static void printTabularHeaderRow(const struct vcfInfoDef *def)
/* Parse the column header parts out of def->description and print as table header row;
* call this only when looksTabular returns TRUE. */
{
regmatch_t substrArr[PATH_LEN];
if (regexMatchSubstr(def->description, COL_DESC_REGEX, substrArr, ArraySize(substrArr)))
{
puts("<TR>");
// Make a copy of the part of def->description that matches the regex,
// then chop by '|' and print out header column tags:
int matchSize = substrArr[0].rm_eo - substrArr[0].rm_so;
char copy[matchSize+1];
safencpy(copy, sizeof(copy), def->description + substrArr[0].rm_so, matchSize);
// Turn '_' into ' ' so description words can wrap inside headers, saving some space
subChar(copy, '_', ' ');
char *words[PATH_LEN];
int descColCount = chopByChar(copy, '|', words, ArraySize(words));
int i;
for (i = 0; i < descColCount; i++)
printf("<TH class='withThinBorder'>%s</TH>", words[i]);
puts("</TR>");
}
else
errAbort("printTabularHeaderRow: code bug, if looksTabular returns true then "
"regex should work here");
}
static char *readHtmlRecursive(char *fileName, char *database)
/* Slurp in an html file. Wherever it contains insertHtmlRegex, recursively slurp that in
* and replace insertHtmlRegex with the contents. */
{
char *html;
readInGulp(fileName, &html, NULL);
if (isEmpty(html))
return html;
regmatch_t substrs[4];
while (regexMatchSubstr(html, insertHtmlRegex, substrs, ArraySize(substrs)))
{
struct dyString *dy = dyStringNew(0);
// All text before the regex match:
dyStringAppendN(dy, html, substrs[0].rm_so);
// Is there an #if before the #insert ?
boolean doInsert = TRUE;
if (substrs[1].rm_so != -1 &&
(! sameStringN(database, html+substrs[2].rm_so, (substrs[2].rm_eo - substrs[2].rm_so))))
doInsert = FALSE;
if (doInsert)
{
// Recursively pull in inserted file contents from relative path, replacing regex match:
char dir[PATH_LEN];
splitPath(fileName, dir, NULL, NULL);
char insertFileName[PATH_LEN+FILENAME_LEN];
safecpy(insertFileName, sizeof(insertFileName), dir);
safencat(insertFileName, sizeof(insertFileName), html+substrs[3].rm_so,
(substrs[3].rm_eo - substrs[3].rm_so));
if (!fileExists(insertFileName))
errAbort("readHtmlRecursive: relative path '%s' (#insert'ed in %s) not found",
insertFileName, fileName);
char *insertedText = readHtmlRecursive(insertFileName, database);
dyStringAppend(dy, insertedText);
freez(&insertedText);
}
// All text after the regex match:
dyStringAppend(dy, html+substrs[0].rm_eo);
freez(&html);
html = dyStringCannibalize(&dy);
}
return html;
}
static void parseMetadataLine(struct vcfFile *vcff, char *line)
/* Parse a VCF header line beginning with "##" that defines a metadata. */
{
char *ptr = line;
if (ptr == NULL && !startsWith(ptr, "##"))
errAbort("Bad line passed to parseMetadataLine");
ptr += 2;
char *firstEq = strchr(ptr, '=');
if (firstEq == NULL)
{
vcfFileErr(vcff, "Metadata line lacks '=': \"%s\"", line);
return;
}
regmatch_t substrs[8];
// Some of the metadata lines are crucial for parsing the rest of the file:
if (startsWith("##fileformat=", line) || startsWith("##format", line))
{
if (regexMatchSubstr(line, fileformatRegex, substrs, ArraySize(substrs)))
{
// substrs[2] is major version #, substrs[3] is set only if there is a minor version,
// and substrs[4] is the minor version #.
vcff->majorVersion = atoi(line + substrs[2].rm_so);
if (substrs[3].rm_so != -1)
vcff->minorVersion = atoi(line + substrs[4].rm_so);
}
else
vcfFileErr(vcff, "##fileformat line does not match expected pattern /%s/: \"%s\"",
fileformatRegex, line);
}
else if (startsWith("##INFO=", line) || startsWith("##FORMAT=", line))
{
boolean isInfo = startsWith("##INFO=", line);
nonAsciiWorkaround(line);
if (regexMatchSubstr(line, infoOrFormatRegex, substrs, ArraySize(substrs)) ||
regexMatchSubstr(line, infoOrFormatRegex3_3, substrs, ArraySize(substrs)))
// substrs[2] is ID/key, substrs[3] is Number, [4] is Type and [5] is Description.
{
struct vcfInfoDef *def = vcfFileAlloc(vcff, sizeof(struct vcfInfoDef));
def->key = vcfFileCloneSubstr(vcff, line, substrs[2]);
char *number = vcfFileCloneSubstr(vcff, line, substrs[3]);
if (sameString(number, ".") || sameString(number, "A") || sameString(number, "G"))
// A is #alts which varies line-to-line; "G" is #genotypes which we haven't
// yet seen. Why is there a G here -- shouldn't such attributes go in the
// genotype columns?
def->fieldCount = -1;
else
def->fieldCount = atoi(number);
def->type = vcfInfoTypeFromSubstr(vcff, line, substrs[4]);
// greedy regex pulls in end quote, trim if found:
if (line[substrs[5].rm_eo-1] == '"')
line[substrs[5].rm_eo-1] = '\0';
def->description = vcfFileCloneSubstr(vcff, line, substrs[5]);
slAddHead((isInfo ? &(vcff->infoDefs) : &(vcff->gtFormatDefs)), def);
}
else
vcfFileErr(vcff, "##%s line does not match expected pattern /%s/ or /%s/: \"%s\"",
(isInfo ? "INFO" : "FORMAT"), infoOrFormatRegex, infoOrFormatRegex3_3, line);
}
else if (startsWith("##FILTER=", line) || startsWith("##ALT=", line))
{
boolean isFilter = startsWith("##FILTER", line);
if (regexMatchSubstr(line, filterOrAltRegex, substrs, ArraySize(substrs)) ||
regexMatchSubstr(line, filterRegex3_3, substrs, ArraySize(substrs)))
{
// substrs[2] is ID/key, substrs[4] is Description.
struct vcfInfoDef *def = vcfFileAlloc(vcff, sizeof(struct vcfInfoDef));
def->key = vcfFileCloneSubstr(vcff, line, substrs[2]);
def->description = vcfFileCloneSubstr(vcff, line, substrs[4]);
slAddHead((isFilter ? &(vcff->filterDefs) : &(vcff->altDefs)), def);
}
else
{
if (isFilter)
vcfFileErr(vcff, "##FILTER line does not match expected pattern /%s/ or /%s/: \"%s\"",
filterOrAltRegex, filterRegex3_3, line);
else
vcfFileErr(vcff, "##ALT line does not match expected pattern /%s/: \"%s\"",
filterOrAltRegex, line);
}
}
}