void parse_conll_data (
const std::string& filename,
std::vector<std::vector<std::string> >& sentences,
std::vector<std::vector<std::pair<unsigned long, unsigned long> > >& chunks,
std::vector<std::vector<unsigned long> >& chunk_labels
)
{
std::vector<labeled_sentence> data = parse_conll_data (filename);
sentences.resize(data.size());
chunks.resize(data.size());
chunk_labels.resize(data.size());
for (unsigned long i = 0; i < data.size(); ++i)
{
convert_sentence(data[i], sentences[i], chunks[i], chunk_labels[i]);
}
}
int read_sentence(FILE *file,sentence_type *sentence)
{
int i;
char wbuf[100],tbuf[100];
int lw,lt;
if(fscanf(file,"%d",&sentence->nws)==EOF)
return 0;
assert(sentence->nws<PMAXWORDS);
for(i=0;i<sentence->nws;i++)
{
fscanf(file,"%s %s ",wbuf,tbuf);
sentence->words[i]=strdup(wbuf);
sentence->tags[i]=strdup(tbuf);
}
convert_sentence(sentence);
return 1;
}
