static int match_stext(fz_context *ctx, fz_stext_page *page, const char *s, int n)
{
int orig = n;
int c;
while (*s)
{
s += fz_chartorune(&c, (char *)s);
if (iswhite(c) && iswhite(charat(ctx, page, n)))
{
const char *s_next;
/* Skip over whitespace in the document */
do
n++;
while (iswhite(charat(ctx, page, n)));
/* Skip over multiple whitespace in the search string */
while (s_next = s + fz_chartorune(&c, (char *)s), iswhite(c))
s = s_next;
}
else
{
if (fz_tolower(c) != fz_tolower(charat(ctx, page, n)))
return 0;
n++;
}
}
return n - orig;
}
//notice that this version of counting sort is a bit different because we are
//using strings and we are trying to get the character for the counts via the
//charat function which can return -1 so in order to have that correctly index
//into counts i need to add one more than usual when dereferencing counts
void counting_sort(const std::vector<std::string> & strings,std::vector<int> &v,
std::vector<int> &aux, int start, int end, int chrindx, int counts[])
{
//do counting sort on the l-th column
for(int i = start; i < end; i++) {
int offset = v[i]; //offset of the ith string
const std::string & currstr = strings[offset];
char c = charat(currstr,chrindx);
counts[c + 1]++;
out("%d) strings[%d]=%s counts[%d + 1]=%d \n",i, offset, currstr.c_str(), c, counts[c + 1]);
}
//accumulate the counts
for(int i = 0; i < R+1; i++) {
if(counts[i+1] > 0)
out("counts[%d]=%d+%d\n",i-1, counts[i + 1], counts[i]);
counts[i + 1] += counts[i];
}
for(int i = start; i < end; i++) {
int offset = v[i]; //offset of the ith string
const std::string & currstr = strings[offset];
char c = charat(currstr,chrindx);
aux[counts[c]++ ] = v[i];
out("aux[%d]=%d\n", start + counts[c], v[i]);
}
out("chrindx=%d\n",chrindx);
for(int i = start; i < end; i++) {
v[i] = aux[i - start];
out("%d) strings[%d]=%s\n", i, v[i], strings[v[i]].c_str());
}
}
//pick a pivot, maybe randomly
//put all points above, less than and equal to in separate places of v[i]
//NOTICE: that we dont swap characters or strings, but rather utilize
//a vector of indices which is the one that determines the sorted order
void threeway_quicksort(const std::vector<char *> & strings,
std::vector<int> &v, int start, int end, int chrindx)
{
int n = end - start;
if(n < 2) return;
if(n < 20) { insertion_sort(strings, v, start, end, chrindx); return; }
//get the pivot
int p = rand() % n;
int c = charat(strings[v[start+p]],chrindx);
int lessthan = start;
int gtthan = end;
int i = start;
while(i < gtthan) {
int lc = charat(strings[v[i]],chrindx);
if(lc < c) {
swap(v[lessthan], v[i]);
lessthan++;
i++;
}
else if(lc > c) {
gtthan--;
swap(v[gtthan], v[i]);
}
else { //equal
i++;
}
}
threeway_quicksort(strings, v, start, lessthan, chrindx);
if(c >=0) threeway_quicksort(strings, v, lessthan, gtthan, chrindx + 1);
threeway_quicksort(strings, v, gtthan, end, chrindx);
}
int main()
{
char s[] = "thisistgk";
char c = 's';
printf("the position of char %c int the string is: %d\n", c, charat(s, c));
}
static int
match(fz_text_page *page, const char *s, int n)
{
int orig = n;
int c;
while (*s) {
s += fz_chartorune(&c, (char *)s);
if (c == ' ' && charat(page, n) == ' ') {
while (charat(page, n) == ' ')
n++;
} else {
if (tolower(c) != tolower(charat(page, n)))
return 0;
n++;
}
}
return n - orig;
}
static int match(char *s, fz_text_span *span, int n)
{
int orig = n;
int c;
while ((c = *s++))
{
if (c == ' ' && charat(span, n) == ' ')
{
while (charat(span, n) == ' ')
n++;
}
else
{
if (tolower(c) != tolower(charat(span, n)))
return 0;
n++;
}
}
return n - orig;
}