//構築したルールを音声認識エンジンにコミットします。
xreturn::r<bool> JuliusPlus::CommitRule()
{
if (! this->IsNeedUpdateRule )
{//アップデートする必要なし
return true;
}
//マイクから入力用
std::ofstream dfa("__temp__regexp_test.dfa");
std::ofstream dict("__temp__regexp_test.dict");
this->MakeJuliusRule(this->Grammer,false, &dfa,&dict);
//ディクテーションフィルター用
std::ofstream dfaFile("__temp__regexp_test_file.dfa");
std::ofstream dictFile("__temp__regexp_test_file.dict");
this->MakeJuliusRule(this->YobikakeRuleHandle,true,&dfaFile,&dictFile);
this->JuliusStop();
this->JuliusFileStart();
this->JuliusStart();
//アップデートが終わったので再びルールに変更が加わるまではアップデートしない。
this->IsNeedUpdateRule = false;
return true;
}
main(int argc, char **argv)
{
int i, sw = 0;
long minbox = 0L, maxbox = 0L, npts, temp;
/* Read and interpret the command line. */
pname = argv[0];
for (i = 1; i < argc && *argv[i] == '-'; i++) {
switch(argv[i][1]) {
case 'd': /* set nfit (the order of the regression fit) */
if ((nfit = atoi(argv[++i])+1) < 2)
error("order must be greater than 0");
break;
case 'i': /* input data are already integrated */
iflag = 0; break;
case 'l': /* set minbox (the minimum box size) */
minbox = atol(argv[++i]); break;
case 'u': /* set maxbox (the maximum box size) */
maxbox = atol(argv[++i]); break;
case 's': /* enable sliding window mode */
sw = 1; break;
case 'h': /* print usage information and quit */
default:
help();
exit(1);
}
}
/* Allocate and fill the input data array seq[]. */
npts = input();
/* Set minimum and maximum box sizes. */
if (minbox < 2*nfit) minbox = 2*nfit;
if (maxbox == 0 || maxbox > npts/4) maxbox = npts/4;
if (minbox > maxbox) {
SWAP(minbox, maxbox);
if (minbox < 2*nfit) minbox = 2*nfit;
}
/* Allocate and fill the box size array rs[]. rscale's third argument
specifies that the ratio between successive box sizes is 2^(1/8). */
nr = rscale(minbox, maxbox, pow(2.0, 1.0/8.0));
/* Allocate memory for dfa() and the functions it calls. */
setup();
/* Measure the fluctuations of the detrended input data at each box size
using the DFA algorithm; fill mse[] with these results. */
dfa(seq, npts, nfit, rs, nr, sw);
/* Output the results. */
for (i = 1; i <= nr; i++)
printf("%g\n", log10(mse[i])/2.0);
/* Release allocated memory. */
cleanup();
exit(0);
}
void eskoTest() {
unsigned step = 5000;
unsigned start = 30000;
unsigned stop = 120000;
unsigned topL = 300;
unsigned minL = 5;
for(unsigned cur = start; cur <= stop; cur += step) {
takeSubstring("data/ecoli.seq", "data/tmpseq.seq", 0, cur);
//takeSubstring("data/ecoli.seq", "data/tmpseq.seq", 278000, 293000);
unsigned L = minL;
unsigned R = topL;
unsigned ans = 0U;
while (L<=R) {
unsigned mid = (L+R) / 2U;
unsigned k = mid-1U;
Genome genome;
readGenome("data/tmpseq.seq", mid, genome);
AhoCorasick * aho = new AhoCorasick(genome);
aho->filterOverlaps(k);
NFA_Automata nfa(*aho, genome, 0, genome.generatedReads.size()-1);
delete aho;
DFA_Automata dfa(nfa, genome, 1.0);
// remove reads since we do not need them anymore
genome.generatedReads.resize(0);
genome.sequence.resize(0);
bool uniqueOk = dfa.isCOAUnique();
if (uniqueOk) {
R = mid-1U;
ans = mid;
} else {
L = mid+1U;
}
}
std::string seq = readSequence("data/tmpseq.seq");
std::cout << cur << " " << ans << " " << getLongestSingleRepeat(seq) << std::endl;
}
}
int main()
{
string re;
cout << "input regex:";
cin >> re;
DFA dfa(re);
string text;
cout << "input text:";
cin >> text;
dfa.match(text);
return 0;
}
int main()
{
char re[20];
tnode *head;
printf("\n enter a RE expression : ");
scanf("%s",re);
head = createTree(re);
printf("\n\n Ind Char NUL First Last Follow ");
listInPostOrder(head);
printf("\n\n");
dfa(head);
return 0;
}
void FSMThingy::finalizeGraph(bool determinize) {
if (Fsm->verticesSize() < 2) {
throw std::runtime_error("No valid patterns were parsed");
}
if (determinize && !Fsm->Deterministic) {
NFAPtr dfa(new NFA(1, 2 * Fsm->verticesSize(), Fsm->edgesSize()));
dfa->TransFac = Fsm->TransFac;
Comp.subsetDFA(*dfa, *Fsm);
Fsm = dfa;
}
Comp.labelGuardStates(*Fsm);
}
void run_nums_test()
{
#define NODES_NUM 10
TestDG *d = create_circular_graph(NODES_NUM);
// B is pointer to the last node, but since the graph is a circle,
// it doesn't matter what BB we'll use
DataFlowA dfa(d->getEntryBB(), no_change);
dfa.run();
for (int i = 0; i < NODES_NUM; ++i) {
check(d->getNode(i)->counter == 1,
"did not go through the node only one time but %d",
d->getNode(i)->counter);
// zero out the counter for next dataflow run
d->getNode(i)->counter = 0;
}
const analysis::DataFlowStatistics& stats = dfa.getStatistics();
check(stats.getBBlocksNum() == NODES_NUM, "wrong number of blocks: %d",
stats.getBBlocksNum());
check(stats.processedBlocks == NODES_NUM,
"processed more blocks than %d - %d", NODES_NUM, stats.processedBlocks);
check(stats.getIterationsNum() == 1, "did wrong number of iterations: %d",
stats.getIterationsNum());
DataFlowA dfa2(d->getEntryBB(), one_change);
dfa2.run();
for (int i = 0; i < NODES_NUM; ++i) {
check(d->getNode(i)->counter == 2,
"did not go through the node only one time but %d",
d->getNode(i)->counter);
}
const analysis::DataFlowStatistics& stats2 = dfa2.getStatistics();
check(stats2.getBBlocksNum() == NODES_NUM, "wrong number of blocks: %d",
stats2.getBBlocksNum());
check(stats2.processedBlocks == 2*NODES_NUM,
"processed more blocks than %d - %d", 2*NODES_NUM, stats2.processedBlocks);
check(stats2.getIterationsNum() == 2, "did wrong number of iterations: %d",
stats2.getIterationsNum());
#undef NODES_NUM
}
void main()
{
//clrscr();
//system("clear");
char str[500];
inpt[0]=NULL;
printf("Enter the postfix expression\n");
scanf("%s",str);
node * root;
int l;
strcat(str,"#.\0");
l=strlen(str);
l--;
int i, j=0;
for(i=0;i<l-1;++i) {
j=0;
while(inpt[j]!=NULL)
{
if(inpt[j]==str[i])
break;
j++;
}
if(inpt[j]!=str[i] && str[i]!='|' && str[i]!='*' && str[i]!='.')
{
inpt[j]=str[i];
inpt[j+1]=NULL;
}
}
int pos=1;
root=create(str,&l);
create_nullable(root,&pos);
printf("NULLABLE TABLE\nElement\tFPOS\tLPOS\n");
print_nullable(root->lc);
print_follow(pos-2);
dfa();
display_dfa();
printf("\n");
}
static void dump_ex(const typename char_state_machine::dfa &dfa_,
ostream &stream_)
{
const std::size_t states_ = dfa_._states.size();
const id_type bol_index_ = dfa_._bol_index;
typename dfa_state::id_type_string_token_map::const_iterator iter_;
typename dfa_state::id_type_string_token_map::const_iterator end_;
for (std::size_t i_ = 0; i_ < states_; ++i_)
{
const dfa_state &state_ = dfa_._states[i_];
state(stream_);
stream_ << i_ << std::endl;
if (state_._end_state)
{
end_state(stream_);
if (state_._push_pop_dfa == dfa_state::push_dfa)
{
push(stream_);
stream_ << state_._push_dfa;
}
else if (state_._push_pop_dfa == dfa_state::pop_dfa)
{
pop(stream_);
}
id(stream_);
stream_ << static_cast<std::size_t>(state_._id);
user_id(stream_);
stream_ << static_cast<std::size_t>(state_._user_id);
dfa(stream_);
stream_ << static_cast<std::size_t>(state_._next_dfa);
stream_ << std::endl;
}
if (i_ == 0 && bol_index_ != char_state_machine::npos())
{
bol(stream_);
stream_ << static_cast<std::size_t>(bol_index_) << std::endl;
}
if (state_._eol_index != char_state_machine::npos())
{
eol(stream_);
stream_ << static_cast<std::size_t>(state_._eol_index) <<
std::endl;
}
iter_ = state_._transitions.begin();
end_ = state_._transitions.end();
for (; iter_ != end_; ++iter_)
{
string_token token_ = iter_->second;
open_bracket(stream_);
if (!iter_->second.any() && iter_->second.negatable())
{
token_.negate();
negated(stream_);
}
string chars_;
typename string_token::range_vector::const_iterator
ranges_iter_ = token_._ranges.begin();
typename string_token::range_vector::const_iterator
ranges_end_ = token_._ranges.end();
for (; ranges_iter_ != ranges_end_; ++ranges_iter_)
{
if (ranges_iter_->first == '-' ||
ranges_iter_->first == '^' ||
ranges_iter_->first == ']')
{
stream_ << '\\';
}
chars_ = string_token::escape_char
(ranges_iter_->first);
if (ranges_iter_->first != ranges_iter_->second)
{
if (ranges_iter_->first + 1 < ranges_iter_->second)
{
chars_ += '-';
}
if (ranges_iter_->second == '-' ||
ranges_iter_->second == '^' ||
ranges_iter_->second == ']')
{
stream_ << '\\';
}
chars_ += string_token::escape_char
(ranges_iter_->second);
}
stream_ << chars_;
}
close_bracket(stream_);
stream_ << static_cast<std::size_t>(iter_->first) <<
std::endl;
}
stream_ << std::endl;
}
}
/**
* Convert an nfa to a dfa
*/
void convert_nfa_dfa( Automata* nfa )
{
std::vector<bool> marked;
std::vector<TempState> dfaStates;
/**
* -- Algorithm --
*
* while unmarked states exist:
* if empty ( dfaStates ):
* curTempState.anchors = e-closure(nfa.Start)
* dfaStates.addState(curTempState)
* markedState = curTempState
* else:
* foreach a in Alphabet:
* newState_i.anchors = e-closure(a-closure(markedState))
* if not dfaStates.contains(newState_i):
* dfaStates.addState(newState_i)
* dfaStates.addTrans(markedState, newstate_i)
* else:
* dfaStates.addTrans(markedState, dfaStates.get(newState_i))
*
* if state was not last:
* mark dfa.nextUnmarkedState
*/
int curMarked = -1;
do
{
if (dfaStates.size() == 0) {
TempState DFAStartState;
DFAStartState.anchors = nfa->EClosure({nfa->Start()});
dfaStates.push_back(DFAStartState);
marked.push_back(false);
curMarked = 0;
// -- Output check --
std::cout << "E-closure(IO) = "; _print_vec(DFAStartState.anchors);
std::cout << " = " << curMarked+1 << "\n";
std::cout << "\nMark " << curMarked+1 << std::endl;
// -- End output check --
}
else
{
for (int i = 0; i < nfa->Alphabet().length()-1; i++) {
TempState cmSymClosure, newState;
cmSymClosure.anchors = nfa->SymClosure(dfaStates[curMarked].anchors,
nfa->Alphabet(i)),
newState.anchors = nfa->EClosure({cmSymClosure.anchors});
// make sure the new state doens't already exist
int spos = _find_state(newState.anchors, dfaStates);
if (newState.anchors.size() > 0 && spos < 0) {
dfaStates.push_back(newState);
marked.push_back(false);
dfaStates[curMarked].next[nfa->Alphabet(i)] = dfaStates.size()-1;
} else {
dfaStates[curMarked].next[nfa->Alphabet(i)] = spos;
}
// -- Output check --
if (newState.anchors.size() != 0) {
_print_vec(dfaStates[curMarked].anchors);
std::cout << "--" << nfa->Alphabet(i) << "--> ";
_print_vec(cmSymClosure.anchors);
std::cout << "\nE-closure";
_print_vec(cmSymClosure.anchors);
std::cout << "= ";
_print_vec(newState.anchors);
std::cout << " = " << (spos >= 0 ? spos+1 : dfaStates.size()) << "\n";
}
// -- End output check --
}
marked[curMarked++] = true;
// -- Output check
if (curMarked < dfaStates.size()) {
std::cout << "\nMark " << curMarked+1 << std::endl;
}
// -- End output check --
}
}
while (std::find(marked.begin(),marked.end(), false) != marked.end());
// Trim Epsilon from alphabet
std::string dfaAlpha = nfa->Alphabet().substr(0,nfa->Alphabet().length()-1);
// Load DFA
int dfaStart = 0;
std::vector<int> dfaFinalStates;
for (int i = 0; i<nfa->FinalStates().size(); i++) {
for (int j = 0; j<dfaStates.size(); j++) {
// auto: because typing iterator types sucks sometimes
auto start = dfaStates[j].anchors.begin(),
end = dfaStates[j].anchors.end();
if (find(start, end, nfa->FinalStates()[i]) != end) {
dfaFinalStates.push_back(j);
}
}
}
// Add transitions
Automata dfa(dfaAlpha, dfaStart, dfaStates.size(), dfaFinalStates);
for (int i = 0; i<dfaStates.size(); i++) {
for (int j = 0; j<dfa.Alphabet().length(); j++) {
if (dfaStates[i].next[dfa.Alphabet().at(j)] >= 0)
dfa.AddTrans(i, dfaStates[i].next[dfa.Alphabet().at(j)], dfa.Alphabet().at(j));
}
}
*nfa = dfa;
}
void run_nums_test_interproc()
{
#define NODES_NUM 5
TestDG *d = create_circular_graph(NODES_NUM);
TestNode *last;
for (auto It : *d) {
TestDG *sub = create_circular_graph(NODES_NUM);
It.second->addSubgraph(sub);
}
// B is pointer to the last node, but since the graph is a circle,
// it doesn't matter what BB we'll use
DataFlowA dfa(d->getEntryBB(), no_change);
dfa.run();
for (int i = 0; i < NODES_NUM; ++i) {
TestNode *n = d->getNode(i);
check(n->counter == 1,
"did not go through the node only one time but %d", n->counter);
// check that subgraphs are untouched by the dataflow
// analysis
for (auto sub : n->getSubgraphs()) {
// iterate over nodes
for (auto It : *sub) {
TestNode *n = It.second;
check(n->counter == 0,
"intrAproc. dataflow went to procedures (%d - %d)",
n->getKey(), n->counter);
TestBBlock *BB = n->getBBlock();
assert(BB);
check(BB->getDFSOrder() == 0, "DataFlow went into subgraph blocks");
}
}
// zero out the counter for next dataflow run
n->counter = 0;
}
// this did not go into the procedures, so we should have only
// the parent graph
const analysis::DataFlowStatistics& stats = dfa.getStatistics();
check(stats.getBBlocksNum() == NODES_NUM, "wrong number of blocks: %d",
stats.getBBlocksNum());
check(stats.processedBlocks == NODES_NUM,
"processed more blocks than %d - %d", NODES_NUM, stats.processedBlocks);
check(stats.getIterationsNum() == 1, "did wrong number of iterations: %d",
stats.getIterationsNum());
DataFlowA dfa2(d->getEntryBB(), one_change,
analysis::DATAFLOW_INTERPROCEDURAL | analysis::DATAFLOW_BB_NO_CALLSITES);
dfa2.run();
for (int i = 0; i < NODES_NUM; ++i) {
TestNode *n = d->getNode(i);
check(n->counter == 2,
"did not go through the node only one time but %d", n->counter);
// check that subgraphs are untouched by the dataflow
// analysis
for (auto sub : n->getSubgraphs()) {
// iterate over nodes
for (auto It : *sub) {
TestNode *n = It.second;
check(n->counter == 2,
"intErproc. dataflow did NOT went to procedures (%d - %d)",
n->getKey(), n->counter);
TestBBlock *BB = n->getBBlock();
assert(BB);
check(BB->getDFSOrder() != 0,
"intErproc DataFlow did NOT went into subgraph blocks");
n->counter = 0;
}
}
// zero out the counter for next dataflow run
n->counter = 0;
}
// we have NODES_NUM nodes and each node has subgraph of the
// same size + the blocks in parent graph
// we don't go through the parameters!
uint64_t blocks_num = (NODES_NUM + 1) * NODES_NUM;
const analysis::DataFlowStatistics& stats2 = dfa2.getStatistics();
check(stats2.getBBlocksNum() == blocks_num, "wrong number of blocks: %d",
stats2.getBBlocksNum());
check(stats2.processedBlocks == 2*blocks_num,
"processed more blocks than %d - %d", 2*blocks_num, stats2.processedBlocks);
check(stats2.getIterationsNum() == 2, "did wrong number of iterations: %d",
stats2.getIterationsNum());
// BBlocks now keep call-sites information, so now
// this should work too
DataFlowA dfa3(d->getEntryBB(), one_change,
analysis::DATAFLOW_INTERPROCEDURAL);
dfa3.run();
for (int i = 0; i < NODES_NUM; ++i) {
TestNode *n = d->getNode(i);
check(n->counter == 2,
"did not go through the node only one time but %d", n->counter);
// check that subgraphs are untouched by the dataflow
// analysis
for (auto sub : n->getSubgraphs()) {
// iterate over nodes
for (auto It : *sub) {
TestNode *n = It.second;
check(n->counter == 2,
"intErproc. dataflow did NOT went to procedures (%d - %d)",
n->getKey(), n->counter);
TestBBlock *BB = n->getBBlock();
assert(BB);
check(BB->getDFSOrder() != 0,
"intErproc DataFlow did NOT went into subgraph blocks");
n->counter = 0;
}
}
// zero out the counter for next dataflow run
n->counter = 0;
}
// we have NODES_NUM nodes and each node has subgraph of the
// same size + the blocks in parent graph
const analysis::DataFlowStatistics& stats3 = dfa3.getStatistics();
check(stats3.getBBlocksNum() == blocks_num, "wrong number of blocks: %d",
stats3.getBBlocksNum());
check(stats3.processedBlocks == 2*blocks_num,
"processed more blocks than %d - %d", 2*blocks_num, stats3.processedBlocks);
check(stats3.getIterationsNum() == 2, "did wrong number of iterations: %d",
stats3.getIterationsNum());
#undef NODES_NUM
}
static void dump (const basic_state_machine<CharT> &state_machine_, ostream &stream_)
{
typename basic_state_machine<CharT>::iterator iter_ =
state_machine_.begin ();
typename basic_state_machine<CharT>::iterator end_ =
state_machine_.end ();
for (std::size_t dfa_ = 0, dfas_ = state_machine_.size ();
dfa_ < dfas_; ++dfa_)
{
const std::size_t states_ = iter_->states;
for (std::size_t i_ = 0; i_ < states_; ++i_)
{
state (stream_);
stream_ << i_ << std::endl;
if (iter_->end_state)
{
end_state (stream_);
stream_ << iter_->id;
dfa (stream_);
stream_ << iter_->goto_dfa;
stream_ << std::endl;
}
if (iter_->bol_index != npos)
{
bol (stream_);
stream_ << iter_->bol_index << std::endl;
}
if (iter_->eol_index != npos)
{
eol (stream_);
stream_ << iter_->eol_index << std::endl;
}
const std::size_t transitions_ = iter_->transitions;
if (transitions_ == 0)
{
++iter_;
}
for (std::size_t t_ = 0; t_ < transitions_; ++t_)
{
std::size_t goto_state_ = iter_->goto_state;
if (iter_->token.any ())
{
any (stream_);
}
else
{
open_bracket (stream_);
if (iter_->token._negated)
{
negated (stream_);
}
string charset_;
CharT c_ = 0;
escape_control_chars (iter_->token._charset,
charset_);
c_ = *charset_.c_str ();
if (!iter_->token._negated &&
(c_ == '^' || c_ == ']'))
{
stream_ << '\\';
}
stream_ << charset_;
close_bracket (stream_);
}
stream_ << goto_state_ << std::endl;
++iter_;
}
stream_ << std::endl;
}
}
}
int main(int argc, char* argv[])
{
std::string regex;
std::string fileName;
std::string dfaFileName;
for( int i = 1 ; i < argc ; ++i){
if( strcmp(argv[i], "-r") == 0 ){
REGEX = true;
regex = argv[++i];
continue;
}
if( strcmp(argv[i], "-f") == 0 ) {
WORDS_FILE = true;
fileName = argv[++i];
continue;
}
if( strcmp(argv[i], "-st") == 0 ) {
SHOW_TREE = true;
continue;
}
if( strcmp(argv[i], "-sa") == 0 ) {
SHOW_DFA = true;
dfaFileName = argv[++i];
continue;
}
HELP = true;
break;
}
if( HELP ) {
std::cout << "HELP" << std::endl;
std::cout << std::endl;
std::cout << "OPCJE : " << std::endl;
std::cout << " -f po tej faldze podać nazwę pliku z lancuchami do sprawdzenia na" << std::endl;
std::cout << " przynaleznosc do języka generowanego przez wyrazenie regularne." << std::endl;
std::cout << " -r po tej faldze podac wyrazenie regularne." << std::endl;
std::cout << " -st flaga oznaczająca wygenerowanie pliku z graficznym przedstawieniem" << std::endl;
std::cout << " drzewa rozbioru wyrazenia regularnego." << std::endl;
std::cout << " -sa flaga oznaczająca wygenerowanie pliku z tekstowym przedstawieniem" << std::endl;
std::cout << " automatu wyrazenia regularnego." << std::endl;
return 0;
}
if( !REGEX ) {
std::cerr << "You have to specify regex to do anything!" << std::endl;
exit(-1);
}
std::cout << "REGEX : " << regex << std::endl;
Scanner scanner(regex);
try {
scanner.tokenize();
} catch (ScannerException & e) {
std::cerr << e.what() << std::endl;
exit(-1);
}
Syntax syntax(scanner.getTokens());
try {
syntax.buildTree();
} catch (SyntaxException & e) {
std::cerr << e.what() << std::endl;
exit(-1);
}
DFA dfa(syntax.getTree());
if(WORDS_FILE){
std::ifstream words;
words.open(fileName);
std::string chain;
while(std::getline(words, chain)) {
std::cout << chain << " => ";
try {
scanner.tokenize(chain);
} catch (ScannerException & e) {
std::cerr << e.what() << std::endl;
continue;
}
if( dfa.checkWord(scanner.getTokens()) ){
std::cout << "TRUE" << std::endl;
} else {
std::cout << "FALSE" << std::endl;
}
}
}
if(SHOW_TREE){
syntax.showTree();
}
if(SHOW_DFA){
std::ofstream file;
file.open(dfaFileName);
file << dfa;
file.close();
}
return 0;
}
