void add_data_from_line(Memory& mem, string& line)
{
regex regex_apostrophe("'");
std::regex_replace(line, regex_apostrophe, "’");
regex regex_word("([a-zA-záéíóúñÁÉÍÓÚÑ'’]+)");
sregex_iterator find_word(line.begin(), line.end(), regex_word);
for (sregex_iterator find_end; find_word != find_end; ++find_word) {
string word_found = find_word->str();
word_found = to_lower_case(word_found);
update_word(mem.word_list, word_found);
}
vector<regex> regex_punc = {
regex("\\."),
regex(","),
regex(";"),
regex(":"),
regex("!"),
regex("\\?"),
regex("(?:[-–—]\\s*){1,2}"),
regex("…|(?:\\.\\s*){3}")
};
for (int i = 0; i < PUNC_NUM; ++i) {
sregex_iterator find_punc(line.begin(), line.end(), regex_punc[i]);
mem.punc_freq[i] += std::distance(find_punc, sregex_iterator());
}
regex regex_split_line("^([^.!?…]*[.!?…])?((?:[^.!?…]*[.!?…])*?)([^.!?…]*)$");
std::sregex_token_iterator split_line_prev(line.begin(), line.end(), regex_split_line, 1);
std::sregex_token_iterator split_line_cont(line.begin(), line.end(), regex_split_line, 2);
std::sregex_token_iterator split_line_xtra(line.begin(), line.end(), regex_split_line, 3);
string sentence_prev = *split_line_prev;
int sentence_prev_len = get_word_count(sentence_prev);
if (sentence_prev_len > 0) {
sentence_prev_len += mem.sentence_carry;
mem.sentence_len.push_back(sentence_prev_len);
mem.sentence_carry = 0;
}
string sentence_cont = *split_line_cont;
regex regex_split_sentence("([^\\.!?…]*)[\\.!?…]");
sregex_iterator find_sentence(sentence_cont.begin(), sentence_cont.end(), regex_split_sentence);
for (sregex_iterator find_end; find_sentence != find_end; ++find_sentence) {
string sentence_found = find_sentence->str();
int sentence_len = get_word_count(sentence_found);
mem.sentence_len.push_back(sentence_len);
mem.sentence_carry = 0;
}
string sentence_xtra = *split_line_xtra;
int sentence_xtra_len = get_word_count(sentence_xtra);
if (sentence_xtra_len > 0) {
mem.sentence_carry += sentence_xtra_len;
}
}
int main()
{
input();
int word_count = get_word_count();
printf("Words = %d\n", word_count);
int num_sum = get_num_sum();
printf("Sum of numbers = %d\n", num_sum);
getch();
return 0;
}
static ssize_t word_count_write(struct file *file, const char __user * buf, size_t count, loff_t *ppos)
{
ssize_t written = count;
copy_from_user(mem, buf, count);
mem[count] = '\0';
word_count=get_word_count(mem);
read_flag='n';
printk("write: %d Byte \n", (int)count);
return written;
}
static ssize_t word_count_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos)
{
ssize_t written = count;
if (copy_from_user(mem, buf, count))
{
return -EINVAL;
}
mem[count] = '\0';
word_count = get_word_count(mem);
printk("write:word count:%d\n", (int) word_count);
return written;
}
//------------------------------------------------------------------------------
str_iter line_state::get_end_word() const
{
int n = get_word_count();
return (n ? get_word(n - 1) : str_iter());
}
//------------------------------------------------------------------------------
bool line_state::get_end_word(str_base& out) const
{
int n = get_word_count();
return (n ? get_word(n - 1, out) : false);
}
