SpellChecker::SpellChecker( QObject * parent )
: QObject(parent)
{
m_spell_config_ru = new_aspell_config();
aspell_config_replace(m_spell_config_ru, "dict-dir", "./dict");
aspell_config_replace(m_spell_config_ru, "encoding", "utf-8");
aspell_config_replace(m_spell_config_ru, "lang", "ru");
AspellCanHaveError * possible_err = new_aspell_speller(m_spell_config_ru);
m_spell_checker_ru = 0;
if (aspell_error_number(possible_err) != 0){
puts(aspell_error_message(possible_err));
}
else{
m_spell_checker_ru = to_aspell_speller(possible_err);
}
m_spell_config_en = new_aspell_config();
aspell_config_replace(m_spell_config_en, "dict-dir", "./dict");
aspell_config_replace(m_spell_config_en, "encoding", "utf-8");
aspell_config_replace(m_spell_config_en, "lang", "en");
possible_err = new_aspell_speller(m_spell_config_en);
m_spell_checker_en = 0;
if (aspell_error_number(possible_err) != 0){
puts(aspell_error_message(possible_err));
}
else{
m_spell_checker_en = to_aspell_speller(possible_err);
}
m_codec = QTextCodec::codecForName("UTF-8");
}
int main(int argc,char **argv)
{
#ifdef _USEQT5_
QApplication app(argc,argv);
#endif
AspellCanHaveError* possible_err;
aspellConfig=new_aspell_config();
possible_err=new_aspell_speller(aspellConfig);
if(aspell_error_number(possible_err)!= 0)
puts(aspell_error_message(possible_err));
else
spellChecker=to_aspell_speller(possible_err);
#ifndef _USEQT5_
gtk_init(&argc,&argv);
buildMainGuiGtk();
gtk_window_stick(GTK_WINDOW(window));
gtk_window_set_keep_above((GtkWindow*)window,true);
gtk_widget_show_all(window);
gtk_main();
#else
holdapp=&app;
buildMainGuiQt();
window->show();
app.exec();
#endif
}
bool SpellChecker::addCheckedLang(QString &name)
{
if (checkers.find(name) != checkers.end())
return true;
aspell_config_replace(spellConfig, "lang", name.toAscii());
// create spell checker using prepared configuration
AspellCanHaveError* possibleErr = new_aspell_speller(spellConfig);
if (aspell_error_number(possibleErr) != 0)
{
MessageBox::msg(aspell_error_message(possibleErr));
return false;
}
else
checkers[name] = to_aspell_speller(possibleErr);
if (checkers.size() == 1)
{
foreach(ChatWidget *chat, ChatWidgetManager::instance()->chats())
chatCreated(chat);
}
return true;
}
//__________________________________________________________________________
void Speller::Aspell::Suggest::init(const std::string& lang,
const std::string& jargon,
const std::string& encoding)
throw( std::invalid_argument, std::runtime_error )
{
// Save aspell configuration values
flang = lang;
fjargon = jargon;
fencoding = encoding;
fconfig = new_aspell_config();
try
{
setConfig();
}
catch( const std::invalid_argument& err )
{
throw err;
}
AspellCanHaveError* ret = new_aspell_speller( fconfig );
delete_aspell_config( fconfig );
if( aspell_error_number( ret ) != 0 )
{
delete_aspell_can_have_error( ret );
throw std::runtime_error( "(Aspell::Speller::Suggest::init"
"): Error in creating speller." );
}
else
{
fspeller = to_aspell_speller( ret );
fconfig = aspell_speller_config( fspeller );
}
}
//__________________________________________________________________________
void Speller::Aspell::Suggest::resetConfig()
throw( std::invalid_argument, std::runtime_error )
{
delete_aspell_config( fconfig );
fconfig = new_aspell_config();
try
{
setConfig();
}
catch( const std::invalid_argument& err )
{
throw err;
}
AspellCanHaveError* ret = new_aspell_speller( fconfig );
if( aspell_error_number( ret ) != 0 )
{
delete_aspell_can_have_error( ret );
throw std::runtime_error( "(Aspell::Speller::Suggest::Reset"
"Config): Error in creating "
"speller." );
}
else
{
// Following statement causes a crash, hence commented out
//delete_aspell_speller( fspeller );
fspeller = to_aspell_speller( ret );
delete_aspell_config( fconfig );
fconfig = aspell_speller_config( fspeller );
}
}
void AspellAdapterImpl::createAspellInstance() {
AspellCanHaveError * possibleError = new_aspell_speller(aspellConfig_);
if (aspell_error_number(possibleError) != 0) {
ERROR("ASPELL CREATION ERROR: " << aspell_error_message(possibleError));
// @todo
// throw PsiException(aspell_error_message(possibleError));
} else {
aspellSpeller_ = to_aspell_speller(possibleError);
}
}
void initAspell()
{
spell_config = new_aspell_config();
aspell_config_replace(spell_config, "lang", "en_US"); //set language
possible_err = new_aspell_speller(spell_config);
spell_checker = 0;
if (aspell_error_number(possible_err) != 0) {
printf("%s ", aspell_error_message(possible_err));
}
else {
printf("Unscrambled words:\n");
spell_checker = to_aspell_speller(possible_err);
}
}
void init(void)
{
char* filename;
#ifdef _ASPELL_
AspellCanHaveError* possible_err;
#endif
lineWrap=true;
highLight=true;
useUnderline=true;
tabWidth=4;
fontAndSize=strdup("mono 10");
terminalCommand=strdup("xterm -e");
windowWidth=800;
windowHeight=400;
windowX=-1;
windowY=-1;
wrapSearch=true;
insensitiveSearch=true;
replaceAll=false;
showLiveSearch=true;
gzipPages=false;
asprintf(&filename,"%s/.ManPageEditor",getenv("HOME"));
g_mkdir_with_parents(filename,493);
g_free(filename);
readConfig();
tmpGzipPages=gzipPages;
tmpHighLight=highLight;
tmpLineWrap=lineWrap;
tmpTabWidth=tabWidth;
tmpUseUnderline=useUnderline;
tmpShowLiveSearch=showLiveSearch;
#ifdef _ASPELL_
aspellConfig=new_aspell_config();
possible_err=new_aspell_speller(aspellConfig);
if(aspell_error_number(possible_err)!= 0)
puts(aspell_error_message(possible_err));
else
spellChecker=to_aspell_speller(possible_err);
#endif
}
bool AspellAdapterImpl::isDictionaryAvailable(const std::string & langCode) {
AspellConfig * tempAspellConfig = new_aspell_config();
aspell_config_replace(tempAspellConfig, "lang", langCode.c_str());
AspellCanHaveError * possibleError = new_aspell_speller(tempAspellConfig);
bool result = true;
if (aspell_error_number(possibleError) != 0) {
result = false;
} else {
AspellSpeller * tempSpeller = to_aspell_speller(possibleError);
delete_aspell_speller(tempSpeller);
}
delete_aspell_config(tempAspellConfig);
return result;
}
void ASpellChecker::setActiveLanguages(const QList<QString>& langs)
{
clearSpellers();
foreach(const QString& lang, langs)
{
AspellConfig* conf = aspell_config_clone(config_);
aspell_config_replace(conf, "lang", lang.toUtf8().constData());
AspellCanHaveError* ret = new_aspell_speller(conf);
if (aspell_error_number(ret) == 0) {
spellers_.append(to_aspell_speller(ret));
}
else {
qDebug() << QString("Aspell error: %1").arg(aspell_error_message(ret));
}
delete_aspell_config(conf);
}
void ASpellChecker::setActiveLanguages(const QSet<LanguageManager::LangId>& langs)
{
clearSpellers();
for(auto const &lang: langs)
{
AspellConfig* conf = aspell_config_clone(config_);
aspell_config_replace(conf, "lang", LanguageManager::toString(lang)
.replace(QLatin1Char('-'),QLatin1Char('_')).toUtf8().constData());
AspellCanHaveError* ret = new_aspell_speller(conf);
if (aspell_error_number(ret) == 0) {
spellers_.append(to_aspell_speller(ret));
}
else {
qDebug() << QString("Aspell error: %1").arg(aspell_error_message(ret));
}
delete_aspell_config(conf);
}
}
ASpellChecker::ASpellChecker()
{
config_ = NULL;
speller_ = NULL;
config_ = new_aspell_config();
aspell_config_replace(config_, "encoding", "utf-8");
#ifdef Q_WS_WIN
aspell_config_replace(config_, "conf-dir", QDir::homeDirPath());
aspell_config_replace(config_, "data-dir", QString("%1/aspell/data").arg(QCoreApplication::applicationDirPath()));
aspell_config_replace(config_, "dict-dir", QString("%1/aspell/dict").arg(QCoreApplication::applicationDirPath()));
#endif
AspellCanHaveError* ret = new_aspell_speller(config_);
if (aspell_error_number(ret) == 0) {
speller_ = to_aspell_speller(ret);
}
else {
qWarning(QString("Aspell error: %1").arg(aspell_error_message(ret)).toAscii());
}
}
/**
* Create a spell checker for a language
* @param language the language code e.g. en_GB or it
* @return a checker or NULL
*/
static checker *checker_create( const char *language )
{
int err = 0;
checker *c = calloc( 1, sizeof(checker) );
if ( c != NULL )
{
strncpy(c->lang,language,24);
c->spell_config = new_aspell_config();
if ( c->spell_config != NULL )
{
aspell_config_replace( c->spell_config, "lang", language );
AspellCanHaveError *possible_err
= new_aspell_speller(c->spell_config);
c->spell_checker = 0;
if (aspell_error_number(possible_err) != 0)
{
fprintf(stderr,"%s\n",aspell_error_message(possible_err));
err = 1;
}
else
{
c->spell_checker = to_aspell_speller(possible_err);
if ( c->spell_checker == NULL )
{
fprintf(stderr,"checker: failed to initialise speller\n");
err = 1;
}
}
if ( err )
{
checker_dispose( c );
c = NULL;
}
}
else
fprintf(stderr,"checker: failed to create speller\n");
}
else
fprintf(stderr,"checker: failed to create object\n");
return c;
}
/**
* create a neew spell-checker for the language 'lang'
*/
void * spellcheck_create(const char * lang)
{
struct linkgrammar_aspell *aspell = NULL;
size_t i = 0;
AspellCanHaveError *spell_err = NULL;
for (i = 0; i < sizeof(spellcheck_lang_mapping)/sizeof(char *); i += 2)
{
if (0 != strcmp(lang, spellcheck_lang_mapping[i])) continue;
aspell = (struct linkgrammar_aspell *)malloc(sizeof(struct linkgrammar_aspell));
if (!aspell) {
prt_error("Error: out of memory. Aspell not used.\n");
aspell = NULL;
break;
}
aspell->config = NULL;
aspell->speller = NULL;
aspell->config = new_aspell_config();
if (aspell_config_replace(aspell->config, ASPELL_LANG_KEY,
spellcheck_lang_mapping[i]) == 0) {
prt_error("Error: failed to set language in aspell: %s\n", lang);
delete_aspell_config(aspell->config);
free(aspell);
aspell = NULL;
break;
}
spell_err = new_aspell_speller(aspell->config);
if (aspell_error_number(spell_err) != 0) {
prt_error("Error: Aspell: %s\n", aspell_error_message(spell_err));
delete_aspell_can_have_error(spell_err);
delete_aspell_config(aspell->config);
free(aspell);
aspell = NULL;
break;
}
aspell->speller = to_aspell_speller(spell_err);
break;
}
return aspell;
}
static gboolean
gtkspell_set_language_internal(GtkSpell *spell, const gchar *lang, GError **error) {
AspellConfig *config;
AspellCanHaveError *err;
if (lang == NULL) {
lang = g_getenv("LANG");
if (lang) {
if (g_strncasecmp(lang, "C", 1) == 0)
lang = NULL;
else if (lang[0] == 0)
lang = NULL;
}
}
config = new_aspell_config();
if (lang)
aspell_config_replace(config, "language-tag", lang);
aspell_config_replace(config, "encoding", "utf-8");
err = new_aspell_speller(config);
delete_aspell_config(config);
if (aspell_error_number(err) != 0) {
#ifdef USING_ASPELL
g_set_error(error, GTKSPELL_ERROR, GTKSPELL_ERROR_BACKEND,
"aspell: %s", aspell_error_message(err));
#elif defined USING_PSPELL
g_set_error(error, GTKSPELL_ERROR, GTKSPELL_ERROR_BACKEND,
"pspell: %s", aspell_error_message(err));
#endif
return FALSE;
}
if (spell->speller)
delete_aspell_speller(spell->speller);
spell->speller = to_aspell_speller(err);
return TRUE;
}
void SpellChecker::setLanguage(const QString &lang)
{
if (lang.isEmpty()) return;
delete_aspell_speller(spell_checker1);
delete_aspell_speller(spell_checker2);
bad_language.clear();
m_lang2 = "en";
m_lang1 = m_map->value(lang, QString("en"));
if (lang.contains("+")) {
QStringList sl = lang.split("+");
m_lang1 = m_map->value(sl[0], QString("en"));
m_lang2 = m_map->value(sl[1], QString("en"));
if ((m_lang1 == "deu")||(lang == "ger")) {
m_lang1 = "de_DE";
}
if ((m_lang2 == "deu")||(lang == "ger")) {
m_lang2 = "de_DE";
}
}
if (lang == "rus_fra") {
m_lang1 = "ru";
m_lang2 = "fr";
} else if (lang == "rus_ger") {
m_lang1 = "ru";
m_lang2 = "de";
} else if (lang == "rus_spa") {
m_lang1 = "ru";
m_lang2 = "es";
}
if ((lang == "deu")||(lang == "ger")) {
m_lang1 = "de_DE";
m_lang2 = "de_AT";
}
if (lang == "ruseng") {
m_lang1 = "ru";
m_lang2 = "en";
}
aspell_config_replace(spell_config1, "lang", m_lang1.toAscii());
aspell_config_replace(spell_config2, "lang", m_lang2.toAscii());
AspellCanHaveError *possible_err = new_aspell_speller(spell_config1);
spell_checker1 = 0;
if (aspell_error_number(possible_err) == 0)
spell_checker1 = to_aspell_speller(possible_err);
else
delete_aspell_can_have_error(possible_err);
possible_err = new_aspell_speller(spell_config2);
spell_checker2 = 0;
if (aspell_error_number(possible_err) == 0)
spell_checker2 = to_aspell_speller(possible_err);
else
delete_aspell_can_have_error(possible_err);
// Check absent dictionary
if (spell_checker1 == 0)
bad_language = m_lang1;
if (spell_checker2 == 0)
bad_language = m_lang2;
}
/**
* Create a userdata object
* @param language the language e.g. "en_GB"
* @param rules recipe file path to recipe file
* @param fmt the format object containing function pointers
* @return a complete userdata object or NULL
*/
userdata *userdata_create( const char *language, char *barefile, recipe *rules,
format *fmt, hh_exceptions *hhe )
{
int err = 0;
userdata *u = calloc( 1, sizeof(userdata) );
if ( u != NULL )
{
u->rules = rules;
if ( hhe != NULL )
u->hhe = hhe;
u->spell_config = new_aspell_config();
if ( u->spell_config != NULL )
{
aspell_config_replace( u->spell_config, "lang", language );
AspellCanHaveError *possible_err
= new_aspell_speller(u->spell_config);
u->spell_checker = 0;
if (aspell_error_number(possible_err) != 0)
{
fprintf(stderr,"%s\n",aspell_error_message(possible_err));
err = 1;
}
else
{
u->spell_checker = to_aspell_speller(possible_err);
if ( u->spell_checker == NULL )
{
fprintf(stderr,"userdata: failed to initialise speller\n");
err = 1;
}
}
u->range_stack = stack_create();
if ( u->range_stack == NULL )
{
err = 1;
fprintf(stderr,
"stripper: failed to allocate store for range stack" );
}
u->ignoring = stack_create();
if ( u->ignoring == NULL )
{
err = 1;
fprintf(stderr,
"stripper: failed to allocate store for ignore stack" );
}
if ( !open_dest_files(u,barefile,fmt) )
{
err = 1;
fprintf(stderr,"stripper: couldn't open dest files\n");
}
}
else
{
fprintf(stderr, "userdata: failed to initialise speller\n");
err = 1;
}
}
else
fprintf(stderr, "userdata:failed to allocate object\n");
if ( err )
{
userdata_dispose( u );
u = NULL;
}
return u;
}
static int get_aspell_hits(const char *word, int len)
{
if (len < 2)
{
log_message(DEBUG, " [-]Skip aspell checking (word is very short)");
return NO_LANGUAGE;
}
AspellConfig *spell_config = new_aspell_config();
for (int lang = 0; lang < xconfig->total_languages; lang++)
{
char *lang_word = (char *) malloc(1 * sizeof(char));
if (lang_word == NULL)
continue;
lang_word[0] = NULLSYM;
for (int i = 0; i < len; i++)
{
KeyCode kc;
int modifier;
main_window->xkeymap->char_to_keycode(main_window->xkeymap, word[i], &kc, &modifier);
char *symbol = keycode_to_symbol(kc, lang, modifier);
if ((symbol == NULL) || (lang_word == NULL))
continue;
lang_word = (char *) realloc(lang_word, (strlen(lang_word) + strlen(symbol) + 1) * sizeof(char));
if (lang_word != NULL)
strcat(lang_word, symbol);
free(symbol);
}
if (lang_word == NULL)
continue;
aspell_config_replace(spell_config, "lang", xconfig->languages[lang].dir);
AspellCanHaveError *possible_err = new_aspell_speller(spell_config);
AspellSpeller *spell_checker = 0;
if (aspell_error_number(possible_err) == 0)
{
spell_checker = to_aspell_speller(possible_err);
int correct = aspell_speller_check(spell_checker, lang_word, strlen(lang_word));
if (correct)
{
log_message(DEBUG, " [+]Found this word in %s aspell dictionary", xconfig->get_lang_name(xconfig, lang));
delete_aspell_speller(spell_checker);
free(lang_word);
return lang;
}
}
else
{
log_message(DEBUG, " [!]Error aspell checking for %s aspell dictionary", xconfig->get_lang_name(xconfig, lang));
}
delete_aspell_speller(spell_checker);
free(lang_word);
}
log_message(DEBUG, " [-]This word has no hits for all aspell dictionaries");
return NO_LANGUAGE;
}
/* Create a new speller *******************************************************/
static PyObject* new_speller(PyObject* self, PyObject* args) {
aspell_AspellObject* newobj;
AspellSpeller* speller = 0;
AspellConfig* config;
AspellCanHaveError* possible_error;
int i;
int n; /* arg count */
char *key, *value;
config = new_aspell_config();
if (config == NULL) {
PyErr_SetString(_AspellModuleException, "can't create config");
return NULL;
}
/* check constructor arguments */
n = PyTuple_Size(args);
switch (n) {
case 0: /* no arguments passed */
break;
case 2: /* constructor is called with single pair: key & value */
if (PyArg_ParseTuple(args, "ss", &key, &value)) {
if (!aspell_config_replace(config, key, value)) {
PyErr_SetString(_AspellConfigException, aspell_config_error_message(config));
goto arg_error;
}
break;
}
PyErr_Clear();
default: /* list of tuples key&value */
for (i=0; i<n; i++) {
if (!PyArg_ParseTuple(PyTuple_GetItem(args, i), "ss", &key, &value)) {
PyErr_Format(PyExc_TypeError, "argument %d: tuple of two strings (key, value) expeced", i);
goto arg_error;
}
if (!aspell_config_replace(config, key, value)) {
PyErr_SetString(_AspellConfigException, aspell_config_error_message(config));
goto arg_error;
}
}
Py_DECREF(args);
break;
}
/* try to create a new speller */
possible_error = new_aspell_speller(config);
delete_aspell_config(config);
if (aspell_error_number(possible_error) == 0)
/* save a speller */
speller = to_aspell_speller(possible_error);
else {
/* or raise an exception */
PyErr_SetString(_AspellSpellerException, aspell_error_message(possible_error));
delete_aspell_can_have_error(possible_error);
return NULL;
}
/* create a new py-object */
newobj = (aspell_AspellObject*)PyObject_New(aspell_AspellObject, &aspell_AspellType);
newobj->speller = speller;
return (PyObject*)newobj;
/* argument error: before return NULL we need to
delete speller's config we've created */
arg_error:
delete_aspell_config(config);
return NULL;
}
// runs in O(t^2) time where t is the number of tokens in the input corpus
// We consider maxK to be fairly constant
void rawr::compile(int maxK)
{
_maxK = maxK;
std::vector<std::vector<token_id>> tokens;
std::set<std::string> thashtags;
std::set<std::string> fv_emoticons;
std::ifstream fvefile("emoticons.txt");
if (fvefile)
{
std::string line;
while (getline(fvefile, line))
{
fv_emoticons.insert(line);
emoticons.forms.add(line);
}
}
fvefile.close();
std::map<std::string, std::string> canonical_form;
AspellConfig* spell_config = new_aspell_config();
AspellCanHaveError* possible_err = new_aspell_speller(spell_config);
if (aspell_error_number(possible_err) != 0)
{
std::cout << "aspell error: " << aspell_error_message(possible_err) << std::endl;
exit(1);
}
AspellSpeller* spell_checker = to_aspell_speller(possible_err);
std::cout << "Reading emojis..." << std::endl;
prefix_search emojis;
std::ifstream emoji_file("emojis.txt");
if (emoji_file)
{
while (!emoji_file.eof())
{
std::string rawmojis;
getline(emoji_file, rawmojis);
if (rawmojis.back() == '\r')
{
rawmojis.pop_back();
}
emojis.add(rawmojis);
}
emoji_file.close();
}
std::cout << "Tokenizing corpus... 0%" << std::flush;
int len = 0;
for (auto c : _corpora)
{
len += c.length();
}
int startper = 0;
int per = 0;
int perprime = 0;
std::cout.fill(' ');
for (int i = 0; i < _corpora.size(); i++)
{
size_t start = 0;
int end = 0;
std::vector<token_id> tkcor;
while (end != std::string::npos)
{
perprime = (startper + end) * 100 / len;
if (perprime != per)
{
per = perprime;
std::cout << "\b\b\b\b" << std::right;
std::cout.width(3);
std::cout << per << "%" << std::flush;
}
end = _corpora[i].find_first_of(" \n", start);
bool emoji = false;
std::string te = _corpora[i].substr(start, (end == std::string::npos) ? std::string::npos : end - start + 1);
std::string t = "";
if (te.compare("") && te.compare(".") && te.compare(" "))
{
if (te.back() == ' ')
{
te.pop_back();
}
// Extract strings of emojis into their own tokens even if they're not space delimited
int m = emojis.match(te);
emoji = m > 0;
if (m == 0) m = 1;
t = te.substr(0,m);
te = te.substr(m);
while (!te.empty())
{
m = emojis.match(te);
if (emoji == (m > 0))
{
if (m == 0) m = 1;
t += te.substr(0,m);
te = te.substr(m);
} else {
end = start + t.length() - 1;
break;
}
}
std::string tc(t);
std::transform(tc.begin(), tc.end(), tc.begin(), ::tolower);
int pst = tc.find_first_not_of("\"([*");
int dst = tc.find_last_not_of("\")]*.,?!\n;:");
std::string canonical("");
if ((pst != std::string::npos) && (dst != std::string::npos))
{
canonical = std::string(tc, pst, dst - pst + 1);
}
word& w = ([&] () -> word& {
// Hashtag freevar
if (canonical[0] == '#')
{
thashtags.insert(canonical);
return hashtags;
}
// Emoticon freevar
if (emoji)
{
emoticons.forms.add(canonical);
return emoticons;
}
if ((pst != std::string::npos) && (dst != std::string::npos))
{
std::string emoticon_canon(t, pst, t.find_last_not_of("\"]*\n.,?!") - pst + 1);
if (fv_emoticons.count(emoticon_canon) == 1)
{
emoticons.forms.add(emoticon_canon);
return emoticons;
}
}
// Basically any other word
if (canonical_form.count(canonical) == 0)
{
if (
// Legacy freevars should be distinct from tokens containing similar words
(canonical.find("$name$") != std::string::npos)
// Words with no letters will be mangled by the spell checker
|| (canonical.find_first_of("ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz") == std::string::npos)
)
{
canonical_form[canonical] = canonical;
words.emplace(canonical, canonical);
} else {
int correct = aspell_speller_check(spell_checker, canonical.c_str(), canonical.size());
if (correct)
{
words.emplace(canonical, canonical);
canonical_form[canonical] = canonical;
} else {
const AspellWordList* suggestions = aspell_speller_suggest(spell_checker, canonical.c_str(), canonical.size());
AspellStringEnumeration* elements = aspell_word_list_elements(suggestions);
const char* replacement = aspell_string_enumeration_next(elements);
if (replacement != NULL)
{
std::string sugrep(replacement);
canonical_form[canonical] = sugrep;
if (words.count(sugrep) == 0)
{
words.emplace(sugrep, sugrep);
}
} else {
words.emplace(canonical, canonical);
canonical_form[canonical] = canonical;
}
delete_aspell_string_enumeration(elements);
}
}
}
word& tw = words.at(canonical_form.at(canonical));
tw.forms.add(canonical);
return tw;
})();
token tk(w);
tk.raw = t;
for (char c : t)
{
if (c == '*')
{
tk.delimiters[{parentype::asterisk, doublestatus::opening}]++;
} else if (c == '[')
{
tk.delimiters[{parentype::square_bracket, doublestatus::opening}]++;
} else if (c == '(')
{
tk.delimiters[{parentype::paren, doublestatus::opening}]++;
} else if (c == '"')
{
tk.delimiters[{parentype::quote, doublestatus::opening}]++;
} else {
break;
}
}
int backtrack = t.find_last_not_of(".,?!])\"*\n;:") + 1;
if (backtrack != t.length())
{
std::string ending = t.substr(backtrack);
std::string suffix;
bool newline = false;
bool terminating = false;
for (char c : ending)
{
if ((c == '.') || (c == ',') || (c == '?') || (c == '!') || (c == ';') || (c == ':'))
{
suffix += c;
terminating = true;
continue;
} else if (c == '\n')
{
newline = true;
terminating = true;
continue;
}
parentype pt = ([&] {
switch (c)
{
case ']': return parentype::square_bracket;
case ')': return parentype::paren;
case '*': return parentype::asterisk;
case '"': return parentype::quote;
}
})();
if (tk.delimiters[{pt, doublestatus::opening}] > 0)
{
tk.delimiters[{pt, doublestatus::opening}]--;
tk.delimiters[{pt, doublestatus::both}]++;
} else {
tk.delimiters[{pt, doublestatus::closing}]++;
}
}
if (terminating)
{
if ((suffix == ",") && (!newline))
{
tk.suffix = suffixtype::comma;
} else {
tk.suffix = suffixtype::terminating;
if (!newline)
{
w.terms.add({suffix, false});
} else {
w.terms.add({".", false});
}
}
}
}
tkcor.push_back(_tokenstore.add(tk));
}
start = ((end > (std::string::npos - 1) ) ? std::string::npos : end + 1);
}
tokens.push_back(tkcor);
startper += _corpora[i].length();
}
std::cout << "\b\b\b\b100%" << std::endl;
delete_aspell_speller(spell_checker);
delete_aspell_config(spell_config);
std::cout << canonical_form.size() << " distinct forms" << std::endl;
std::cout << words.size() << " distinct words" << std::endl;
// Time to condense the distribution stuff for the words
std::cout << "Compiling token histograms..." << std::endl;
for (auto& it : words)
{
it.second.forms.compile();
it.second.terms.compile();
}
// Hashtag freevar is not frequency distributed
for (auto& it : thashtags)
{
hashtags.forms.add(it);
}
hashtags.forms.compile();
hashtags.terms.compile();
// Compile other freevars
emoticons.forms.compile();
emoticons.terms.compile();
// Compile the interned tokens.
_tokenstore.compile();
// kgram distribution
std::cout << "Creating markov chain... 0%" << std::flush;
std::map<kgram, std::map<token_id, token_data> > tstats;
len = 0;
for (auto c : tokens)
{
len += (maxK-1) * c.size();
}
startper = 0;
per = 0;
perprime = 0;
int corpid = 0;
for (auto corpus : tokens)
{
for (int k=0; k<maxK && k<corpus.size(); k++)
{
// The zero'th token should be a terminator.
token_id fid = corpus[k];
const token& f = _tokenstore.get(fid);
kgram term_prefix(corpus.begin(), corpus.begin()+k);
term_prefix.push_front(wildcardQuery);
if (tstats[term_prefix].count(fid) == 0)
{
tstats[term_prefix].emplace(fid, fid);
}
token_data& td2 = tstats[term_prefix].at(fid);
td2.all++;
td2.corpora.insert(corpid);
if (std::find_if(f.raw.begin(), f.raw.end(), ::islower) == f.raw.end())
{
td2.uppercase++;
} else if (isupper(f.raw[0]))
{
td2.titlecase++;
}
}
for (int k=1; k<maxK && k<corpus.size(); k++)
{
for (int i=0; i<(corpus.size() - k); i++)
{
perprime = (startper+i) * 100 / len;
if (perprime != per)
{
per = perprime;
std::cout << "\b\b\b\b" << std::right;
std::cout.width(3);
std::cout << per << "%" << std::flush;
}
kgram prefix(corpus.begin()+i, corpus.begin()+i+k);
token_id fid = corpus[i+k];
const token& f = _tokenstore.get(fid);
if (tstats[prefix].count(fid) == 0)
{
tstats[prefix].emplace(fid, fid);
}
token_data& td = tstats[prefix].at(fid);
td.all++;
td.corpora.insert(corpid);
if (std::find_if(f.raw.begin(), f.raw.end(), ::islower) == f.raw.end())
{
td.uppercase++;
} else if (isupper(f.raw[0]))
{
td.titlecase++;
}
const token& startTok = _tokenstore.get(std::begin(prefix)->tok);
if (startTok.suffix == suffixtype::terminating)
{
kgram term_prefix(prefix);
term_prefix.pop_front();
term_prefix.push_front(wildcardQuery);
if (tstats[term_prefix].count(fid) == 0)
{
tstats[term_prefix].emplace(fid, fid);
}
token_data& td2 = tstats[term_prefix].at(fid);
td2.all++;
td2.corpora.insert(corpid);
if (std::find_if(f.raw.begin(), f.raw.end(), ::islower) == f.raw.end())
{
td2.uppercase++;
} else if (isupper(f.raw[0]))
{
td2.titlecase++;
}
}
}
startper += corpus.size();
}
corpid++;
}
std::cout << "\b\b\b\b100%" << std::endl;
// Condense the kgram distribution
std::cout << "Compiling kgram distributions... 0%";
len = tstats.size();
per = 0;
perprime = 0;
int indicator = 0;
for (auto& it : tstats)
{
indicator++;
perprime = indicator * 100 / len;
if (per != perprime)
{
per = perprime;
std::cout << "\b\b\b\b" << std::right;
std::cout.width(3);
std::cout << per << "%" << std::flush;
}
kgram klist = it.first;
auto& probtable = it.second;
auto& distribution = _stats[klist];
int max = 0;
for (auto& kt : probtable)
{
max += kt.second.all;
distribution.emplace(max, kt.second);
}
}
std::cout << "\b\b\b\b100%" << std::endl;
_compiled = true;
}
