diff --git a/maca_export/src/maca_lemmatizer_export.cc b/maca_export/src/maca_lemmatizer_export.cc
index 508ab7cfe30c90ebe3d26e81bf5eaca68f01458a..af4ee502b7fa7448bd08d05829a0045de8663d32 100644
--- a/maca_export/src/maca_lemmatizer_export.cc
+++ b/maca_export/src/maca_lemmatizer_export.cc
@@ -27,7 +27,6 @@ MacaonTransLemmatizer::MacaonTransLemmatizer(char *lg, char *mcd) {
ctx = context_read_options(5, argv);
form_pos_ht = hash_new(1000000);
-
//maca_lemmatizer_check_options(ctx);
maca_lemmatizer_set_linguistic_resources_filenames(ctx);
@@ -56,9 +55,9 @@ const char *MacaonTransLemmatizer::lemmatizemcf(const char *mcfString) {
FILE *f = fmemopen ((void *)mcfString, strlen(mcfString), "r");
config *c = config_new(f, ctx->mcd_struct, 5);
- char lemma[200];
- char form[200];
- char pos[200];
+ //char lemma[200];
+ //char form[200];
+ //char pos[200];
size_t size;
if (resultstring != NULL) {
@@ -86,7 +85,8 @@ const char *MacaonTransLemmatizer::lemmatizemcf(const char *mcfString) {
fclose(outstream);
return resultstring;
}
-
+/** taken from maca_trans_lemmatizer and added FILE * outstream to write the result to
+ */
void MacaonTransLemmatizer::print_word(word *w, mcd *mcd_struct, char *lemma, FILE *outstream) {
char *buffer = NULL;
char *token = NULL;
@@ -116,7 +116,7 @@ void MacaonTransLemmatizer::print_word(word *w, mcd *mcd_struct, char *lemma, FI
}
-// taken as is from maca_lemmatizer.c
+/** taken as is from maca_lemmatizer.c since it is not included in libtransparse.a */
void MacaonTransLemmatizer::maca_lemmatizer_set_linguistic_resources_filenames(context *ctx) {
char absolute_filename[500];
@@ -131,6 +131,7 @@ void MacaonTransLemmatizer::maca_lemmatizer_set_linguistic_resources_filenames(c
}
}
+/** taken as is from maca_trans_lemmatizer.c since it is not included in libtransparse.a */
char **MacaonTransLemmatizer::read_fplm_file(char *fplm_filename, hash *form_pos_ht, int debug_mode, int *lemma_array_size) {
char form[1000];
char pos[1000];
@@ -182,9 +183,9 @@ char **MacaonTransLemmatizer::read_fplm_file(char *fplm_filename, hash *form_pos
fclose(f);
return lemma_array;
}
-
+/** taken as is from maca_trans_lemmatizer.c since it is not included in libtransparse.a */
char *MacaonTransLemmatizer::lookup_lemma(char *form, char *pos, hash *form_pos_ht, char **lemma_array, int verbose) {
- char form_pos[1000];
+ //char form_pos[1000];
int index_form_pos;
strcpy(form_pos, form);
diff --git a/maca_export/src/maca_lemmatizer_export.h b/maca_export/src/maca_lemmatizer_export.h
index fd676eea6f743d3f617db92371ad29a6de39b1bb..afa63bc1071a69a8e52df08ba1985cb540ef83c7 100644
--- a/maca_export/src/maca_lemmatizer_export.h
+++ b/maca_export/src/maca_lemmatizer_export.h
@@ -4,15 +4,17 @@ class MacaonTransLemmatizer {
public:
/**
creates instance, assumes that the environment variable MACAON_DIR
- is defined
+ is defined. One instance for each thread has to be created (this means loading resources for each thread).
@param lg language to be used (in the sense of sub-dir in MACAON_DIR)
@param mcd the filename of the mcd definitions
*/
MacaonTransLemmatizer(char *lg, char *mcd);
+
~MacaonTransLemmatizer();
+
/** call lemmatizer
@param mcfString a string containing the sentence to be analysed in mcf format
- (at least the columns form, pos, lemma must be present
+ (at least the columns form, pos must be present
@return the parser output
*/
const char *lemmatizemcf(const char *mcfString);
@@ -25,6 +27,13 @@ class MacaonTransLemmatizer {
/// keeps last result (or NULL)
char *resultstring;
+ /// variables used during lemmatization
+ char form_pos[1000];
+ char lemma[200];
+ char form[200];
+ char pos[200];
+
+ /// variables to stock data
hash *form_pos_ht = NULL;
char **lemma_array = NULL;
int lemma_array_size;
diff --git a/maca_export/src/maca_trans_tagger_export.cc b/maca_export/src/maca_trans_tagger_export.cc
index 33fe8e81352a08cd744b93b892999724dd8bfb0b..a791372ddf7d6977ed6b34418e0a7fa84514ac5b 100644
--- a/maca_export/src/maca_trans_tagger_export.cc
+++ b/maca_export/src/maca_trans_tagger_export.cc
@@ -7,11 +7,9 @@ extern "C"{
#endif
#include "context.h"
-#include "feat_fct.h"
#include "config2feat_vec.h"
-#include "feature_table.h"
-#include "dico.h"
#include "movement_tagger.h"
+#include "simple_decoder_tagger.h"
#ifdef __cplusplus
}
#endif
@@ -57,7 +55,7 @@ const char *MacaonTransTagger::tagmcf(const char *mcfString) {
}
-// taken as is
+/** taken as is from maca_trans_tagger.c */
void MacaonTransTagger::decode_tagger_set_linguistic_resources_filenames(context *ctx) {
char absolute_filename[500];
@@ -101,7 +99,7 @@ void MacaonTransTagger::decode_tagger_set_linguistic_resources_filenames(context
}
}
-// taken from simple_decoder_tagger.c and modified to read from string
+/** taken from simple_decoder_tagger.c and modified to read from string and write to string */
void MacaonTransTagger::simple_decoder_tagger(context *ctx, const char *mcfString) {
config *c;
feat_vec *fv = feat_vec_new(feature_types_nb);
@@ -167,6 +165,8 @@ void MacaonTransTagger::simple_decoder_tagger(context *ctx, const char *mcfStrin
fclose(f);
}
+/** taken from simple_decoder_tagger.c and modified (paramater FILE *outstreal)
+ */
void MacaonTransTagger::print_word(word *w, mcd *mcd_struct, dico *dico_pos, int postag, FILE *outstream) {
char *buffer = NULL;
char *token = NULL;
@@ -193,20 +193,3 @@ void MacaonTransTagger::print_word(word *w, mcd *mcd_struct, dico *dico_pos, int
free(buffer);
}
}
-
-void MacaonTransTagger::add_signature_to_words_in_word_buffer(word_buffer *bf, form2pos *f2p) {
- int i;
- word *w;
- char lower_form[100];
-
- for(i = word_buffer_get_nbelem(bf) - 1; i >=0 ; i--){
- w = word_buffer_get_word_n(bf, i);
- if(word_get_signature(w) != -1) break;
- w->signature = form2pos_get_signature(f2p, w->form);
- if(w->signature == -1){
- strcpy(lower_form, w->form);
- to_lower_string(lower_form);
- w->signature = form2pos_get_signature(f2p, lower_form);
- }
- }
-}
diff --git a/maca_export/src/maca_trans_tagger_export.h b/maca_export/src/maca_trans_tagger_export.h
index 62663dd69d066d28f0b262097aeec8d87b443c23..1d3aefe500aa178d27b5e7c6c543de92dde1714e 100644
--- a/maca_export/src/maca_trans_tagger_export.h
+++ b/maca_export/src/maca_trans_tagger_export.h
@@ -4,15 +4,17 @@ class MacaonTransTagger {
public:
/**
creates instance, assumes that the environment variable MACAON_DIR
- is defined
+ is defined. One instance for each thread has to be created (this means loading resources for each thread).
@param lg language to be used (in the sense of sub-dir in MACAON_DIR)
@param mcd the filename of the mcd definitions
*/
MacaonTransTagger(char *lg, char *mcd);
+
~MacaonTransTagger();
+
/** call tagger
- @param mcfString a string containing the sentence to be analysed in mcf format
- (at least the columns form, pos, lemma must be present
+ @param mcfString a string containing the sentence to be analysed in mcf format.
+ At least the column form must be present
@return the parser output
*/
const char *tagmcf(const char *mcfString);
@@ -24,13 +26,12 @@ class MacaonTransTagger {
context *ctx;
/// keeps last result (or NULL)
char *resultstring;
- void decode_tagger_set_linguistic_resources_filenames(context *ctx);
+ // import functions which are not available in libtransparse.a or are modified
+ void decode_tagger_set_linguistic_resources_filenames(context *ctx);
void simple_decoder_tagger(context *ctx, const char *mcf);
void print_word(word *w, mcd *mcd_struct, dico *dico_pos, int postag, FILE *stream);
- void add_signature_to_words_in_word_buffer(word_buffer *bf, form2pos *f2p);
};
-//extern void add_signature_to_words_in_word_buffer(word_buffer *bf, form2pos *f2p);
diff --git a/maca_trans_parser/src/simple_decoder_tagger.h b/maca_trans_parser/src/simple_decoder_tagger.h
index 28f4f1091d01c97638b0a891c76176c68042992a..b5d5b2aa019629a560011a9b7f575228a7cb3353 100644
--- a/maca_trans_parser/src/simple_decoder_tagger.h
+++ b/maca_trans_parser/src/simple_decoder_tagger.h
@@ -1,6 +1,6 @@
#ifndef __SIMPLE_DECODER_TAGGER__
#define __SIMPLE_DECODER_TAGGER__
-
+void add_signature_to_words_in_word_buffer(word_buffer *bf, form2pos *f2p);
void simple_decoder_tagger(context *ctx);
#endif