int makeMilli(char *s, struct lineFile *lf)
/* Convert ascii floating point to parts per thousand representation. */
{
/* Cope with -0.0 and -0.2 etc.*/
if (s[0] == '-')
{
if (!sameString(s, "-0.0"))
warn("Strange perc. field %s line %d of %s", s, lf->lineIx, lf->fileName);
s = "0.0";
}
if (!isdigit(s[0]))
badFormat(lf,1);
return round(10.0*atof(s));
}
// checkFormat: Checks the format of the input and prints an
// error and exits if one is found.
int checkFormat(char s[MAXS][MAXLEN]) {
int state;
int ops = 0, nums = 0;
int exp_nums = 0;
if(si == 0) {
badFormat();
}
if(si == 1) {
if(isNumber(s[0]) == 0) {
return 0;
}
else {
badFormat();
}
}
// Start by counting the number of operators
state = OPSTATE;
for(int i = si - 1; i >= 0; i--) {
if(isNumber(s[i]) != 0 && isOp(s[i]) != 0) {
badFormat();
}
switch(state) {
case OPSTATE:
if(isOp(s[i]) == 0) {
ops++;
state = OPSTATE;
}
else {
exp_nums = ops + 1;
nums++;
state = NUMSTATE;
}
break;
case NUMSTATE:
if(isNumber(s[i]) == 0) {
nums++;
state = NUMSTATE;
}
else {
badFormat();
}
break;
default:
badFormat();
}
}
if(nums != exp_nums) {
badFormat();
}
return 0;
}
void makeMotifs(char *inFile, struct hash *tfHash, char *outFile)
/* Parse input motifs and save them to outFile in dnaMotif format. */
{
struct lineFile *lf = lineFileOpen(inFile, TRUE);
FILE *f = mustOpen(outFile, "w");
struct hashEl *hel;
for (;;)
{
char *line;
char *words[256], *word;
int wordCount;
struct dnaMotif *motif;
if (!lineFileSkipTo(lf, "Probability matrix for"))
break;
lineFileNeedNext(lf, &line, NULL);
wordCount = chopLine(line, words);
if (wordCount >= ArraySize(words))
errAbort("Line %d of %s is too long\n", lf->lineIx, lf->fileName);
if (!sameString(words[0], "#"))
badFormat(lf);
AllocVar(motif);
motif->columnCount = wordCount-1;
readBaseProbs(lf, words, "#A", &motif->aProb, motif->columnCount);
readBaseProbs(lf, words, "#C", &motif->cProb, motif->columnCount);
readBaseProbs(lf, words, "#T", &motif->tProb, motif->columnCount);
readBaseProbs(lf, words, "#G", &motif->gProb, motif->columnCount);
if (!lineFileSkipTo(lf, "Source:"))
lineFileUnexpectedEnd(lf);
lineFileReuse(lf);
lineFileNeedNext(lf, &line, NULL);
word = nextWord(&line);
word = nextWord(&line);
if (word == NULL)
errAbort("Short Source: line %d of %s", lf->lineIx, lf->fileName);
motif->name = cloneString(word);
hel = hashLookup(tfHash, motif->name);
if (hel == NULL)
errAbort("%s in %s but not GFFs", motif->name, lf->fileName);
hel->val = motif;
dnaMotifTabOut(motif, f);
}
carefulClose(&f);
lineFileClose(&lf);
}
void wigAsciiToBinary( int argc, char *argv[] )
{
int i = 0; /* general purpose int counter */
struct lineFile *lf; /* for line file utilities */
char * fileName; /* the basename of the input file */
char *line = (char *) NULL; /* to receive data input line */
char *words[4]; /* to split data input line */
int wordCount = 0; /* result of split */
int validLines = 0; /* counting only lines with data */
unsigned long long previousOffset = 0; /* for data missing detection */
double dataValue = 0.0; /* from data input */
char *wigfile = (char *) NULL; /* file name of wiggle database file */
boolean firstInChrom; /* Is this the first line in chromosome? */
/* for each input data file */
for (i = 1; i < argc; ++i)
{
verbose(2, "translating file: %s\n", argv[i]);
fileName = basename(argv[i]);
if (name) /* Is the name of this feature specified ? */
{
safef( featureName, sizeof(featureName) - 1, "%s", name);
}
if (chrom) /* Is the chrom name specified ? */
{
chromName = cloneString(chrom);
if (! name) /* that names the feature too if not already */
safef( featureName, sizeof(featureName) - 1, "%s", chrom);
}
/* Name mangling to determine output file name */
if (wibFile) /* when specified, simply use it */
{
binfile = addSuffix(wibFile, ".wib");
wigfile = addSuffix(wibFile, ".wig");
} else { /* not specified, construct from input names */
if (startsWith("chr",fileName))
{
char *tmpString;
tmpString = cloneString(fileName);
chopSuffix(tmpString);
binfile = addSuffix(tmpString, ".wib");
wigfile = addSuffix(tmpString, ".wig");
if (! chrom) /* if not already taken care of */
chromName = cloneString(tmpString);
if (! name && ! chrom) /* if not already done */
safef(featureName, sizeof(featureName) - 1, "%s", tmpString);
freeMem(tmpString);
} else {
errAbort("Can not determine output file name, no -wibFile specified\n");
}
}
verbose(2, "output files: %s, %s\n", binfile, wigfile);
validLines = 0; /* to count only lines with data */
rowCount = 0; /* to count rows output */
bincount = 0; /* to count up to binsize */
fileOffset = 0; /* current location within binary data file */
fileOffsetBegin = 0;/* location in binary data file where this bin starts*/
firstInChrom = TRUE;
freeMem(data_values);
freeMem(validData);
data_values = needMem( (size_t) (binsize * sizeof(double)));
validData = needMem( (size_t) (binsize * sizeof(unsigned char)));
overallLowerLimit = 1.0e+300; /* for the complete set of data */
overallUpperLimit = -1.0e+300; /* for the complete set of data */
binout = mustOpen(binfile,"w"); /* binary data file */
wigout = mustOpen(wigfile,"w"); /* table row definition file */
lf = lineFileOpen(argv[i], TRUE); /* input file */
while (lineFileNextReal(lf, &line))
{
boolean readingFrameSlipped;
char *valEnd;
char *val;
++validLines;
wordCount = chopByWhite(line, words, ArraySize(words));
if (wordCount == 1)
{
Offset += 1;
val = words[0];
}
else if (wordCount == 2)
{
Offset = atoll(words[0]) - 1;
val = words[1];
}
else if (wordCount == 3)
{
char *newChrom = words[0];
boolean sameChrom = (chromName == NULL || sameString(chromName, newChrom));
Offset = atoll(words[1]) - 1;
val = words[2];
if (!sameChrom)
{
output_row();
firstInChrom = TRUE;
freez(&chromName);
}
if (chromName == NULL)
chromName = cloneString(newChrom);
}
else
{
val = NULL;
badFormat(lf);
}
if (Offset < 0)
errAbort("Illegal offset %llu at line %d of %s", Offset+1, lf->lineIx,
lf->fileName);
dataValue = strtod(val, &valEnd);
if(trimVals)
{
dataValue = max(minVal, dataValue);
dataValue = min(maxVal, dataValue);
}
if ((*val == '\0') || (*valEnd != '\0'))
errAbort("Not a valid float at line %d: %s\n", lf->lineIx, val);
/* see if this is the first time through, establish chromStart */
if (firstInChrom) {
chromStart = Offset;
verbose(2, "first offset: %llu\n", chromStart);
}
else if (!firstInChrom && (Offset <= previousOffset))
errAbort("ERROR: chrom positions not in order. line %d of %s\n"
"previous: %llu >= %llu <-current",
lf->lineIx, lf->fileName, previousOffset+1, Offset+1);
/* 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 (!firstInChrom && (dataSpan > 1))
{
int skippedBases;
int spansSkipped;
skippedBases = Offset - previousOffset;
spansSkipped = skippedBases / dataSpan;
if ((spansSkipped * dataSpan) != skippedBases)
readingFrameSlipped = TRUE;
}
if (readingFrameSlipped)
{
verbose(2, "data not spanning %llu bases, prev: %llu, this: %llu, at line: %d\n", dataSpan, previousOffset, Offset, lf->lineIx);
output_row();
chromStart = Offset; /* a full reset here */
}
/* Check to see if data is being skipped */
else if ( (!firstInChrom) && (Offset > (previousOffset + dataSpan)) )
{
unsigned long long off;
unsigned long long fillSize; /* number of bytes */
verbose(2, "missing data offsets: %llu - %llu\n",
previousOffset+1,Offset-1);
/* 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 at bincount: %llu\n", fillSize, bincount);
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 at bincount: %llu\n", fillSize, bincount);
/* 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.
*/
data_values[bincount] = dataValue;
validData[bincount] = TRUE;
++fileOffset;
++bincount; /* count scores in this bin */
/* Is it time to output a row definition ? */
if (bincount >= binsize)
{
output_row();
}
previousOffset = Offset;
firstInChrom = FALSE;
} /* reading file input loop end */
/* Done with input file, any data points left in this bin ? */
if (bincount)
{
output_row();
}
verbose(2, "fini: %s, read %d lines, table rows: %llu, data bytes: %lld\n",
argv[i], lf->lineIx, rowCount, fileOffset);
verbose(1, "data limits: [%g:%g], range: %g\n",
overallLowerLimit, overallUpperLimit,
overallUpperLimit - overallLowerLimit);
lineFileClose(&lf);
fclose(binout);
fclose(wigout);
freeMem(binfile);
freeMem(wigfile);
freeMem(chromName);
binfile = (char *) NULL;
wigfile = (char *) NULL;
chromName = (char *) NULL;
}
return;
}
