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 bedMergeOverlappingBlocks(char *inBed, char *outBed)
/* bedMergeOverlappingBlocks - Fix faulty BED 12 files with illegal overlapping blocks. Also reports a summary of the changes.. */
{
int badBeds = 0;
FILE *log = NULL;
FILE *newBedFile = mustOpen(outBed, "w");
char *logName = optionVal("report", NULL);
struct lineFile *lf = lineFileOpen(inBed, TRUE);
char *line, *row[12];
boolean isItemRgb = FALSE;
if (logName)
log = mustOpen(logName, "w");
while (lineFileNext(lf, &line, NULL))
{
struct bed *bed;
int numFields = chopByWhite(line, row, ArraySize(row));
/* strange it's reading empty lines... whatever */
if (numFields == 0)
continue;
if (numFields < 12)
errAbort("file %s doesn't appear to be in blocked-bed format. At least 12 fields required, got %d", inBed, numFields);
if (bedParseRgb(row[8]))
isItemRgb = TRUE;
bed = bedLoadN(row, numFields);
badBeds += fixBed(bed, lf->lineIx, log);
if (isItemRgb)
bedTabOutNitemRgb(bed, numFields, newBedFile);
else
bedTabOutN(bed, numFields, newBedFile);
}
lineFileClose(&lf);
if (log)
{
fprintf(log, "Fixed %d bad beds in all.\n", badBeds);
carefulClose(&log);
}
carefulClose(&newBedFile);
}
struct encodePeak *encodePeakLineFileLoad(char **row, enum encodePeakType pt, struct lineFile *lf)
/* From a linefile line, load an encodePeak row. Errors outputted */
/* have line numbers, etc. Does more error checking as well. */
{
struct encodePeak *peak;
if (!pt)
errAbort("Unknown peak type set for track");
AllocVar(peak);
peak->chrom = cloneString(row[0]);
peak->chromStart = lineFileNeedNum(lf, row, 1);
peak->chromEnd = lineFileNeedNum(lf, row, 2);
peak->peak = -1;
if (peak->chromEnd < 1)
lineFileAbort(lf, "chromEnd less than 1 (%d)", peak->chromEnd);
if (peak->chromEnd < peak->chromStart)
lineFileAbort(lf, "chromStart after chromEnd (%d > %d)",
peak->chromStart, peak->chromEnd);
peak->name = cloneString(row[3]);
peak->score = lineFileNeedNum(lf, row, 4);
safecpy(peak->strand, sizeof(peak->strand), row[5]);
if (peak->strand[0] != '+' && peak->strand[0] != '-' && peak->strand[0] != '.')
lineFileAbort(lf, "Expecting +, -, or . in strand");
if (pt != gappedPeak)
/* deal with signalValue, pValue, qValue, and peak */
{
peak->signalValue = (float)lineFileNeedDouble(lf, row, 6);
peak->pValue = (float)lineFileNeedDouble(lf, row, 7);
peak->qValue = (float)lineFileNeedDouble(lf, row, 8);
if ((pt == narrowPeak) || (pt == encodePeak))
{
peak->peak = lineFileNeedNum(lf, row, 9);
if (peak->peak >= (int)peak->chromEnd)
lineFileAbort(lf, "peak site past chromEnd (%d > %d)", peak->peak, peak->chromEnd);
}
}
else /* must be gappedPeak */
/* deal with thickStart, thickEnd, itemRgb even though they're not used */
{
int thickStart = lineFileNeedNum(lf, row, 6);
int thickEnd = lineFileNeedNum(lf, row, 7);
int itemRgb = 0;
char *comma;
/* Allow comma separated list of rgb values here */
comma = strchr(row[8], ',');
if (comma)
itemRgb = bedParseRgb(row[8]);
else
itemRgb = lineFileNeedNum(lf, row, 8);
if ((thickStart != 0) || (thickEnd != 0) || (itemRgb != 0))
lineFileAbort(lf, "thickStart, thickEnd, and itemRgb columns not used in gappedPeak type, set all to 0");
}
/* Deal with blocks */
if ((pt == gappedPeak) || (pt == encodePeak))
{
int i, count;
int lastEnd, lastStart;
int blockCountIx, blockSizesIx, blockStartsIx;
if (pt == gappedPeak)
{
blockCountIx = 9;
blockSizesIx = 10;
blockStartsIx = 11;
}
else
{
blockCountIx = 10;
blockSizesIx = 11;
blockStartsIx = 12;
}
peak->blockCount = lineFileNeedNum(lf, row, blockCountIx);
sqlUnsignedDynamicArray(row[blockSizesIx], &peak->blockSizes, &count);
if (count != peak->blockCount)
lineFileAbort(lf, "expecting %d elements in array", peak->blockCount);
sqlUnsignedDynamicArray(row[blockStartsIx], &peak->blockStarts, &count);
if (count != peak->blockCount)
lineFileAbort(lf, "expecting %d elements in array", peak->blockCount);
// tell the user if they appear to be using absolute starts rather than
// relative... easy to forget! Also check block order, coord ranges...
lastStart = -1;
lastEnd = 0;
for (i=0; i < peak->blockCount; i++)
{
if (peak->blockStarts[i]+peak->chromStart >= peak->chromEnd)
{
if (peak->blockStarts[i] >= peak->chromStart)
lineFileAbort(lf,
"BED blockStarts offsets must be relative to chromStart, "
"not absolute. Try subtracting chromStart from each offset "
"in blockStarts.");
else
lineFileAbort(lf,
"BED blockStarts[i]+chromStart must be less than chromEnd.");
}
lastStart = peak->blockStarts[i];
lastEnd = peak->chromStart + peak->blockStarts[i] + peak->blockSizes[i];
}
if (peak->blockStarts[0] != 0)
lineFileAbort(lf,
"BED blocks must span chromStart to chromEnd. "
"BED blockStarts[0] must be 0 (==%d) so that (chromStart + "
"blockStarts[0]) equals chromStart.", peak->blockStarts[0]);
i = peak->blockCount-1;
if ((peak->chromStart + peak->blockStarts[i] + peak->blockSizes[i]) !=
peak->chromEnd)
{
lineFileAbort(lf,
"BED blocks must span chromStart to chromEnd. (chromStart + "
"blockStarts[last] + blockSizes[last]) must equal chromEnd.");
}
}
if (pt == gappedPeak)
/* deal with final three columns of a gappedPeak */
{
peak->signalValue = (float)lineFileNeedDouble(lf, row, 12);
peak->pValue = (float)lineFileNeedDouble(lf, row, 13);
peak->qValue = (float)lineFileNeedDouble(lf, row, 14);
}
return peak;
}
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);
}
