",
lf->lineIx, dupe, bed->chromStart, bed->chromEnd);
}
bed->line = dupe;
slAddHead(pList, bed);
}
}
void verifyJoinedFormat(char *s)
/* Verify that s consists of lines with two tab-separated fields,
* and that the second field has some n/a and some comma-separated lists. */
{
char *e;
int lineIx = 0;
boolean gotCommas = FALSE, gotNa = FALSE;
while (s != NULL && s[0] != 0)
{
char *row[3];
int fieldCount;
++lineIx;
e = strchr(s, '\n');
if (e != NULL)
*e++ = 0;
if (s[0] != '#')
{
fieldCount = chopTabs(s, row);
if (fieldCount != 2)
{
qaStatusSoftError(tablesTestList->status,
"Got %d fields line %d of joined result, expected 2",
fieldCount, lineIx);
break;
}
if (sameString(row[1], "n/a"))
gotNa = TRUE;
if (countChars(s, ',') >= 2)
gotCommas = TRUE;
}
s = e;
}
if (!gotCommas)
qaStatusSoftError(tablesTestList->status,
"Expected some rows in join to have comma separated lists.");
if (!gotNa)
qaStatusSoftError(tablesTestList->status,
"Expected some rows in joint to have n/a.");
}
void motifFinder(char *database, char *name, int fileCount, char *files[])
/* motifFinder - find largest scoring motif in bed items. */
{
struct sqlConnection *conn = sqlConnect(database);
int fileNum;
char where[256];
struct chromInfo *ci = createChromInfoList(NULL, database);
sqlSafefFrag(where, sizeof(where), "name = '%s'", name);
struct dnaMotif *motif = dnaMotifLoadWhere(conn, motifTable, where);
if(markovTable != NULL)
dnaMotifMakeLog2(motif);
if(motif == NULL)
errAbort("couldn't find motif '%s'", name);
for (fileNum = 0; fileNum < fileCount; fileNum++)
{
char *words[64], *line;
char **row;
struct lineFile *lf = lineFileOpen(files[fileNum], TRUE);
while (lineFileNextReal(lf, &line))
{
int dnaLength, i, j, rowOffset, length, wordCount = chopTabs(line, words);
unsigned chromSize;
boolean markovFound = FALSE;
double mark0[5];
double mark2[5][5][5];
struct dnaSeq *seq = NULL;
char *dupe = NULL;
if (0 == wordCount)
continue;
lineFileExpectAtLeast(lf, 3, wordCount);
dupe = cloneString(line);
char *chrom = words[0];
int chromStart = lineFileNeedNum(lf, words, 1);
if(markovTable != NULL)
chromStart = max(2, chromStart);
unsigned chromEnd = lineFileNeedNum(lf, words, 2);
if (chromEnd < 1)
errAbort("ERROR: line %d:'%s'\nchromEnd is less than 1\n",
lf->lineIx, dupe);
if (chromStart > chromEnd)
errAbort("ERROR: line %d:'%s'\nchromStart after chromEnd (%d > %d)\n",
lf->lineIx, dupe, chromStart, chromEnd);
length = chromEnd - chromStart;
chromSize = getChromSize(ci, chrom);
if(markovTable == NULL)
{
dnaLength = length;
seq = hDnaFromSeq(database, chrom, chromStart, chromEnd, dnaUpper);
if(uniformBackground)
{
int i;
mark0[0] = 1;
for(i = 1; i <= 4; i++)
mark0[i] = 0.25;
}
else
{
dnaMark0(seq, mark0, NULL);
}
}
else
{
dnaLength = length + 4;
if(chromStart - 2 + dnaLength > chromSize)
// can't do analysis for potential peak hanging off the end of the chrom
continue;
seq = hDnaFromSeq(database, chrom, chromStart - 2, chromEnd + 2, dnaUpper);
struct sqlResult *sr = hRangeQuery(conn, markovTable, chrom, chromStart,
chromStart + 1, NULL, &rowOffset);
if((row = sqlNextRow(sr)) != NULL)
{
dnaMark2Deserialize(row[rowOffset + 3], mark2);
dnaMarkMakeLog2(mark2);
markovFound = TRUE;
}
else
errAbort("markov table '%s' is missing; non-markov analysis is current not supported", markovTable);
sqlFreeResult(&sr);
}
struct bed6FloatScore *hits = NULL;
for (i = 0; i < 2; i++)
{
double mark0Copy[5];
char strand = i == 0 ? '+' : '-';
for (j = 0; j <= 4; j++)
mark0Copy[j] = mark0[j];
if(strand == '-')
{
// reverse markov table too!
double tmp;
reverseComplement(seq->dna, dnaLength);
tmp = mark0Copy[1];
mark0Copy[1] = mark0Copy[3];
mark0Copy[3] = tmp;
tmp = mark0Copy[2];
mark0Copy[2] = mark0Copy[4];
mark0Copy[4] = tmp;
}
for (j = 0; j < length - motif->columnCount + 1; j++)
// tricky b/c if(markovFound) then seq includes the two bytes on either side of actual sequence.
{
double score;
if(markovFound)
score = dnaMotifBitScoreWithMarkovBg(motif, seq->dna + j, mark2);
else
score = dnaMotifBitScoreWithMark0Bg(motif, seq->dna + j, mark0Copy);
if(score >= minScoreCutoff)
{
int start;
if(strand == '-')
start = (chromEnd - j) - motif->columnCount;
else
start = chromStart + j;
struct bed6FloatScore *hit = NULL;
// Watch out for overlapping hits (on either strand; yes, I've seen that happen);
// we report only the highest scoring hit in this case.
// O(n^2) where n == number of motifs in a peak, but I expect n to be almost always very small.
if(!originalCoordinates)
{
for (hit = hits; hit != NULL; hit = hit->next)
{
if(hit->chromEnd > start && hit->chromStart <= (start + motif->columnCount))
{
verbose(3, "found overlapping hits: %d-%d overlaps with %d-%d\n", start, start + motif->columnCount, hit->chromStart, hit->chromEnd);
break;
}
}
}
if(hit == NULL || hit->score < score)
{
if(hit == NULL)
{
AllocVar(hit);
slAddHead(&hits, hit);
hit->chrom = cloneString(chrom);
}
hit->chromStart = originalCoordinates ? chromStart : start;
hit->chromEnd = originalCoordinates ? chromEnd : start + motif->columnCount;
hit->score = score;
hit->strand[0] = strand;
}
}
verbose(3, "j: %d; score: %.2f\n", j, score);
}
}
slSort(&hits, bed6FloatCmpDesc);
int count;
float currentPrior = prior;
for(count = 1; hits != NULL; count++, hits = hits->next)
{
if(topOnly && count > topOnly)
break;
// Use a progressively weaker prior for hits with lower scores
verbose(3, "count: %d; score: %.2f; prior: %.2f; log2(prior / (1 - prior)): %.2f\n", count, hits->score, currentPrior, log2(currentPrior / (1 - currentPrior)));
if(hits->score >= minScoreCutoff - log2(currentPrior / (1 - currentPrior)))
{
printf("%s\t%d\t%d\t%s\t%.2f\t%c\n", chrom, originalCoordinates ? chromStart : hits->chromStart,
originalCoordinates ? chromEnd : hits->chromStart + motif->columnCount, name, hits->score, hits->strand[0]);
currentPrior = count == 1 ? priorBackoff : currentPrior * priorBackoff;
if(count > 2)
verbose(3, "hit for count: %d at %s:%d-%d\n", count, chrom, hits->chromStart, hits->chromStart + motif->columnCount);
}
else
break;
}
freeDnaSeq(&seq);
freeMem(dupe);
}
lineFileClose(&lf);
}
sqlDisconnect(&conn);
}
static void bigBedClick(char *fileName, struct trackDb *tdb,
char *item, int start, int end, int bedSize)
/* Handle click in generic bigBed track. */
{
boolean showUrl = FALSE;
char *chrom = cartString(cart, "c");
/* Open BigWig file and get interval list. */
struct bbiFile *bbi = bigBedFileOpen(fileName);
struct lm *lm = lmInit(0);
int ivStart = start, ivEnd = end;
if (start == end)
{
// item is an insertion; expand the search range from 0 bases to 2 so we catch it:
ivStart = max(0, start-1);
ivEnd++;
}
struct bigBedInterval *bbList = bigBedIntervalQuery(bbi, chrom, ivStart, ivEnd, 0, lm);
/* Get bedSize if it's not already defined. */
if (bedSize == 0)
{
bedSize = bbi->definedFieldCount;
showUrl = TRUE;
}
char *scoreFilter = cartOrTdbString(cart, tdb, "scoreFilter", NULL);
int minScore = 0;
if (scoreFilter)
minScore = atoi(scoreFilter);
/* Find particular item in list - matching start, and item if possible. */
boolean found = FALSE;
boolean firstTime = TRUE;
struct bigBedInterval *bb;
for (bb = bbList; bb != NULL; bb = bb->next)
{
if (!(bb->start == start && bb->end == end))
continue;
if (bedSize > 3)
{
char *name = cloneFirstWordByTab(bb->rest);
boolean match = sameString(name, item);
freez(&name);
if (!match)
continue;
}
found = TRUE;
if (firstTime)
printf("<BR>\n");
int seq1Seq2Fields = 0;
// check for seq1 and seq2 in columns 7+8 (eg, pairedTagAlign)
boolean seq1Seq2 = sameOk(trackDbSetting(tdb, BASE_COLOR_USE_SEQUENCE), "seq1Seq2");
if (seq1Seq2 && bedSize == 6)
seq1Seq2Fields = 2;
char *fields[bedSize+seq1Seq2Fields];
char startBuf[16], endBuf[16];
char *rest = cloneString(bb->rest);
int bbFieldCount = bigBedIntervalToRow(bb, chrom, startBuf, endBuf, fields,
bedSize+seq1Seq2Fields);
if (bbFieldCount != bedSize+seq1Seq2Fields)
{
errAbort("Disagreement between trackDb field count (%d) and %s fieldCount (%d)",
bedSize, fileName, bbFieldCount);
}
struct bed *bed = bedLoadN(fields, bedSize);
if (bedSize >= 6 && scoreFilter && bed->score < minScore)
continue;
if (showUrl && (bedSize >= 4))
printCustomUrl(tdb, item, TRUE);
bedPrintPos(bed, bedSize, tdb);
// display seq1 and seq2
if (seq1Seq2 && bedSize+seq1Seq2Fields == 8)
printf("<table><tr><th>Sequence 1</th><th>Sequence 2</th></tr>"
"<tr><td> %s </td><td> %s </td></tr></table>", fields[6], fields[7]);
else if (isNotEmpty(rest))
{
char *restFields[256];
int restCount = chopTabs(rest, restFields);
int restBedFields = bedSize - 3;
if (restCount > restBedFields)
{
if (0 == extraFieldsPrint(tdb,NULL,restFields + restBedFields,restCount - restBedFields))
{
int i;
char label[20];
safef(label, sizeof(label), "nonBedFieldsLabel");
printf("<B>%s </B>",
trackDbSettingOrDefault(tdb, label, "Non-BED fields:"));
for (i = restBedFields; i < restCount; i++)
printf("%s%s", (i > 0 ? "\t" : ""), restFields[i]);
printf("<BR>\n");
}
}
}
if (isCustomTrack(tdb->track))
{
time_t timep = bbiUpdateTime(bbi);
printBbiUpdateTime(&timep);
}
}
if (!found)
{
printf("No item %s starting at %d\n", emptyForNull(item), start);
}
lmCleanup(&lm);
bbiFileClose(&bbi);
}
void processRefSeq(char *database, char *faFile, char *raFile, char *pslFile, char *loc2refFile,
char *pepFile, char *mim2locFile)
/* hgRefSeqMrna - Load refSeq mRNA alignments and other info into
* refSeqGene table. */
{
struct lineFile *lf;
struct hash *raHash, *rsiHash = newHash(0);
struct hash *loc2mimHash = newHash(0);
struct refSeqInfo *rsiList = NULL, *rsi;
char *s, *line, *row[5];
int wordCount, dotMod = 0;
int noLocCount = 0;
int rsiCount = 0;
int noProtCount = 0;
struct psl *psl;
struct sqlConnection *conn = hgStartUpdate(database);
struct hash *productHash = loadNameTable(conn, "productName", 16);
struct hash *geneHash = loadNameTable(conn, "geneName", 16);
char *kgName = "refGene";
FILE *kgTab = hgCreateTabFile(".", kgName);
FILE *productTab = hgCreateTabFile(".", "productName");
FILE *geneTab = hgCreateTabFile(".", "geneName");
FILE *refLinkTab = hgCreateTabFile(".", "refLink");
FILE *refPepTab = hgCreateTabFile(".", "refPep");
FILE *refMrnaTab = hgCreateTabFile(".", "refMrna");
struct exon *exonList = NULL, *exon;
char *answer;
char cond_str[200];
/* Make refLink and other tables table if they don't exist already. */
sqlMaybeMakeTable(conn, "refLink", refLinkTableDef);
sqlUpdate(conn, "NOSQLINJ delete from refLink");
sqlMaybeMakeTable(conn, "refGene", refGeneTableDef);
sqlUpdate(conn, "NOSQLINJ delete from refGene");
sqlMaybeMakeTable(conn, "refPep", refPepTableDef);
sqlUpdate(conn, "NOSQLINJ delete from refPep");
sqlMaybeMakeTable(conn, "refMrna", refMrnaTableDef);
sqlUpdate(conn, "NOSQLINJ delete from refMrna");
/* Scan through locus link to omim ID file and put in hash. */
{
char *row[2];
printf("Scanning %s\n", mim2locFile);
lf = lineFileOpen(mim2locFile, TRUE);
while (lineFileRow(lf, row))
{
hashAdd(loc2mimHash, row[1], intToPt(atoi(row[0])));
}
lineFileClose(&lf);
}
/* Scan through .ra file and make up start of refSeqInfo
* objects in hash and list. */
printf("Scanning %s\n", raFile);
lf = lineFileOpen(raFile, TRUE);
while ((raHash = hashNextRa(lf)) != NULL)
{
if (clDots > 0 && ++dotMod == clDots )
{
dotMod = 0;
dotOut();
}
AllocVar(rsi);
slAddHead(&rsiList, rsi);
if ((s = hashFindVal(raHash, "acc")) == NULL)
errAbort("No acc near line %d of %s", lf->lineIx, lf->fileName);
rsi->mrnaAcc = cloneString(s);
if ((s = hashFindVal(raHash, "siz")) == NULL)
errAbort("No siz near line %d of %s", lf->lineIx, lf->fileName);
rsi->size = atoi(s);
if ((s = hashFindVal(raHash, "gen")) != NULL)
rsi->geneName = cloneString(s);
//!!!else
//!!! warn("No gene name for %s", rsi->mrnaAcc);
if ((s = hashFindVal(raHash, "cds")) != NULL)
parseCds(s, 0, rsi->size, &rsi->cdsStart, &rsi->cdsEnd);
else
rsi->cdsEnd = rsi->size;
if ((s = hashFindVal(raHash, "ngi")) != NULL)
rsi->ngi = atoi(s);
rsi->geneNameId = putInNameTable(geneHash, geneTab, rsi->geneName);
s = hashFindVal(raHash, "pro");
if (s != NULL)
rsi->productName = cloneString(s);
rsi->productNameId = putInNameTable(productHash, productTab, s);
hashAdd(rsiHash, rsi->mrnaAcc, rsi);
freeHashAndVals(&raHash);
}
lineFileClose(&lf);
if (clDots) printf("\n");
/* Scan through loc2ref filling in some gaps in rsi. */
printf("Scanning %s\n", loc2refFile);
lf = lineFileOpen(loc2refFile, TRUE);
while (lineFileNext(lf, &line, NULL))
{
char *mrnaAcc;
if (line[0] == '#')
continue;
wordCount = chopTabs(line, row);
if (wordCount < 5)
errAbort("Expecting at least 5 tab-separated words line %d of %s",
lf->lineIx, lf->fileName);
mrnaAcc = row[1];
mrnaAcc = accWithoutSuffix(mrnaAcc);
if (mrnaAcc[2] != '_')
warn("%s is and odd name %d of %s",
mrnaAcc, lf->lineIx, lf->fileName);
if ((rsi = hashFindVal(rsiHash, mrnaAcc)) != NULL)
{
rsi->locusLinkId = lineFileNeedNum(lf, row, 0);
rsi->omimId = ptToInt(hashFindVal(loc2mimHash, row[0]));
rsi->proteinAcc = cloneString(accWithoutSuffix(row[4]));
}
}
lineFileClose(&lf);
/* Report how many seem to be missing from loc2ref file.
* Write out knownInfo file. */
printf("Writing %s\n", "refLink.tab");
for (rsi = rsiList; rsi != NULL; rsi = rsi->next)
{
++rsiCount;
if (rsi->locusLinkId == 0)
++noLocCount;
if (rsi->proteinAcc == NULL)
++noProtCount;
fprintf(refLinkTab, "%s\t%s\t%s\t%s\t%u\t%u\t%u\t%u\n",
emptyForNull(rsi->geneName),
emptyForNull(rsi->productName),
emptyForNull(rsi->mrnaAcc),
emptyForNull(rsi->proteinAcc),
rsi->geneNameId, rsi->productNameId,
rsi->locusLinkId, rsi->omimId);
}
if (noLocCount)
printf("Missing locusLinkIds for %d of %d\n", noLocCount, rsiCount);
if (noProtCount)
printf("Missing protein accessions for %d of %d\n", noProtCount, rsiCount);
/* Process alignments and write them out as genes. */
lf = pslFileOpen(pslFile);
dotMod = 0;
while ((psl = pslNext(lf)) != NULL)
{
if (hashFindVal(rsiHash, psl->qName) != NULL)
{
if (clDots > 0 && ++dotMod == clDots )
{
dotMod = 0;
dotOut();
}
sqlSafefFrag(cond_str, sizeof cond_str, "extAC='%s'", psl->qName);
answer = sqlGetField(proteinDB, "spXref2", "displayID", cond_str);
if (answer == NULL)
{
fprintf(stderr, "%s NOT FOUND.\n", psl->qName);
fflush(stderr);
}
if (answer != NULL)
{
struct genePred *gp = NULL;
exonList = pslToExonList(psl);
fprintf(kgTab, "%s\t%s\t%c\t%d\t%d\t",
psl->qName, psl->tName, psl->strand[0], psl->tStart, psl->tEnd);
rsi = hashMustFindVal(rsiHash, psl->qName);
gp = genePredFromPsl(psl, rsi->cdsStart, rsi->cdsEnd, genePredStdInsertMergeSize);
if (!gp)
errAbort("Cannot convert psl (%s) to genePred.\n", psl->qName);
fprintf(kgTab, "%d\t%d\t", gp->cdsStart, gp->cdsEnd);
fprintf(kgTab, "%d\t", slCount(exonList));
fflush(kgTab);
for (exon = exonList; exon != NULL; exon = exon->next)
fprintf(kgTab, "%d,", exon->start);
fprintf(kgTab, "\t");
for (exon = exonList; exon != NULL; exon = exon->next)
fprintf(kgTab, "%d,", exon->end);
fprintf(kgTab, "\n");
slFreeList(&exonList);
}
}
else
{
fprintf(stderr, "%s found in psl, but not in .fa or .ra data files.\n", psl->qName);
fflush(stderr);
}
}
if (clDots) printf("\n");
if (!clTest)
{
writeSeqTable(pepFile, refPepTab, FALSE, TRUE);
writeSeqTable(faFile, refMrnaTab, FALSE, FALSE);
}
carefulClose(&kgTab);
carefulClose(&productTab);
carefulClose(&geneTab);
carefulClose(&refLinkTab);
carefulClose(&refPepTab);
carefulClose(&refMrnaTab);
if (!clTest)
{
printf("Loading database with %s\n", kgName);
fflush(stdout);
hgLoadTabFile(conn, ".", kgName, NULL);
printf("Loading database with %s\n", "productName");
fflush(stdout);
hgLoadTabFile(conn, ".", "productName", NULL);
printf("Loading database with %s\n", "geneName");
fflush(stdout);
hgLoadTabFile(conn, ".", "geneName", NULL);
printf("Loading database with %s\n", "refLink");
fflush(stdout);
hgLoadTabFile(conn, ".", "refLink", NULL);
printf("Loading database with %s\n", "refPep");
fflush(stdout);
hgLoadTabFile(conn, ".", "refPep", NULL);
printf("Loading database with %s\n", "refMrna");
fflush(stdout);
hgLoadTabFile(conn, ".", "refMrna", NULL);
}
}
static void writeBlocks(struct bbiChromUsage *usageList, struct lineFile *lf, struct asObject *as,
int itemsPerSlot, struct bbiBoundsArray *bounds,
int sectionCount, boolean doCompress, FILE *f,
int resTryCount, int resScales[], int resSizes[],
struct bbExIndexMaker *eim, int bedCount,
bits16 fieldCount, bits32 *retMaxBlockSize)
/* Read through lf, writing it in f. Save starting points of blocks (every itemsPerSlot)
* to boundsArray */
{
int maxBlockSize = 0;
struct bbiChromUsage *usage = usageList;
char *line, *row[fieldCount+1];
int lastField = fieldCount-1;
int itemIx = 0, sectionIx = 0;
bits64 blockStartOffset = 0;
int startPos = 0, endPos = 0;
bits32 chromId = 0;
struct dyString *stream = dyStringNew(0);
/* Will keep track of some things that help us determine how much to reduce. */
bits32 resEnds[resTryCount];
int resTry;
for (resTry = 0; resTry < resTryCount; ++resTry)
resEnds[resTry] = 0;
boolean atEnd = FALSE, sameChrom = FALSE;
bits32 start = 0, end = 0;
char *chrom = NULL;
struct bed *bed;
AllocVar(bed);
/* Help keep track of which beds are in current chunk so as to write out
* namedChunks to eim if need be. */
long sectionStartIx = 0, sectionEndIx = 0;
for (;;)
{
/* Get next line of input if any. */
if (lineFileNextReal(lf, &line))
{
/* Chop up line and make sure the word count is right. */
int wordCount;
if (tabSep)
wordCount = chopTabs(line, row);
else
wordCount = chopLine(line, row);
lineFileExpectWords(lf, fieldCount, wordCount);
loadAndValidateBed(row, bedN, fieldCount, lf, bed, as, FALSE);
chrom = bed->chrom;
start = bed->chromStart;
end = bed->chromEnd;
sameChrom = sameString(chrom, usage->name);
}
else /* No next line */
{
atEnd = TRUE;
}
/* Check conditions that would end block and save block info and advance to next if need be. */
if (atEnd || !sameChrom || itemIx >= itemsPerSlot)
{
/* Save stream to file, compressing if need be. */
if (stream->stringSize > maxBlockSize)
maxBlockSize = stream->stringSize;
if (doCompress)
{
size_t maxCompSize = zCompBufSize(stream->stringSize);
// keep around an area of scratch memory
static int compBufSize = 0;
static char *compBuf = NULL;
// check to see if buffer needed for compression is big enough
if (compBufSize < maxCompSize)
{
// free up the old not-big-enough piece
freez(&compBuf); // freez knows bout NULL
// get new scratch area
compBufSize = maxCompSize;
compBuf = needLargeMem(compBufSize);
}
int compSize = zCompress(stream->string, stream->stringSize, compBuf, maxCompSize);
mustWrite(f, compBuf, compSize);
}
else
mustWrite(f, stream->string, stream->stringSize);
dyStringClear(stream);
/* Save block offset and size for all named chunks in this section. */
if (eim != NULL)
{
bits64 blockEndOffset = ftell(f);
bbExIndexMakerAddOffsetSize(eim, blockStartOffset, blockEndOffset-blockStartOffset,
sectionStartIx, sectionEndIx);
sectionStartIx = sectionEndIx;
}
/* Save info on existing block. */
struct bbiBoundsArray *b = &bounds[sectionIx];
b->offset = blockStartOffset;
b->range.chromIx = chromId;
b->range.start = startPos;
b->range.end = endPos;
++sectionIx;
itemIx = 0;
if (atEnd)
break;
}
/* Advance to next chromosome if need be and get chromosome id. */
if (!sameChrom)
{
usage = usage->next;
assert(usage != NULL);
assert(sameString(chrom, usage->name));
for (resTry = 0; resTry < resTryCount; ++resTry)
resEnds[resTry] = 0;
}
chromId = usage->id;
/* At start of block we save a lot of info. */
if (itemIx == 0)
{
blockStartOffset = ftell(f);
startPos = start;
endPos = end;
}
/* Otherwise just update end. */
{
if (endPos < end)
endPos = end;
/* No need to update startPos since list is sorted. */
}
/* Save name into namedOffset if need be. */
if (eim != NULL)
{
bbExIndexMakerAddKeysFromRow(eim, row, sectionEndIx);
sectionEndIx += 1;
}
/* Write out data. */
dyStringWriteOne(stream, chromId);
dyStringWriteOne(stream, start);
dyStringWriteOne(stream, end);
if (fieldCount > 3)
{
int i;
/* Write 3rd through next to last field and a tab separator. */
for (i=3; i<lastField; ++i)
{
char *s = row[i];
dyStringAppend(stream, s);
dyStringAppendC(stream, '\t');
}
/* Write last field and terminal zero */
char *s = row[lastField];
dyStringAppend(stream, s);
}
dyStringAppendC(stream, 0);
itemIx += 1;
/* Do zoom counting. */
for (resTry = 0; resTry < resTryCount; ++resTry)
{
bits32 resEnd = resEnds[resTry];
if (start >= resEnd)
{
resSizes[resTry] += 1;
resEnds[resTry] = resEnd = start + resScales[resTry];
}
while (end > resEnd)
{
resSizes[resTry] += 1;
resEnds[resTry] = resEnd = resEnd + resScales[resTry];
}
}
}
assert(sectionIx == sectionCount);
freez(&bed);
*retMaxBlockSize = maxBlockSize;
}
void liftTabbed(char *destFile, struct hash *liftHash,
int sourceCount, char *sources[],
int ctgWord, int startWord, int endWord,
boolean doubleLift, int ctgWord2, int startWord2, int endWord2,
int startOffset, int strandWord)
/* Generic lift a tab-separated file with contig, start, and end fields.
* If doubleLift is TRUE, also lift second set of coordinated.*/
{
int minFieldCount = max3(startWord, endWord, ctgWord) + 1;
int wordCount, lineSize;
char *words[128], *line, *source;
struct lineFile *lf;
FILE *f = mustOpen(destFile, "w");
int i,j;
int start = 0;
int end = 0;
int start2 = 0;
int end2 = 0;
char *contig, *chrom = NULL, *chrom2 = NULL;
struct liftSpec *spec;
static char buf[1024*16];
char *s;
int len;
struct bedInfo *biList = NULL, *bi;
boolean anyHits = FALSE;
if (doubleLift)
{
int min2 = max3(ctgWord2, startWord2, endWord2);
minFieldCount = max(minFieldCount, min2);
}
for (i=0; i<sourceCount; ++i)
{
source = sources[i];
lf = lineFileOpen(source, TRUE);
verbose(1, "Lifting %s\n", source);
while (lineFileNext(lf, &line, &lineSize))
{
if (line[0] == '#')
continue;
wordCount = chopTabs(line, words);
if (wordCount == 0)
continue;
if (wordCount < minFieldCount)
errAbort("Expecting at least %d words line %d of %s",
minFieldCount, lf->lineIx, lf->fileName);
contig = words[ctgWord];
contig = rmChromPrefix(contig);
if (startWord >= 0)
start = lineFileNeedNum(lf, words, startWord);
if (endWord >= 0)
end = lineFileNeedNum(lf, words, endWord);
spec = findLift(liftHash, contig, lf);
if (spec == NULL)
{
if (how == carryMissing)
chrom = cloneString(contig);
else
continue;
}
else
{
chrom = spec->newName;
if (spec->strand == '-')
{
int s = start - startOffset, e = end;
start = spec->oldSize - e + startOffset;
end = spec->oldSize - s;
if (strandWord >= 0 && strandWord < wordCount)
{
char strand = words[strandWord][0];
if (strand == '+')
words[strandWord] = "-";
else if (strand == '-')
words[strandWord] = "+";
}
}
start += spec->offset;
end += spec->offset;
}
if (doubleLift)
{
contig = words[ctgWord2];
start2 = lineFileNeedNum(lf, words, startWord2);
end2 = lineFileNeedNum(lf, words, endWord2);
spec = findLift(liftHash, contig, lf);
if (spec == NULL)
{
if (how == carryMissing)
chrom2 = cloneString(contig);
else
errAbort("Couldn't find second contig in lift file at line %d of %s\n", lf->lineIx, lf->fileName);
}
else
{
cantHandleSpecRevStrand(spec);
chrom2 = spec->newName;
start2 += spec->offset;
end2 += spec->offset;
}
}
anyHits = TRUE;
s = buf;
for (j=0; j<wordCount; ++j)
{
if (s + 128 >= buf + sizeof(buf))
errAbort("Line %d too long in %s", lf->lineIx, lf->fileName);
if (j != 0)
*s++ = '\t';
if (j == ctgWord)
s += sprintf(s, "%s", chrom);
else if (j == startWord)
s += sprintf(s, "%d", start);
else if (j == endWord)
s += sprintf(s, "%d", end);
else if (doubleLift && j == ctgWord2)
s += sprintf(s, "%s", chrom2);
else if (doubleLift && j == startWord2)
s += sprintf(s, "%d", start2);
else if (doubleLift && j == endWord2)
s += sprintf(s, "%d", end2);
else
s += sprintf(s, "%s", words[j]);
}
*s = 0;
if (nosort)
{
fprintf(f, "%s\n", buf);
}
else
{
len = s-buf;
bi = needMem(sizeof(*bi) + len);
bi->chrom = chrom;
bi->start = start;
bi->end = end;
memcpy(bi->line, buf, len);
slAddHead(&biList, bi);
}
}
lineFileClose(&lf);
if (dots)
verbose(1, "\n");
}
if (!nosort)
{
slSort(&biList, bedInfoCmp);
for (bi = biList; bi != NULL; bi = bi->next)
{
fprintf(f, "%s\n", bi->line);
}
}
if (ferror(f))
errAbort("error writing %s", destFile);
fclose(f);
if (!anyHits)
errAbort("No lines lifted!");
}
void loadOneBed(struct lineFile *lf, int bedSize, struct bedStub **pList)
/* Load one bed file. Make sure all lines have the correct number of fields.
* Put results in *pList. */
{
char *words[64], *line, *dupe;
int wordCount;
struct bedStub *bed;
struct asObject *asObj = getAsObj(bedSize);
int fieldCount = getFieldCount(bedSize, asObj);
struct bed *validateBed;
AllocVar(validateBed);
verbose(1, "Reading %s\n", lf->fileName);
while (lineFileNextReal(lf, &line))
{
if (hasBin)
nextWord(&line);
dupe = cloneString(line);
if (strictTab)
wordCount = chopTabs(line, words);
else
wordCount = chopLine(line, words);
/* ignore empty lines */
if (0 == wordCount)
continue;
lineFileExpectWords(lf, fieldCount, wordCount);
if (type)
// TODO also, may need to add a flag to the validateBed() interface to support -allowNegativeScores when not isCt
// although can probably get away without it since usually -allowNegativeScores is used by ct which has already verified it.
// thus -allowNegativeScores is unlikely to be used with -type.
{
loadAndValidateBed(words, typeBedN, fieldCount, lf, validateBed, asObj, FALSE);
checkChromNameAndSize(lf, validateBed->chrom, validateBed->chromEnd);
}
AllocVar(bed);
bed->chrom = cloneString(words[0]);
bed->chromStart = lineFileNeedNum(lf, words, 1);
bed->chromEnd = lineFileNeedNum(lf, words, 2);
if (! noStrict)
{
if ((bed->chromEnd < 1) && !allowStartEqualEnd)
errAbort("ERROR: line %d:'%s'\nchromEnd is less than 1\n",
lf->lineIx, dupe);
if (bed->chromStart == bed->chromEnd && !allowStartEqualEnd)
errAbort("ERROR: line %d:'%s'\nchromStart == chromEnd (%d) (zero-length item)\n"
"Use -allowStartEqualEnd if that is legit (e.g. for insertion point).\n",
lf->lineIx, dupe, bed->chromStart);
if (bed->chromStart > bed->chromEnd)
errAbort("ERROR: line %d:'%s'\nchromStart after chromEnd (%d > %d)\n",
lf->lineIx, dupe, bed->chromStart, bed->chromEnd);
}
bed->line = dupe;
slAddHead(pList, bed);
}
if (asObj)
asObjectFreeList(&asObj);
freez(&validateBed);
}
void writeBedTab(char *fileName, struct bedStub *bedList)
/* Write out bed list to tab-separated file. */
{
struct bedStub *bed;
FILE *f = mustOpen(fileName, "w");
char *words[64];
int i, wordCount;
for (bed = bedList; bed != NULL; bed = bed->next)
{
if (!noBin)
{
// allow for zero-length at start of seq [bin code can't handle 0-0]
unsigned end = (bed->chromEnd > 0) ? bed->chromEnd : 1;
if (fprintf(f, "%u\t", hFindBin(bed->chromStart, end)) <= 0)
writeFailed(fileName);
}
if (strictTab)
wordCount = chopTabs(bed->line, words);
else
wordCount = chopLine(bed->line, words);
for (i=0; i<wordCount; ++i)
{
/* new definition for old "reserved" field, now itemRgb */
/* and when itemRgb, it is a comma separated string r,g,b */
if (itemRgb && (i == 8))
{
char *comma;
/* Allow comma separated list of rgb values here */
comma = strchr(words[8], ',');
if (comma)
{
int itemRgb = 0;
if (-1 == (itemRgb = bedParseRgb(words[8])))
errAbort("ERROR: expecting r,g,b specification, "
"found: '%s'", words[8]);
else
if (fprintf(f, "%d", itemRgb) <= 0)
writeFailed(fileName);
verbose(2, "itemRgb: %s, rgb: %#x\n", words[8], itemRgb);
}
else
if (fputs(words[i], f) == EOF)
writeFailed(fileName);
}
else if ((dotIsNull > 0) && (dotIsNull == i) && sameString(words[i],"."))
/* If the . was used to represent NULL, replace with -1 in the tables */
{
if (fputs("-1", f) == EOF)
writeFailed(fileName);
}
else
if (fputs(words[i], f) == EOF)
writeFailed(fileName);
if (i == wordCount-1)
{
if (fputc('\n', f) == EOF)
writeFailed(fileName);
}
else
if (fputc('\t', f) == EOF)
writeFailed(fileName);
}
}
fclose(f);
}
struct tagStorm *idfToStormTop(char *fileName)
/* Convert an idf.txt format file to a tagStorm with a single top-level stanza */
{
/* Create a tag storm with one as yet empty stanza */
struct tagStorm *storm = tagStormNew(fileName);
struct tagStanza *stanza = tagStanzaNew(storm, NULL);
/* Some stuff to help turn File_Data1, File_Data2, etc to a comma separated list */
char *additionalFilePrefix = "idf.Comment_AdditionalFile_Data";
struct dyString *additionalFileDy = dyStringNew(0);
/* There can be multiple secondary accession tags, so handle these too */
char *secondaryAccessionTag = "idf.Comment_SecondaryAccession";
struct dyString *secondaryAccessionDy = dyStringNew(0);
/* Parse lines from idf file into stanza */
struct lineFile *lf = lineFileOpen(fileName, TRUE);
char *line;
struct dyString *dyVal = dyStringNew(0);
while (lineFileNextReal(lf, &line))
{
/* Erase trailing tab... */
eraseTrailingSpaces(line);
/* Parse line into tab-separated array and make sure it's a reasonable size */
char *row[256];
int rowSize = chopTabs(line, row);
if (rowSize == ArraySize(row))
errAbort("Line %d of %s has too many fields", lf->lineIx, lf->fileName);
if (rowSize < 2)
continue;
/* Convert first element to tagName */
char tagName[256];
aeFieldToNormalField("idf.", trimSpaces(row[0]), tagName, sizeof(tagName));
/* Special case where we already are a comma separated list */
if (sameString(tagName, "idf.Publication_Author_List"))
{
tagStanzaAppend(storm, stanza, tagName, row[1]);
}
else if (startsWith(additionalFilePrefix, tagName))
{
csvEscapeAndAppend(additionalFileDy, row[1]);
}
else if (sameString(secondaryAccessionTag, tagName))
{
csvEscapeAndAppend(secondaryAccessionDy, row[1]);
}
else
{
/* Convert rest of elements to possibly comma separated values */
dyStringClear(dyVal);
int i;
for (i=1; i<rowSize; ++i)
csvEscapeAndAppend(dyVal, row[i]);
tagStanzaAppend(storm, stanza, tagName, dyVal->string);
}
}
if (additionalFileDy->stringSize != 0)
tagStanzaAppend(storm, stanza, additionalFilePrefix, additionalFileDy->string);
if (secondaryAccessionDy->stringSize != 0)
tagStanzaAppend(storm, stanza, secondaryAccessionTag, secondaryAccessionDy->string);
dyStringFree(&secondaryAccessionDy);
dyStringFree(&additionalFileDy);
dyStringFree(&dyVal);
lineFileClose(&lf);
return storm;
}
