Merge branch 'deep'

Author: kosloot

src/checkmbma.cxx

@@ -79,12 +79,13 @@ void check_word( const UnicodeString& _word, bool doMor ){
   if ( uword != ls ){
     return;
   }
-  myMbma.Classify( ls );
-  vector<UnicodeString> anas = myMbma.getResult();
+  myMbma.Classify( ls, "" );
+  vector<pair<UnicodeString,string>> anas = myMbma.getResults(true);
   set<UnicodeString> fails;
   for ( const auto& ana : anas ){
+    UnicodeString flat = flatten(ana.first);
     bool lem_found = false;
-    vector<UnicodeString> mors = TiCC::split_at_first_of( ana, "[]" );
+    vector<UnicodeString> mors = TiCC::split_at_first_of( flat, "[]" );
     bool first = true;
     for ( const auto& mor : mors ){
       UnicodeString mor1 = mor;

src/froggen.cxx

@@ -368,12 +368,12 @@ void create_mblem_trainfile( const multimap<UnicodeString, map<UnicodeString, ma
     exit( EXIT_FAILURE );
   }
   UnicodeString outLine;
-  for ( const auto& it : data ){
-    UnicodeString wordform = it.first;
+  for ( const auto& data_it : data ){
+    UnicodeString wordform = data_it.first;
     UnicodeString safeInstance;
     if ( !outLine.isEmpty() ){
       string out = UnicodeToUTF8(outLine);
-      out.erase( out.length()-1 );
+      out.erase( out.length()-1 ); // remove the final '|'
       os << out << endl;
       outLine.remove();
     }
@@ -410,7 +410,7 @@ void create_mblem_trainfile( const multimap<UnicodeString, map<UnicodeString, ma
       outLine = instance;
     }
     multimap<size_t, multimap<UnicodeString,UnicodeString>,std::greater<size_t>> sorted;
-    for ( const auto& it2 : it.second ){
+    for ( const auto& it2 : data_it.second ){
       for ( const auto& it3: it2.second ){
 	multimap<UnicodeString,UnicodeString> mm;
 	mm.insert(make_pair(it3.first,it2.first));
@@ -441,11 +441,10 @@ void create_mblem_trainfile( const multimap<UnicodeString, map<UnicodeString, ma
 	  thisform = wordform;
 	  if ( tag.indexOf(it.first.c_str()) >= 0 ){
 	    // the POS tag matches, so potentially yes
-	    int part_pos = -1;
 	    UnicodeString part;
 	    for ( const auto& p : it.second ){
 	      // loop over potential particles.
-	      part_pos = thisform.indexOf(p.c_str());
+	      int part_pos = thisform.indexOf(p.c_str());
 	      if ( part_pos != -1 ){
 		part = p.c_str();
 		if ( debug ){

src/makembma.cxx

@@ -164,7 +164,7 @@ int main(int argc, char * const argv[] ) {
       exit(1);
     }
     parts.erase(parts.begin());
-    vector<Rule *> r = myMbma.execute( word, parts );
+    vector<Rule *> r = myMbma.execute( word, "", parts );
     if ( r.empty() ){
       cerr << "problems with entry: '" << line << "'" << endl;
       continue;

src/morgen.cxx

@@ -146,7 +146,7 @@ void create_instance_file( const string& inpname, const string& outname ){
       exit(1);
     }
     parts.erase(parts.begin());
-    vector<Rule *> r = myMbma.execute( word, parts );
+    vector<Rule *> r = myMbma.execute( word, "", parts );
     if ( r.empty() ){
       cerr << "problems with entry: '" << line << "'" << endl;
       continue;
@@ -240,13 +240,13 @@ int main(int argc, char * const argv[] ) {
   TiCC::Configuration frog_config = use_config;
   //  frog_config.clearatt( "configDir", "global" );
   string inpname = names[0];
-  string outname = outputdir + base_name + ".data";
+  string data_out_name = outputdir + base_name + ".data";
   string treename = use_config.lookUp( "treeFile", "mbma" );
   if ( treename.empty() ){
     treename = base_name + ".tree";
   }
 
-    string cgndir;
+  string cgndir;
   bool cgn_opt = opts.extract( "cgn", cgndir );
   if ( cgn_opt ){
     use_config.setatt( "cgnDir", cgndir, "mbma" );
@@ -295,11 +295,10 @@ int main(int argc, char * const argv[] ) {
 
   frog_config.setatt( "treeFile", treename, "mbma" );
   string full_treename = outputdir + treename;
-  create_instance_file( inpname, outname );
-  create_instance_base( outname, full_treename );
+  create_instance_file( inpname, data_out_name );
+  create_instance_base( data_out_name, full_treename );
 
   frog_config.clearatt( "baseName", "mbma" );
-  string mbma_set_name = use_config.lookUp( "set", "mbma" );
 
   string cfg_out;
   if ( configfile.empty() ){

src/nerv.cxx

@@ -17,7 +17,6 @@ int main(){
   string ner = parts[1];
   string prev = "_";
   string cur = parts[2];
-  string next;
   while ( getline( cin, line ) ){
     parts = TiCC::split_at( line, "\t" );
     if ( parts.empty() ){
@@ -37,17 +36,15 @@ int main(){
 	ner = parts[1];
 	prev = cur;
 	cur = parts[2];
-	next = "_";
       }
     }
     else {
-      next = parts[2];
-      cout << word << "\t" << "\t" << ner << "\t" << prev << "\t" << cur << "\t" << next << endl;
+      cout << word << "\t" << "\t" << ner << "\t" << prev << "\t"
+	   << cur << "\t" << parts[2] << endl;
       word = parts[0];
       ner = parts[1];
       prev = cur;
       cur = parts[2];
-      next = "ERR";
     }
   }
 }

src/testmbma.cxx

@@ -145,13 +145,13 @@ void Test( istream& in, bool deep ){
     UnicodeString uWord = parts[0];
     uWord.toLower();
     parts.erase(parts.begin());
-    vector<Rule *> rules = myMbma.execute( uWord, parts );
+    vector<Rule *> rules = myMbma.execute( uWord, "", parts );
     if ( rules.empty() ){
       cout << "no rule matched: " << line << endl;
     }
     else {
       for ( auto const& r : rules ){
-	cout << uWord << "==> " << r->morpheme_string( deep )
+	cout << uWord << "==> " << r->pretty_string( deep )
 	     << " " << r->tag << endl;
 	delete r;
       }