void dump_trie(Trie *v, int detail)
{
int i;
dump_node(v, detail);
for (i = 0; i < v->nchildren; i++)
dump_trie(&v->children[i], detail);
}
void dump_trie (FILE *fp, int features, int lvl, uint64_t refpos, uint64_t pos, unsigned char *buffer, unsigned char *bp)
{
uint64_t son, curpos, cnt;
int len, end;
unsigned char *cp;
if (!pos) return;
curpos = pos;
trie_node_read (fp, features, &refpos, &curpos, &cnt, &len, &end, &son, bp);
if (!len) return;
printf ("%010ld", (long)pos);
printf (" %s", (end ? "*" : " "));
for (cp = buffer; cp < bp; cp++) putchar (*cp == '\t' ? '\t' : ' ');
fwrite (bp, len, 1, stdout);
printf ("\n");
dump_trie (fp, features, lvl + 1, son, son, buffer, bp + len);
dump_trie (fp, features, lvl, refpos, curpos, buffer, bp);
}
void serialize_trie(const std::string& filename, Trie* root)
{
auto file = fopen(filename.c_str(), "w+b");
auto offset_map = std::map<Trie*, size_t>();
fseek(file, 0, SEEK_SET);
unsigned root_adr = 0;
fwrite(&root_adr, sizeof(unsigned), 1, file);
dump_trie(root, file, offset_map, sizeof(unsigned));
root_adr = offset_map[root];
fseek(file, 0, SEEK_SET);
fwrite(&root_adr, sizeof(unsigned), 1, file);
fclose(file);
}
int main (int argc, char *argv[])
{
FILE *fp;
uint64_t pos;
char *trie = NULL;
int version, features;
static unsigned char buffer[1024];
for (;;) {
int option_index = 0;
int c;
static struct option long_options[] = {
{"trie", required_argument, 0, 't'},
{0, 0, 0, 0}
};
c = getopt_long(argc, argv, "t:",
long_options, &option_index);
if (c == -1)
break;
switch (c) {
case 't':
trie = optarg;
break;
default:
printf("?? getopt returned character code 0%o ??\n", c);
}
}
if (optind < argc) error (argv[0], "Excess argument");
if (trie == NULL) error (argv[0], "Missing argument --trie");
fp = fopen (trie, "r");
if (fp == NULL) error (argv[0], "Unable to open file \"%s\"\n", argv[1]);
pos = trie_hdr_read (fp, &version, &features);
dump_trie (fp, features, 0, pos, pos, buffer, buffer);
return 0;
}
size_t dump_trie(Trie* trie, FILE* file, std::map<Trie*, size_t>& offset_map,
size_t offset)
{
for (const auto& prefix : *trie)
{
// Recursive call on every child
offset = dump_trie(prefix.second.get(), file, offset_map, offset);
}
Datas d;
d.is_word = trie->is_word();
d.nb_children = trie->nb_children();
offset_map[trie] = offset;
// Write the datas that will be used for recognition by the app
fwrite(&d, sizeof(Datas), 1, file);
offset += sizeof(Datas);
if (trie->is_word())
{
// If we are on a word (not just a prefix) we write the frequency of it
fwrite(&(trie->freq_), sizeof(int), 1, file);
offset += sizeof(int);
}
for (const auto& prefix : *trie)
{
// Write the word and the offset of each chile of the current trie
fwrite(prefix.first.c_str(), prefix.first.length() + 1, 1, file);
fwrite(&(offset_map[prefix.second.get()]), sizeof(unsigned), 1, file);
offset += prefix.first.length() + 1 + sizeof(unsigned);
}
return offset;
}
