int
main(int argc, char *argv[])
{
ECArgs args( argc, argv );
/* o = basic, but not debugging, output.
l = length of sentence to be proceeds 0-40 is default
n = work on each #'th line.
d = print out debugging info at level #
W = use wwclasses
R = use rwclasses
t = report timings (requires o)
s = maximum sleep time
f = f# says multiply ctl2 counts by #
p = p# use prepFactor #
P = which types of prob models to use */
// prevent core file creation;
struct rlimit core_limits;
core_limits.rlim_cur = 0;
core_limits.rlim_max = 0;
setrlimit( RLIMIT_CORE, &core_limits );
params.init( args );
if(args.isset('s'))
{
int maxDelay = atoi(args.value('s').c_str());
srand(params.whichSent());
int randN = rand();
int delay = randN%maxDelay;
sleep(delay);
}
if(args.isset('T'))
{
int fac = atoi(args.value('T').c_str());
float ffac = (float)fac;
ffac /= 10;
Bchart::timeFactor = ffac;
}
int maxSentLen = 70;
if(args.isset('l'))
{
maxSentLen = atoi(args.value('l').c_str());
}
int totEdges = 0;
int totPopedEdges = 0;
double totAccessTime = 0;
double totParseTime = 0;
double totSemParseTime = 0;
clock_t lastTime, currTime;
double lastTimeSec, currTimeSec, elapsedTime;
endFactor = 1.2;
midFactor = (1.0 - (.3684 * endFactor))/(1.0 - .3684);
if( args.nargs() > 2 || args.nargs() == 0 ) // require path name
error( "Need exactly two arg." );
ECString path( args.arg( 0 ) );
readHeadInfo(path);
Term::init( path );
InputTree::init();
ECString testSString( args.arg(1) );
ewDciTokStrm testSStream(testSString);
//ifstream testSStream(testSString.c_str());
if( !testSStream ) error( "No testSstream" );
int sentenceCount = 0; //counts all sentences so we can use 1/50;
ECString probSumString( path );
probSumString += "pSgT.txt";
ifstream probSumStream( probSumString.c_str() );
if( !probSumStream ) error( "Failed to find probSum file" );
Bchart::readTermProbs(path);
if( args.isset('d') )
{
int lev = atoi(args.value('d').c_str());
Bchart::printDebug() = lev;
}
int totSents = 0;
int totUnparsed = 0;
MeChart::init(path);
Bchart::setPosStarts();
for( ; !(!testSStream) ; )
{
SentRep sr(testSStream, SentRep::SGML);
int len = sr.length();
if(len == 0) continue;
if(len > maxSentLen) continue;
if( !params.field().in(sentenceCount) )
{
sentenceCount++;
continue;
}
if(len == 1)
{
if(sr[0].lexeme() == "</DOC>")
{
continue;
}
}
sentenceCount++;
//SentRep orgsr( wtList ); // used in precision calc;
if( args.isset('t') ) lastTime = clock();
if(args.isset('t') )
{
currTime = clock();
lastTimeSec = (double)lastTime/(double)CLOCKS_PER_SEC;
currTimeSec = (double)currTime/(double)CLOCKS_PER_SEC;
elapsedTime = currTimeSec - lastTimeSec;
if(elapsedTime < 0) elapsedTime += 2147;
cerr << "Reading data time = " << elapsedTime << endl;
totAccessTime += elapsedTime;
lastTime = currTime;
}
MeChart* chart = new MeChart( sr );
curChart = chart;
chart->ruleCountTimeout() = 250000;
totSents++;
if(args.isset('t') )
lastTime = clock();
double tmpCrossEnt = chart->parse( );
Item* topS = chart->topS();
if(!topS)
{
if(len == 1)
{
delete chart;
continue;
}
Edge::DemFac = .9;
delete chart;
chart = new MeChart(sr);
chart->ruleCountTimeout() = 350000;
curChart = chart;
tmpCrossEnt = chart->parse( );
topS = chart->topS();
Edge::DemFac = .999;
if(!topS)
{
totUnparsed++;
cerr << "Parse failed on: " << sr << endl;
delete chart;
continue;
}
}
// compute the outside probabilities on the items so that we can
// skip doing detailed computations on the really bad ones
if(args.isset('t') )
{
currTime = clock();
lastTimeSec = (double)lastTime/(double)CLOCKS_PER_SEC;
currTimeSec = (double)currTime/(double)CLOCKS_PER_SEC;
elapsedTime = currTimeSec - lastTimeSec;
if(elapsedTime < 0) elapsedTime += 2147;
cerr << "Parsing time = " << elapsedTime
<< "\tEdges created = " << chart->totEdgeCountAtS()
<< "\tEdges poped = " << chart->popedEdgeCountAtS() << endl;
totParseTime += elapsedTime;
//totEdges += chart->totEdgeCountAtS();
//totPopedEdges += chart->popedEdgeCountAtS();
totEdges += chart->totEdgeCountAtS();
totPopedEdges += chart->popedEdgeCountAtS();
lastTime = clock();
}
chart->set_Alphas();
AnswerTree* at = chart->findMapParse();
if( !at )
{
totUnparsed++;
cerr << "MapParse failed on: " << sr << endl;
delete chart;
continue;
}
InputTree* mapparse = inputTreeFromAnswerTree(at,topS);
//at->deleteSubTrees();
//delete at;
cout << *mapparse << endl;
delete mapparse;
if(args.isset('t') )
{
currTime = clock();
lastTimeSec = (double)lastTime/(double)CLOCKS_PER_SEC;
currTimeSec = (double)currTime/(double)CLOCKS_PER_SEC;
elapsedTime = currTimeSec - lastTimeSec;
if(elapsedTime < 0) elapsedTime += 2147;
cerr << "Sem Parsing time = " << elapsedTime << endl;
totSemParseTime += elapsedTime;
}
delete chart;
}
if( args.isset('t') )
cout << "Av access time = " << totAccessTime/totSents
<< "\t Av parse time = "
<< totParseTime/totSents
<< "\t Av stats time = "
<< totSemParseTime/totSents
<< "\nAv edges created = "
<< (float)totEdges/totSents
<< "\tAv edges poped = "
<< (float)totPopedEdges/totSents
<< endl;
return 0;
}
int
main(int argc, char *argv[])
{
ECArgs args( argc, argv );
//AnsTreeHeap::print = true;
/* o = basic, but not debugging, output.
l = length of sentence to be proceeds 0-40 is default
n = work on each #'th line.
d = print out debugging info at level #
t = report timings (requires o)
*/
int i;
params.init( args );
//cerr << "Starting wwBCTest " << Feature::sLM << endl;
if( args.nargs() > 2 || args.nargs() < 2 ) // require path name
error( "Need exactly two args.");
ECString path( args.arg( 0 ) );
generalInit(path);
for(int i = 0 ; i < 500 ; i++) histPoints[i] = false;
histPoints[0] = true;
if (Bchart::Nth == 50 || Bchart::Nth == 500 || Bchart::Nth == 1000)
histPoints[1] = histPoints[9] = histPoints[24] = histPoints[49] = true;
if (Bchart::Nth == 500 || Bchart::Nth == 1000)
histPoints[99] = histPoints[249] = histPoints[499] = true;
if(Bchart::Nth == 1000)
histPoints[749] = histPoints[999] = true;
TimeIt timeIt;
ECString testSString = args.arg(1);
ifstream testSStream(testSString.c_str());
if( !testSStream )
{
cerr << "Could not find " << testSString << endl;
error( "No testSstream");
}
ECString pstatStreamName( params.fileString());
pstatStreamName += "PStatInfo/pStat";
pstatStreamName += params.numString();
pstatStreamName += ".txt";
ofstream pstatStream( pstatStreamName.c_str(), ios::out);
if( !pstatStream )
{
cerr << "Looking to output to " << pstatStreamName << endl;
error( "unable to open pstat stream");
}
for(i = 0 ; i < MAXNUMTHREADS ; i++) globalGi[i] =NULL;
pthread_t thread[MAXNUMTHREADS];
loopArg lA[MAXNUMTHREADS];
for(i = 0 ; i < numThreads ; i++){
lA[i].id = i;
lA[i].inpt=&testSStream;
lA[i].outpt=&pstatStream;
pthread_create(&thread[i],0,mainLoop, (void*)&lA[i]);
}
for(i=0; i<numThreads; i++){
pthread_join(thread[i],0);
}
for(int i = 0 ; i < Bchart::Nth ; i++)
if(histPoints[i])
{
cerr << i << " " << totPst[i].fMeasure() << "\t";
}
cerr << endl;
if(Feature::isLM)
{
cerr << pow(2.0,totGram/(double)totWords);
cerr <<"\t" << pow(2.0,totTri/(double)totWords);
cerr << "\t" << pow(2.0,totMix/(double)(totWords));
cerr << endl;
}
if( args.isset('t') ) timeIt.finish(sentenceCount);
pthread_exit(0);
return 0;
}
