/*!
* Explicit constructor.
*/
AnnotatorPerformanceAnnotation::AnnotatorPerformanceAnnotation(CAS& aCas, const UnicodeString& annotatorName, long elapsedMillis)
: AnnotationWrapper(aCas)
{
FSIndexRepository& indexRep = aCas.getIndexRepository();
annotation = aCas.createAnnotation(tAnnotatorPerformanceAnnotation, 0, 0);
setComponentName(annotatorName);
setElapsedTime(elapsedMillis);
indexRep.addFS(annotation);
}
SentenceAnnotation::SentenceAnnotation(CAS& aCas, size_t begin, size_t end, const TokenAnnotation& firstToken, const TokenAnnotation& lastToken) :
ContextAreaAnnotation(aCas)
{
FSIndexRepository& indexRep = aCas.getIndexRepository();
annotation = aCas.createAnnotation(tSentenceAnnotation, begin, end);
setFirstToken(firstToken);
setLastToken(lastToken);
indexRep.addFS(annotation);
}
void testCallingSequence1(uima::util::ConsoleUI& rclConsole, const TCHAR* cpszConfigFilename)
/* ----------------------------------------------------------------------- */
{
ErrorInfo errInfo;
uima::TextAnalysisEngine* pEngine = TextAnalysisEngine::createTextAnalysisEngine(cpszConfigFilename, errInfo);
failIfNotTrue(errInfo.getErrorId() == UIMA_ERR_NONE);
failIfNotTrue(pEngine != NULL);
CAS* cas = pEngine->newCAS();
failIfNotTrue(cas != NULL);
uima::UnicodeStringRef us(UnicodeString("a"));
// UnicodeStringRef uRef(us);
rclConsole.formatHeader(_TEXT("testing Engine CallingSequence1"));
cas->setDocumentText(us.getBuffer(), us.length());
cas->getDocumentAnnotation().setLanguage("en");
failIfNotTrue(pEngine->process(*cas) == UIMA_ERR_NONE);
failIfNotTrue(cas->reset() == UIMA_ERR_NONE);
failIfNotTrue(pEngine->destroy() == UIMA_ERR_NONE);
cas->setDocumentText(us.getBuffer(), us.length());
cas->getDocumentAnnotation().setLanguage("en");
failIfNotTrue(pEngine->process(*cas) == UIMA_ERR_ENGINE_INVALID_CALLING_SEQUENCE);
TyErrorId deInitRC = pEngine->destroy();
rclConsole.format("RC of deInit()", deInitRC);
failIfNotTrue(deInitRC == UIMA_ERR_ENGINE_INVALID_CALLING_SEQUENCE);
rclConsole.formatBool(_TEXT("testing Engine CallingSequence1 OK"),
true); //lint !e944: argument for operator '!' always evaluates to False
delete cas;
delete pEngine;
} //lint !e715: cpszConfigFilename (line 99) not referenced
void testCasMultiplier(uima::util::ConsoleUI& rclConsole)
/* ----------------------------------------------------------------------- */
{
rclConsole.info("testCasMultiplier start.");
uima::TextAnalysisEngine* pEngine;
ErrorInfo errInfo;
UnicodeString filename("SimpleTextSegmenter.xml");
UnicodeString fn = ResourceManager::resolveFilename(filename, filename);
pEngine = TextAnalysisEngine::createTextAnalysisEngine(UnicodeStringRef(fn).asUTF8().c_str(), errInfo);
failIfNotTrue(errInfo.getErrorId() == UIMA_ERR_NONE);
failIfNotTrue(pEngine != NULL);
//test operational properties settings
failIfNotTrue(pEngine->getAnalysisEngineMetaData().getOperationalProperties()->getOutputsNewCASes() == true);
failIfNotTrue(pEngine->getAnalysisEngineMetaData().getOperationalProperties()->getModifiesCas() == false);
failIfNotTrue(
pEngine->getAnalysisEngineMetaData().getOperationalProperties()->isMultipleDeploymentAllowed() == true);
CAS* cas = pEngine->newCAS();
cas->setDocumentText(
UnicodeString("This is the first sentence. This is the second sentence. This is the third sentence."));
CASIterator iter = pEngine->processAndOutputNewCASes(*cas);
int num = 0;
while (iter.hasNext()) {
num++;
CAS& seg = iter.next();
failIfNotTrue(seg.getDocumentText().length() > 0);
pEngine->getAnnotatorContext().releaseCAS(seg);
}
failIfNotTrue(num == 3);
delete pEngine;
delete cas;
rclConsole.info("testCasMultiplier finished.");
}
void testProcessTerm(uima::util::ConsoleUI& rclConsole,
uima::TextAnalysisEngine& rclEngine,
///const uima::CCSID & crclCCSID,
const char* crclCCSID,
const uima::Language& crclLanguage,
const TCHAR* cpszInpTerm)
/* ----------------------------------------------------------------------- */
{
TyErrorId utErrorId;
failIfNotTrue(EXISTS(cpszInpTerm));
rclConsole.format(_TEXT("Input term"), cpszInpTerm);
DocBuffer docBuffer;
docBuffer.addDocPart(cpszInpTerm, strlen(cpszInpTerm), crclCCSID);
//? assert(false);
CAS* cas = rclEngine.newCAS();
failIfNotTrue(cas != NULL);
// For terms we always add a term annotation for the whole "document"
/* since we already added a complete doc, we may not add anything else */
cas->setDocumentText(docBuffer.getDocBuffer(), docBuffer.getLength());
cas->getDocumentAnnotation().setLanguage(crclLanguage);
utErrorId = rclEngine.process(*cas);
uimaToolHandleErrorId(rclConsole, utErrorId, rclEngine.getAnnotatorContext().getLogger().getLastErrorAsCStr(),
_TEXT("uima::Engine::processDocument"), gs_lExpectedProcessDocumentRc);
if (utErrorId == UIMA_ERR_NONE && gs_bDoIterTest) {
failIfNotTrue(false);
// iteratorTest(rclConsole, rclEngine);
}
utErrorId = cas->reset();
uimaToolHandleErrorId(rclConsole, utErrorId, rclEngine.getAnnotatorContext().getLogger().getLastErrorAsCStr(),
_TEXT("uima::Engine::resetDocument"));
delete cas;
}
TyErrorId
AnnotatorDump::process(CAS & tcas,
const ResultSpecification &
) {
TyErrorId tyErrId;
// in append mode all data in a session/collection is dumped into one file
// otherwise the same dump file is deleted and rewritten for each document
// in the session/collection
if (!iv_bAppendFile) {
tyErrId = openOutputFile();
if (tyErrId != UIMA_ERR_NONE) {
return tyErrId;
}
}
assert(iv_clOutputStream.good());
if (iv_bDumpDocBuffer) {
// Dumping the Document Buffer
UnicodeStringRef doc = tcas.getDocumentText();
outputDocBuffer(doc);
}
uima::CASWriterABase * writer = NULL;
switch (iv_enOutputStyle) {
case Xml:
writer = new uima::XMLDumpWriter(tcas, iv_bDumpDocBuffer);
break;
case XCas:
writer = new uima::XCASWriter(tcas, iv_bDumpDocBuffer);
break;
default:
assert(false);
}
assert( EXISTS(writer) );
auto_ptr<CASWriterABase> apWriter( writer );
apWriter->write(iv_clOutputStream);
// in append mode all data in a session/collection is dumped into one file
// otherwise the same dump file is deleted and rewritten for each document
// in the session/collection
if (!iv_bAppendFile) {
closeOutputFile();
}
return(TyErrorId)UIMA_ERR_NONE;
}
void testCallingSequence3(uima::util::ConsoleUI& rclConsole, const TCHAR* cpszConfigFilename)
/* ----------------------------------------------------------------------- */
{
uima::TextAnalysisEngine* pEngine = NULL;
uima::Language clLanguage(MAIN_DEFAULT_LANG);
const char* clCCSID = MAIN_DEFAULT_CCSID_STR;
TyErrorId utErrorId;
UnicodeString us("a");
UnicodeStringRef uref(us);
rclConsole.formatHeader(_TEXT("testing Engine CallingSequence3"));
ErrorInfo errInfo;
pEngine = TextAnalysisEngine::createTextAnalysisEngine(cpszConfigFilename, errInfo);
failIfNotTrue(errInfo.getErrorId() == UIMA_ERR_NONE);
failIfNotTrue(pEngine != NULL);
CAS* cas = pEngine->newCAS();
failIfNotTrue(cas != NULL);
/* test for NULL ptrs */
UnicodeStringRef uref2(NULL);
cas->setDocumentText(uref2.getBuffer(), uref2.length());
cas->getDocumentAnnotation().setLanguage("en");
failIfNotTrue(pEngine->process(*cas) == UIMA_ERR_NONE);
failIfNotTrue(cas->reset() == UIMA_ERR_NONE);
/* test for subsequent processes */
cas->setDocumentText(uref2.getBuffer(), uref2.length());
cas->getDocumentAnnotation().setLanguage("en");
failIfNotTrue(pEngine->process(*cas) == UIMA_ERR_NONE);
failIfNotTrue(pEngine->process(*cas) == UIMA_ERR_NONE);
utErrorId = pEngine->destroy();
failIfNotTrue(utErrorId == UIMA_ERR_NONE);
delete cas;
delete pEngine;
rclConsole.formatBool(_TEXT("testing Engine CallingSequence3 OK"),
true); //lint !e944: argument for operator '!' always evaluates to False
}
int main(int argc, char* argv[]) {
try {
int loglevel = -1;
/* check the number of command line args */
if (argc != 3 && argc != 5) {
tell();
return 1;
}
if (argc == 5) {
if (!strcmp(argv[3], "-l")) {
loglevel = atoi(argv[4]);
if (loglevel < LogStream::EnMessage) {
cerr << "LogLevel less than minimum value (Message) = " << LogStream::EnMessage << endl;
return 1;
}
if (loglevel > LogStream::EnError) {
cerr << "LogLevel greater than maximum value (Error) = " << LogStream::EnError << endl;
return 1;
}
} else {
cerr << "Inexpected option: " << argv[3] << endl;
tell();
return 1;
}
}
/* Create/link up to a resource manager instance (singleton) */
(void) ResourceManager::createInstance("UIMACPP_EXAMPLE_APPLICATION");
if (loglevel >= 0) {
ResourceManager::getInstance().setLoggingLevel((LogStream::EnEntryType) loglevel);
}
std::string a = "abc";
UnicodeString b = "wxyz";
UChar c = 'c';
TyErrorId utErrorId; // Variable to store return codes
ErrorInfo errorInfo; // Variable to stored detailed error info
/* Initialize engine with filename of config-file */
AnalysisEngine* pEngine = Framework::createAnalysisEngine(argv[1], errorInfo);
CheckError(errorInfo);
/* Get a new CAS */
CAS* tcas = pEngine->newCAS();
/* process input xcas */
util::DirectoryWalk dirwalker(argv[2]);
if (dirwalker.isValid()) {
util::Filename infile(argv[2], "FilenamePlaceHolder");
while (dirwalker.isValid()) {
// Process all files or just the ones with matching suffix
if (dirwalker.isFile() &&
dirwalker.matchesWildcardPattern("*.txt")) {
infile.setNewName(dirwalker.getNameWithoutPath());
std::string afile(infile.getAsCString());
//process the file
processFile(afile, pEngine, tcas);
//reset the cas
tcas->reset();
}
//get the next xcas file in the directory
dirwalker.setToNext();
}
} else {
/* If has no directory entries then probably a file */
cout << "ExampleApplication: processing file " << argv[2] << endl;
std::string afile(argv[2]);
//process the cas
processFile(afile, pEngine, tcas);
}
/* call collectionProcessComplete */
utErrorId = pEngine->collectionProcessComplete();
/* Free ressorces */
utErrorId = pEngine->destroy();
CheckError(utErrorId, *pEngine);
delete tcas;
delete pEngine;
} catch (Exception e) {
cout << "ExampleApplication " << e << endl;
}
/* If we got this far everything went OK */
cout << "ExampleApplication: processing finished sucessfully! " << endl;
return (0);
}
// Look for "EnglishDocument" sofa and read it as a stream
TyErrorId process(CAS & rCas, ResultSpecification const & crResultSpecification) {
cout << "SofaDataAnnotator: process() begins" << endl;
/** get the CAS view of the sofa */
CAS * tcas = rCas.getView("EnglishDocument");
/** get the handle to the index repository */
FSIndexRepository & indexRep = tcas->getIndexRepository();
/** get the default text sofa */
SofaFS textSofa = tcas->getSofa();
/** get the handle to the sofa data stream */
SofaDataStream * pStream = textSofa.getSofaDataStream();
/** open the stream */
int rc = pStream->open();
if (rc != 0) {
cout << "open failed " << rc << endl;
return (TyErrorId)UIMA_ERR_USER_ANNOTATOR_COULD_NOT_PROCESS;
}
/** get the total stream size */
size_t streamSize = pStream->getTotalStreamSizeInBytes();
/** read file contents into a buffer */
char * pBuffer = new char[streamSize+1];
memset(pBuffer,'\n' ,streamSize+1);
int elementsize=1;
pStream->read(pBuffer, elementsize, streamSize);
cout << endl;
cout.write(pBuffer, streamSize);
cout << endl;
/** convert to unicode */
UnicodeString ustrInputText(pBuffer, streamSize+1, "utf-8");
/** find tokens and annotate */
UnicodeString delim(" ");
UChar *myLocalSaveState;
UChar * pInputText = (UChar*) ustrInputText.getBuffer();
const UChar * pToken = pInputText;
const UChar * pNextToken = u_strtok_r((UChar*) pInputText, delim.getBuffer(), &myLocalSaveState);
int start = 1;
int tokenlength=0;
int nTokens = 0;
while ( (pNextToken=u_strtok_r(NULL, delim.getBuffer(), &myLocalSaveState)) ) {
tokenlength = pNextToken - pToken;
AnnotationFS annotFS = tcas->createAnnotation(annot, start, start+tokenlength-2);
indexRep.addFS(annotFS);
++nTokens;
start += tokenlength;
pToken = pNextToken;
}
/* last token */
tokenlength = pNextToken - pToken;
AnnotationFS annotFS = tcas->createAnnotation(annot, start, streamSize);
indexRep.addFS(annotFS);
++nTokens;
cout << endl << " Annotated " << nTokens << " tokens." << endl << endl;
/** close the stream */
pStream->close();
delete pStream;
delete[] pBuffer;
cout << "SofaDataAnnotator: process() ends" << endl;
return (TyErrorId)UIMA_ERR_NONE;
}
void testProcessDocu(uima::util::ConsoleUI& rclConsole,
uima::TextAnalysisEngine& rclEngine,
const char* crclCCSID,
const uima::Language& crclLanguage)
/* ----------------------------------------------------------------------- */
{
TyErrorId utErrorId;
string clstrInputFileContent;
size_t uiNumOfInputDocs = 0;
uima::DocBuffer docBuffer;
CAS* cas = rclEngine.newCAS();
failIfNotTrue(cas != NULL);
/* iterate through all doc specs on command line */
for (rclConsole.setToFirst(); rclConsole.isValid(); rclConsole.setToNext()) {
////uima::util::Filename clInputFilename(rclConsole.getAsCString());
//replaced with a hard wired data file
UnicodeString filename("tdoc_001_enus_850.asc");
UnicodeString fn = ResourceManager::resolveFilename(filename, filename);
uima::util::Filename clInputFilename(UnicodeStringRef(fn).asUTF8().c_str());
size_t uiSize;
if (!clInputFilename.isExistent()) {
rclConsole.fatal(1, _TEXT("Input file not found"), clInputFilename.getAsCString());
}
if (crclCCSID == NULL) /**** (!crclCCSID.isValid()) ***/
{
rclConsole.fatal(1, _TEXT("Invalid CCSID specified - cannot load document"),
crclCCSID /**crclCCSID.getName() **/);
}
rclConsole.format(_TEXT("Adding Document"), clInputFilename.getAsCString());
uiSize = ftool_ReadFileToString(clInputFilename, clstrInputFileContent);
docBuffer.addDocPart(clstrInputFileContent.data(), uiSize, crclCCSID);
// For real file based documents we only add a term annotation for the
// whole "document" if the appropriate switch is set
if (gs_bDocIsTerm) {
assert(false);
}
UnicodeString ustrInputFileContent(clstrInputFileContent.data(), uiSize, crclCCSID);
/* since we already added a complete doc, we may not add anything else */
/// failIfNotTrue(rclEngine.addDocPart(ustrInputFileContent) == UIMA_ERR_ENGINE_INVALID_CALLING_SEQUENCE);
/// failIfNotTrue(rclEngine.addDoc(ustrInputFileContent) == UIMA_ERR_ENGINE_INVALID_CALLING_SEQUENCE);
cas->setDocumentText(docBuffer.getDocBuffer(), docBuffer.getLength());
cas->getDocumentAnnotation().setLanguage(crclLanguage);
utErrorId = rclEngine.process(*cas);
uimaToolHandleErrorId(rclConsole, utErrorId, rclEngine.getAnnotatorContext().getLogger().getLastErrorAsCStr(),
_TEXT("uima::Engine::processDocument"), gs_lExpectedProcessDocumentRc);
if (utErrorId == UIMA_ERR_NONE && gs_bDoIterTest) {
failIfNotTrue(false);
// iteratorTest(rclConsole, rclEngine);
}
utErrorId = cas->reset();
uimaToolHandleErrorId(rclConsole, utErrorId, rclEngine.getAnnotatorContext().getLogger().getLastErrorAsCStr(),
_TEXT("uima::Engine::resetDocument"));
++uiNumOfInputDocs;
}
if (uiNumOfInputDocs == 0) {
rclConsole.warning(_TEXT("No input file(s) specified"));
}
delete cas;
}
TyErrorId process(CAS & rCAS, ResultSpecification const & crResultSpecification) {
CAS *engTcas, *germTcas;
UChar *myLocalSaveState;
// Look for english document and "translate" to German
cout << "SofaExampleAnnotator: process() begins" << endl;
// get English view
engTcas = rCAS.getView("EnglishDocument");
DocumentFS adocFS = engTcas->getDocumentAnnotation();
UnicodeStringRef aengText = adocFS.getCoveredText();
cout << " English Input: " << aengText << endl;
// Create the output German text Sofa and open CAS view
germTcas = rCAS.createView("GermanDocument");
// Get pointer to the English text document
DocumentFS docFS = engTcas->getDocumentAnnotation();
UnicodeStringRef engText = docFS.getCoveredText();
// make copy of document for the u_strtok_r function (100 character limit!)
UChar uWork[100];
u_strncpy(uWork, engText.getBuffer(), 99);
// Setup for translated text
int germBegin = 0;
int germEnd = 0;
UChar translation[400];
translation[0]=0;
// get two IR handles for adding annotations to the appropriate view
FSIndexRepository & engIndexRep = engTcas->getIndexRepository();
FSIndexRepository & germIndexRep = germTcas->getIndexRepository();
// Parse the English text
UChar uDelim[2];
UnicodeString delimUS(" ");
u_strncpy(uDelim, delimUS.getBuffer(), 1);
uDelim[1] = 0;
UChar * next = u_strtok_r(uWork, uDelim, &myLocalSaveState);
while (next) {
// Create annotation on source text
AnnotationFS engAnnot =
engTcas->createAnnotation(annot, next-uWork, (next-uWork)+u_strlen(next));
engIndexRep.addFS(engAnnot);
// Translate word-by-word
const UChar * gword = translate(next);
// Accumulate the total translated document
if (germBegin > 0) {
// if not the first word, add space before
u_strncat(translation, uDelim, 1);
germBegin += 1;
}
u_strcat(translation, gword);
// Create annotation on output text
germEnd = germBegin + u_strlen(gword);
AnnotationFS germAnnot = germTcas->createAnnotation(cross, germBegin, germEnd);
germIndexRep.addFS(germAnnot);
// add link to English text
germAnnot.setFSValue(other, engAnnot);
germBegin = germEnd;
next = u_strtok_r(NULL, uDelim, &myLocalSaveState);
}
// set documentText with accumulated transation
germTcas->setDocumentText( translation, u_strlen(translation), true );
cout << " German(!) Output: " << germTcas->getDocumentText() << endl;
cout << "SofaExampleAnnotator: process() ends" << endl;
return (TyErrorId)UIMA_ERR_NONE;
}
void CASDeserializer::deserializeIndexedFSs(vector<SerializedCAS::TyNum> & crIndexFSs,
uima::CAS & rCAS) {
uima::internal::CASImpl & rCASImpl = uima::internal::CASImpl::promoteCAS(rCAS);
uima::lowlevel::FSHeap & crHeap = rCASImpl.getHeap();
uima::lowlevel::IndexRepository * crIndexRep = &rCASImpl.getIndexRepository();
uima::lowlevel::FSHeap::TyFSHeap const & rTempFSHeap = crHeap.iv_clTemporaryHeap;
SerializedCAS::TyNum iMaxOffset = rTempFSHeap.getTopOfHeap();
vector<SerializedCAS::TyNum>::const_iterator cit, loopit;
vector<SerializedCAS::TyNum> perLoopIndexedFSs;
cit = crIndexFSs.begin();
int numViews = *cit++;
int loopSize = *cit;
crIndexRep->reset();
// deserialize base CAS
if (loopSize > 0) {
lastSegmentUsed = 0;
perLoopIndexedFSs.insert(perLoopIndexedFSs.end(), cit+1, cit+1+loopSize);
cit += loopSize + 1;
for (loopit = perLoopIndexedFSs.begin(); loopit != perLoopIndexedFSs.end(); ++loopit) {
assert( *loopit < iMaxOffset );
crIndexRep->add( *loopit );
}
}
// book keeping for all Sofas
rCAS.getBaseCas()->iv_sofaCount = 1; // reserve for initial view
FSIndex fsIdx = crIndexRep->getIndex(CAS::INDEXID_SOFA);
FSIterator fsIt = fsIdx.iterator();
while (fsIt.isValid()) {
SofaFS aSofa = (SofaFS) fsIt.get();
if ( 0 == aSofa.getSofaID().compare(UnicodeString(CAS::NAME_DEFAULT_SOFA)) ) {
rCAS.registerInitialSofa();
} else {
// only bump sofa count if not initial View
rCAS.bumpSofaCount();
}
rCAS.getView(aSofa)->registerView(aSofa);
fsIt.moveToNext();
}
for (int view = 1; view <= numViews; view++) {
// Check if sofa's index has anything in it
loopSize = *cit;
if (0 == loopSize) {
cit++;
continue;
}
CAS* tcas = rCAS.getViewBySofaNum(view);
uima::internal::CASImpl & crTCASImpl = uima::internal::CASImpl::promoteCAS(*tcas);
crIndexRep = &crTCASImpl.getIndexRepository();
crIndexRep->reset();
perLoopIndexedFSs.clear();
perLoopIndexedFSs.insert(perLoopIndexedFSs.end(), cit+1, cit+1+loopSize);
cit += loopSize + 1;
for (loopit = perLoopIndexedFSs.begin(); loopit != perLoopIndexedFSs.end(); ++loopit) {
assert( *loopit < iMaxOffset );
crIndexRep->add( *loopit );
}
tcas->pickupDocumentAnnotation();
}
}
