int main()
{ int i,j;
char str[20] = "do1.txt";
text_t * text = text_new(str);
// if (text != NULL)
//text_print(text);
sentense_t * tmp = NULL;
int flag = 0;
for ( i = 0; i < text_sentence_count(text); i++)
{
tmp = text_get_sentence(text, i);
for ( j = 0; j < sentence_word_count(tmp); j++)
{
flag = 0;
while (strcmp(word_get_str(sentence_get_word(tmp, j)), "not") == 0)
{
j++;
if (j == sentence_word_count(tmp))
break;
flag = 1;
}
if (flag == 1)
flag = 0;
else
{
sentence_del_word(tmp, j);
j--;
}
}
}
FILE * file_out = NULL;
file_out = fopen("posle.txt", "w");
if (file_out == NULL)
return 1;
for ( i = 0; i < text_sentence_count(text); i++)
{
tmp = text_get_sentence(text, i);
for ( j = 0; j < sentence_word_count(tmp); j++)
{
if (j < sentence_word_count(tmp) - 1)
fprintf(file_out, "%s, ", word_get_str(sentence_get_word(tmp, j)));
else
fprintf(file_out, "%s\n", word_get_str(sentence_get_word(tmp, j)));
}
}
text_print(text);
text_free(text);
fclose(file_out);
return 0;
}
/*
* call-seq:
* sentence.word( idx ) -> str
*
* Returns the spelling of the n-th word in the sentence as it appears after
* tokenization.
*/
static VALUE
rlink_sentence_word( VALUE self, VALUE n ) {
struct rlink_sentence *ptr = get_sentence( self );
const char *word;
if ( !RTEST(ptr->parsed_p) )
rlink_sentence_parse( 0, 0, self );
word = sentence_get_word( (Sentence)ptr->sentence, FIX2INT(n) );
return rb_str_new2( word );
}
/*
* Class: LinkGrammar
* Method: getWord
* Signature: (I)Ljava/lang/String;
*/
JNIEXPORT jstring JNICALL
Java_org_linkgrammar_LinkGrammar_getWord(JNIEnv *env, jclass cls, jint i)
{
per_thread_data *ptd = get_ptd(env, cls);
/* Does not need to be freed, points into sentence */
const char * w = sentence_get_word(ptd->sent, i);
/* FWIW, j will be null if w is utf8-encoded Japanese or Chinese.
* I guess my JVM is not capable of handling Chinese/Japanese ??
* Maybe some special java thing needs to be installed?
*/
jstring j = (*env)->NewStringUTF(env, w);
return j;
}
std::map<int, EdgeDescription> LinkParserAdapterImpl::parseSentence(std::string sentenceStr) {
Parse_Options parseOptions = parse_options_create();
int verbosity = psi_logger.getLoggingPriority() / 100 - 4;
parse_options_set_verbosity(parseOptions, verbosity);
starts_.clear();
ends_.clear();
edgeDescriptions_.clear();
freeSentence();
sentence_ = sentence_create(sentenceStr.c_str(), dictionary_);
if (!sentence_) {
std::stringstream errorSs;
errorSs << "Link-parser failed to tokenize the input text.";
throw ParserException(errorSs.str());
}
boost::algorithm::to_lower(sentenceStr);
if (sentence_parse(sentence_, parseOptions)) {
size_t currentPos = 0;
size_t foundPos = 0;
int wordNo = 0;
while (wordNo < sentence_length(sentence_)) {
std::string word(sentence_get_word(sentence_, wordNo));
boost::algorithm::to_lower(word);
foundPos = sentenceStr.find(word, currentPos);
if (foundPos != std::string::npos) {
starts_[wordNo] = foundPos;
ends_[wordNo] = currentPos = foundPos + word.length();
}
++wordNo;
}
Linkage linkage = linkage_create(0, sentence_, parseOptions);
CNode * ctree = linkage_constituent_tree(linkage);
extractEdgeDescriptions(ctree, linkage);
linkage_free_constituent_tree(ctree);
linkage_delete(linkage);
} else {
std::stringstream errorSs;
errorSs << "Link-parser failed to parse the input text.\n"
<< "Your input text is probably not a correct sentence.";
WARN(errorSs.str());
}
return edgeDescriptions_;
}
/*
* call-seq:
* sentence.words -> array
*
* Returns the words of the sentence as they appear after tokenization.
*
* sentence = LinkParser::Dictionary.new.parse( "The dogs barks." )
* sentence.words #->
*/
static VALUE
rlink_sentence_words( VALUE self ) {
struct rlink_sentence *ptr = get_sentence( self );
const char *word;
int i, length;
VALUE words = rb_ary_new();
if ( !RTEST(ptr->parsed_p) )
rlink_sentence_parse( 0, 0, self );
length = sentence_length( (Sentence)ptr->sentence );
for ( i = 0; i < length; i++ ) {
word = sentence_get_word( (Sentence)ptr->sentence, i );
debugMsg(( "Word %d: <%s>", i, word ));
rb_ary_push( words, rb_str_new2(word) );
}
return words;
}
