std::vector<SubstringOccurence> IndexSearcher::simpleSearch(
boost::shared_ptr<HashGenerator> hashGenerator,
boost::shared_ptr<std::vector<sauchar_t> > T,
boost::shared_ptr<std::vector<SUFFIX_MARKER_TYPE> > markers,
boost::shared_ptr<std::vector<saidx_t> > SA,
const std::string & pattern) throw(ConcordiaException) {
std::vector<SubstringOccurence> result;
int left;
std::vector<INDEX_CHARACTER_TYPE> hash =
hashGenerator->generateHash(pattern);
saidx_t patternLength = hash.size()*sizeof(INDEX_CHARACTER_TYPE);
sauchar_t * patternArray = Utils::indexVectorToSaucharArray(hash);
int size = sa_search(T->data(), (saidx_t) T->size(),
(const sauchar_t *) patternArray, patternLength,
SA->data(), (saidx_t) SA->size(), &left);
for (int i = 0; i < size; ++i) {
saidx_t resultPos = SA->at(left + i);
if (resultPos % sizeof(INDEX_CHARACTER_TYPE) == 0) {
// As we are looking for a pattern in an array of higher
// resolution than the hashed index file, we might
// obtain accidental results exceeding the boundaries
// of characters in hashed index. The above check
// removes these accidental results.
saidx_t actualResultPos = resultPos / sizeof(INDEX_CHARACTER_TYPE);
SUFFIX_MARKER_TYPE marker = markers->at(actualResultPos);
result.push_back(SubstringOccurence(marker));
}
}
delete[] patternArray;
return result;
}
static
void
_search_file(const sauchar_t *T, saidx_t Tsize,
const sauchar_t *P, saidx_t Psize,
const saidx_t *SA, saidx_t SAsize,
searchoption_t *option) {
saidx_t size, left;
size = sa_search(T, Tsize, P, Psize, SA, SAsize, &left);
if(0 <= option->maxcount) {
if(option->maxcount == 0) {
return;
}
if(option->maxcount < size) {
size = option->maxcount;
}
}
option->func(T, Tsize, P, Psize, SA, SAsize, left, size, option);
}
int
main(int argc, const char *argv[]) {
FILE *fp;
const char *P;
sauchar_t *T;
saidx_t *SA;
LFS_OFF_T n;
size_t Psize;
saidx_t i, size, left;
if((argc == 1) ||
(strcmp(argv[1], "-h") == 0) ||
(strcmp(argv[1], "--help") == 0)) { print_help(argv[0], EXIT_SUCCESS); }
if(argc != 4) { print_help(argv[0], EXIT_FAILURE); }
P = argv[1];
Psize = strlen(P);
/* Open a file for reading. */
#if HAVE_FOPEN_S
if(fopen_s(&fp, argv[2], "rb") != 0) {
#else
if((fp = LFS_FOPEN(argv[2], "rb")) == NULL) {
#endif
fprintf(stderr, "%s: Cannot open file `%s': ", argv[0], argv[2]);
perror(NULL);
exit(EXIT_FAILURE);
}
/* Get the file size. */
if(LFS_FSEEK(fp, 0, SEEK_END) == 0) {
n = LFS_FTELL(fp);
rewind(fp);
if(n < 0) {
fprintf(stderr, "%s: Cannot ftell `%s': ", argv[0], argv[2]);
perror(NULL);
exit(EXIT_FAILURE);
}
} else {
fprintf(stderr, "%s: Cannot fseek `%s': ", argv[0], argv[2]);
perror(NULL);
exit(EXIT_FAILURE);
}
/* Allocate 5n bytes of memory. */
T = (sauchar_t *)malloc((size_t)n * sizeof(sauchar_t));
SA = (saidx_t *)malloc((size_t)n * sizeof(saidx_t));
if((T == NULL) || (SA == NULL)) {
fprintf(stderr, "%s: Cannot allocate memory.\n", argv[0]);
exit(EXIT_FAILURE);
}
/* Read n bytes of data. */
if(fread(T, sizeof(sauchar_t), (size_t)n, fp) != (size_t)n) {
fprintf(stderr, "%s: %s `%s': ",
argv[0],
(ferror(fp) || !feof(fp)) ? "Cannot read from" : "Unexpected EOF in",
argv[2]);
perror(NULL);
exit(EXIT_FAILURE);
}
fclose(fp);
/* Open the SA file for reading. */
#if HAVE_FOPEN_S
if(fopen_s(&fp, argv[3], "rb") != 0) {
#else
if((fp = LFS_FOPEN(argv[3], "rb")) == NULL) {
#endif
fprintf(stderr, "%s: Cannot open file `%s': ", argv[0], argv[3]);
perror(NULL);
exit(EXIT_FAILURE);
}
/* Read n * sizeof(saidx_t) bytes of data. */
if(fread(SA, sizeof(saidx_t), (size_t)n, fp) != (size_t)n) {
fprintf(stderr, "%s: %s `%s': ",
argv[0],
(ferror(fp) || !feof(fp)) ? "Cannot read from" : "Unexpected EOF in",
argv[3]);
perror(NULL);
exit(EXIT_FAILURE);
}
fclose(fp);
/* Search and print */
size = sa_search(T, (saidx_t)n,
(const sauchar_t *)P, (saidx_t)Psize,
SA, (saidx_t)n, &left);
for(i = 0; i < size; ++i) {
fprintf(stdout, "%" PRIdSAIDX_T "\n", SA[left + i]);
}
/* Deallocate memory. */
free(SA);
free(T);
return 0;
}
