int main( int argc, const char* argv[])
{
int rt = 0;
std::auto_ptr<strus::ErrorBufferInterface> errorBuffer( strus::createErrorBuffer_standard( 0, 2));
if (!errorBuffer.get())
{
std::cerr << _TXT("failed to create error buffer") << std::endl;
return -1;
}
strus::ProgramOptions opt;
bool printUsageAndExit = false;
try
{
opt = strus::ProgramOptions(
argc, argv, 11,
"h,help", "v,version", "license", "t,tokenizer:", "n,normalizer:",
"m,module:", "M,moduledir:", "q,quot:", "p,plain",
"R,resourcedir:", "T,trace:");
if (opt( "help")) printUsageAndExit = true;
std::auto_ptr<strus::ModuleLoaderInterface> moduleLoader( strus::createModuleLoader( errorBuffer.get()));
if (!moduleLoader.get()) throw strus::runtime_error(_TXT("failed to create module loader"));
if (opt("moduledir"))
{
std::vector<std::string> modirlist( opt.list("moduledir"));
std::vector<std::string>::const_iterator mi = modirlist.begin(), me = modirlist.end();
for (; mi != me; ++mi)
{
moduleLoader->addModulePath( *mi);
}
moduleLoader->addSystemModulePath();
}
if (opt("module"))
{
std::vector<std::string> modlist( opt.list("module"));
std::vector<std::string>::const_iterator mi = modlist.begin(), me = modlist.end();
for (; mi != me; ++mi)
{
if (!moduleLoader->loadModule( *mi))
{
throw strus::runtime_error(_TXT("error failed to load module %s"), mi->c_str());
}
}
}
if (opt("license"))
{
std::vector<std::string> licenses_3rdParty = moduleLoader->get3rdPartyLicenseTexts();
std::vector<std::string>::const_iterator ti = licenses_3rdParty.begin(), te = licenses_3rdParty.end();
if (ti != te) std::cout << _TXT("3rd party licenses:") << std::endl;
for (; ti != te; ++ti)
{
std::cout << *ti << std::endl;
}
std::cerr << std::endl;
if (!printUsageAndExit) return 0;
}
if (opt( "version"))
{
std::cout << _TXT("Strus utilities version ") << STRUS_UTILITIES_VERSION_STRING << std::endl;
std::cout << _TXT("Strus module version ") << STRUS_MODULE_VERSION_STRING << std::endl;
std::cout << _TXT("Strus rpc version ") << STRUS_RPC_VERSION_STRING << std::endl;
std::cout << _TXT("Strus trace version ") << STRUS_TRACE_VERSION_STRING << std::endl;
std::cout << _TXT("Strus analyzer version ") << STRUS_ANALYZER_VERSION_STRING << std::endl;
std::cout << _TXT("Strus base version ") << STRUS_BASE_VERSION_STRING << std::endl;
std::vector<std::string> versions_3rdParty = moduleLoader->get3rdPartyVersionTexts();
std::vector<std::string>::const_iterator vi = versions_3rdParty.begin(), ve = versions_3rdParty.end();
if (vi != ve) std::cout << _TXT("3rd party versions:") << std::endl;
for (; vi != ve; ++vi)
{
std::cout << *vi << std::endl;
}
if (!printUsageAndExit) return 0;
}
else if (!printUsageAndExit)
{
if (opt.nofargs() > 1)
{
std::cerr << _TXT("too many arguments") << std::endl;
printUsageAndExit = true;
rt = 1;
}
if (opt.nofargs() < 1)
{
std::cerr << _TXT("too few arguments") << std::endl;
printUsageAndExit = true;
rt = 2;
}
}
if (printUsageAndExit)
{
std::cout << _TXT("usage:") << " strusAnalyze [options] <phrasepath>" << std::endl;
std::cout << "<phrasepath> = " << _TXT("path to phrase to analyze ('-' for stdin)") << std::endl;
std::cout << "description: " << _TXT("tokenizes and normalizes a text segment") << std::endl;
std::cout << " " << _TXT("and prints the result to stdout.") << std::endl;
std::cout << _TXT("options:") << std::endl;
std::cout << "-h|--help" << std::endl;
std::cout << " " << _TXT("Print this usage and do nothing else") << std::endl;
std::cout << "-v|--version" << std::endl;
std::cout << " " << _TXT("Print the program version and do nothing else") << std::endl;
std::cout << "--license" << std::endl;
std::cout << " " << _TXT("Print 3rd party licences requiring reference") << std::endl;
std::cout << "-m|--module <MOD>" << std::endl;
std::cout << " " << _TXT("Load components from module <MOD>") << std::endl;
std::cout << "-M|--moduledir <DIR>" << std::endl;
std::cout << " " << _TXT("Search modules to load first in <DIR>") << std::endl;
std::cout << "-R|--resourcedir <DIR>" << std::endl;
std::cout << " " << _TXT("Search resource files for analyzer first in <DIR>") << std::endl;
std::cout << "-t|--tokenizer <CALL>" << std::endl;
std::cout << " " << _TXT("Use the tokenizer <CALL> (default 'content')") << std::endl;
std::cout << "-n|--normalizer <CALL>" << std::endl;
std::cout << " " << _TXT("Use the normalizer <CALL> (default 'orig')") << std::endl;
std::cout << "-q|--quot <STR>" << std::endl;
std::cout << " " << _TXT("Use the string <STR> as quote for the result (default \"\'\")") << std::endl;
std::cout << "-p|--plain" << std::endl;
std::cout << " " << _TXT("Do not print position and define default quotes as empty") << std::endl;
std::cout << "-T|--trace <CONFIG>" << std::endl;
std::cout << " " << _TXT("Print method call traces configured with <CONFIG>") << std::endl;
std::cout << " " << strus::string_format( _TXT("Example: %s"), "-T \"log=dump;file=stdout\"") << std::endl;
return rt;
}
// Declare trace proxy objects:
typedef strus::Reference<strus::TraceProxy> TraceReference;
std::vector<TraceReference> trace;
if (opt("trace"))
{
std::vector<std::string> tracecfglist( opt.list("trace"));
std::vector<std::string>::const_iterator ti = tracecfglist.begin(), te = tracecfglist.end();
for (; ti != te; ++ti)
{
trace.push_back( new strus::TraceProxy( moduleLoader.get(), *ti, errorBuffer.get()));
}
}
std::string resultQuot = "'";
bool resultPlain = false;
if (opt( "plain"))
{
resultPlain = true;
resultQuot.clear();
}
if (opt( "quot"))
{
resultQuot = opt[ "quot"];
}
std::string docpath = opt[0];
std::string tokenizer( "content");
if (opt( "tokenizer"))
{
tokenizer = opt[ "tokenizer"];
}
std::string normalizer( "orig");
if (opt( "normalizer"))
{
normalizer = opt[ "normalizer"];
}
// Set paths for locating resources:
if (opt("resourcedir"))
{
std::vector<std::string> pathlist( opt.list("resourcedir"));
std::vector<std::string>::const_iterator
pi = pathlist.begin(), pe = pathlist.end();
for (; pi != pe; ++pi)
{
moduleLoader->addResourcePath( *pi);
}
}
// Create root object for analyzer:
std::auto_ptr<strus::AnalyzerObjectBuilderInterface>
analyzerBuilder( moduleLoader->createAnalyzerObjectBuilder());
if (!analyzerBuilder.get()) throw strus::runtime_error(_TXT("failed to create analyzer object builder"));
// Create proxy objects if tracing enabled:
{
std::vector<TraceReference>::const_iterator ti = trace.begin(), te = trace.end();
for (; ti != te; ++ti)
{
strus::AnalyzerObjectBuilderInterface* proxy = (*ti)->createProxy( analyzerBuilder.get());
analyzerBuilder.release();
analyzerBuilder.reset( proxy);
}
}
// Create objects for analyzer:
std::auto_ptr<strus::QueryAnalyzerInterface>
analyzer( analyzerBuilder->createQueryAnalyzer());
if (!analyzer.get()) throw strus::runtime_error(_TXT("failed to create analyzer"));
const strus::TextProcessorInterface* textproc = analyzerBuilder->getTextProcessor();
if (!textproc) throw strus::runtime_error(_TXT("failed to get text processor"));
// Create phrase type (tokenizer and normalizer):
std::string phraseType;
if (!strus::loadQueryAnalyzerPhraseType(
*analyzer, textproc, phraseType, "", normalizer, tokenizer, errorBuffer.get()))
{
throw strus::runtime_error(_TXT("failed to load analyze phrase type"));
}
// Load the phrase:
std::string phrase;
if (docpath == "-")
{
unsigned int ec = strus::readStdin( phrase);
if (ec) throw strus::runtime_error( _TXT( "error reading input from stdin (errno %u)"), ec);
}
else
{
unsigned int ec = strus::readFile( docpath, phrase);
if (ec) throw strus::runtime_error( _TXT( "error reading input file '%s' (errno %u)"), docpath.c_str(), ec);
}
// Analyze the phrase and print the result:
std::vector<strus::analyzer::Term> terms
= analyzer->analyzePhrase( phraseType, phrase);
std::sort( terms.begin(), terms.end(), TermOrder());
std::vector<strus::analyzer::Term>::const_iterator
ti = terms.begin(), te = terms.end();
for (; ti != te; ++ti)
{
if (!resultPlain)
{
std::cout << ti->pos() << " ";
}
std::cout << resultQuot << ti->value() << resultQuot << std::endl;
}
if (errorBuffer->hasError())
{
throw strus::runtime_error(_TXT("error in analyze phrase"));
}
return 0;
}
catch (const std::bad_alloc&)
{
std::cerr << _TXT("ERROR ") << _TXT("out of memory") << std::endl;
}
catch (const std::runtime_error& e)
{
const char* errormsg = errorBuffer->fetchError();
if (errormsg)
{
std::cerr << _TXT("ERROR ") << e.what() << ": " << errormsg << std::endl;
}
else
{
std::cerr << _TXT("ERROR ") << e.what() << std::endl;
}
}
catch (const std::exception& e)
{
std::cerr << _TXT("EXCEPTION ") << e.what() << std::endl;
}
return -1;
}
// This doesn't rewrite changes through properly so needs to have a substitution
// applied to its output.
ASTNode PropagateEqualities::propagate(const ASTNode& a, ArrayTransformer* at)
{
ASTNode output;
// if the variable has been solved for, then simply return it
if (simp->InsideSubstitutionMap(a, output))
return output;
if (!alreadyVisited.insert(a.GetNodeNum()).second)
{
return a;
}
output = a;
// traverse a and populate the SubstitutionMap
const Kind k = a.GetKind();
if (SYMBOL == k && BOOLEAN_TYPE == a.GetType())
{
bool updated = simp->UpdateSubstitutionMap(a, ASTTrue);
output = updated ? ASTTrue : a;
}
else if (NOT == k)
{
bool updated = searchXOR(a[0], ASTFalse);
output = updated ? ASTTrue : a;
}
else if (IFF == k || EQ == k)
{
const ASTVec& c = a.GetChildren();
if (c[0] == c[1])
return ASTTrue;
bool updated = simp->UpdateSubstitutionMap(c[0], c[1]);
if (updated)
{
// fill the arrayname readindices vector if e0 is a
// READ(Arr,index) and index is a BVCONST
int to;
if ((to = TermOrder(c[0], c[1])) == 1 && c[0].GetKind() == READ)
at->FillUp_ArrReadIndex_Vec(c[0], c[1]);
else if (to == -1 && c[1].GetKind() == READ)
at->FillUp_ArrReadIndex_Vec(c[1], c[0]);
}
if (!updated)
updated = searchTerm(c[0], c[1]);
if (!updated)
updated = searchTerm(c[1], c[0]);
output = updated ? ASTTrue : a;
}
else if (XOR == k)
{
bool updated = searchXOR(a, ASTTrue);
output = updated ? ASTTrue : a;
if (updated)
return output;
// The below block should be subsumed by the searchXOR function which
// generalises it.
// So the below block should never do anything..
#ifndef NDEBUG
if (a.Degree() != 2)
return output;
int to = TermOrder(a[0], a[1]);
if (0 == to)
{
if (a[0].GetKind() == NOT && a[0][0].GetKind() == EQ &&
a[0][0][0].GetValueWidth() == 1 && a[0][0][1].GetKind() == SYMBOL)
{
// (XOR (NOT(= (1 v))) ... )
const ASTNode& symbol = a[0][0][1];
const ASTNode newN = nf->CreateTerm(
ITE, 1, a[1], a[0][0][0], nf->CreateTerm(BVNEG, 1, a[0][0][0]));
if (simp->UpdateSolverMap(symbol, newN))
{
assert(false);
output = ASTTrue;
}
}
else if (a[1].GetKind() == NOT && a[1][0].GetKind() == EQ &&
a[1][0][0].GetValueWidth() == 1 &&
a[1][0][1].GetKind() == SYMBOL)
{
const ASTNode& symbol = a[1][0][1];
const ASTNode newN = nf->CreateTerm(
ITE, 1, a[0], a[1][0][0], nf->CreateTerm(BVNEG, 1, a[1][0][0]));
if (simp->UpdateSolverMap(symbol, newN))
{
assert(false);
output = ASTTrue;
}
}
else if (a[0].GetKind() == EQ && a[0][0].GetValueWidth() == 1 &&
a[0][1].GetKind() == SYMBOL)
{
// XOR ((= 1 v) ... )
const ASTNode& symbol = a[0][1];
const ASTNode newN = nf->CreateTerm(
ITE, 1, a[1], nf->CreateTerm(BVNEG, 1, a[0][0]), a[0][0]);
if (simp->UpdateSolverMap(symbol, newN))
{
assert(false);
output = ASTTrue;
}
}
else if (a[1].GetKind() == EQ && a[1][0].GetValueWidth() == 1 &&
a[1][1].GetKind() == SYMBOL)
{
const ASTNode& symbol = a[1][1];
const ASTNode newN = nf->CreateTerm(
ITE, 1, a[0], nf->CreateTerm(BVNEG, 1, a[1][0]), a[1][0]);
if (simp->UpdateSolverMap(symbol, newN))
{
assert(false);
output = ASTTrue;
}
}
else
return output;
}
else
{
ASTNode symbol, rhs;
if (to == 1)
{
symbol = a[0];
rhs = a[1];
}
else
{
symbol = a[1];
rhs = a[0];
}
assert(symbol.GetKind() == SYMBOL);
if (simp->UpdateSolverMap(symbol, nf->CreateNode(NOT, rhs)))
{
assert(false);
output = ASTTrue;
}
}
#endif
}
else if (AND == k)
{
const ASTVec& c = a.GetChildren();
ASTVec o;
o.reserve(c.size());
for (ASTVec::const_iterator it = c.begin(), itend = c.end(); it != itend;
it++)
{
if (always_true)
simp->UpdateAlwaysTrueFormSet(*it);
ASTNode aaa = propagate(*it, at);
if (ASTTrue != aaa)
{
if (ASTFalse == aaa)
return ASTFalse;
else
o.push_back(aaa);
}
}
if (o.size() == 0)
output = ASTTrue;
else if (o.size() == 1)
output = o[0];
else if (o != c)
output = nf->CreateNode(AND, o);
else
output = a;
}
return output;
}