adding empty lefff

git-svn-id: https://scm.gforge.inria.fr/authscm/cfourrie/svn/lingwb/metagger/trunk@2700 dc05b511-7f1d-0410-9f1c-d6f32a2df9e4

bin/Makefile.am

@@ -2,26 +2,29 @@
 
 AUTOMAKE_OPTIONS = foreign
 
-bin_SCRIPTS =	build_word_list.py	\
-		build_tag_dict.py	\
-		learning_curves.py	\
-		pos_tag.py		\
-		lemmatizer.pl		\
-		eval.py
+bin_SCRIPTS =	build_word_list.py		\
+		build_tag_dict.py		\
+		learning_curves.py		\
+		pos_tag.py			\
+		lemmatizer.pl			\
+		eval.py				\
+		run_learning_curves_exp.py	
 
-CLEANFILES =	build_word_list.py	\
-		build_tag_dict.py	\
-		learning_curves.py	\
-		pos_tag.py		\
-		lemmatizer.pl		\
-		eval.py
+CLEANFILES =	build_word_list.py		\
+		build_tag_dict.py		\
+		learning_curves.py		\
+		pos_tag.py			\
+		lemmatizer.pl			\
+		eval.py				\
+		run_learning_curves_exp.py
 
-BUILT_SOURCES =	build_word_list.py	\
-		build_tag_dict.py	\
-		learning_curves.py	\
-		pos_tag.py		\
-		lemmatizer.pl		\
-		eval.py
+BUILT_SOURCES =	build_word_list.py		\
+		build_tag_dict.py		\
+		learning_curves.py		\
+		pos_tag.py			\
+		lemmatizer.pl			\
+		eval.py				\
+		run_learning_curves_exp.py
 
 EDIT = perl -pe "s|\@datadir\@|\$(pkgdatadir)|g; 	\
 		 s|\@libdir\@|\$(pkglibdir)|g; 		\

bin/run_learning_curves_exp.py.in

+#!/usr/bin/env python
+
+import sys
+import os
+import re
+import optparse
+from metagger.pos_tagger import POSTagger
+from metagger.result_sink import AccuracySink, compare_files
+from metagger.utils import tag_dict, word_list
+
+#### I/O 
+usage = "usage: %prog [options] <input_file>"
+parser = optparse.OptionParser(usage=usage)
+parser.add_option("-t", "--training", action="store", help="training data directory")
+parser.add_option("-l", "--lefff", action="store", help="lefff directory")
+parser.add_option("-v", "--values", action="store", help="different values for precision priors")
+parser.add_option("-g", "--gold", action="store", help="gold file")
+(options, args) = parser.parse_args()
+
+infile = args[0]
+
+# values = [1, 0.1, 0.01, 0.001, 10, 100]
+values = [1]
+if options.values:
+    values = map(int, options.values.split(','))
+
+#### get training files 
+train_files = [ os.path.join( options.training, f ) for f in os.listdir( options.training ) if not f.startswith('.')]
+train_files.sort( key=lambda x:int(os.path.getsize( x )) )
+
+#### get lefff files
+lefff_files = [ os.path.join( options.lefff, f ) for f in os.listdir( options.lefff ) if not f.startswith('.')]
+lefff_files.sort( key=lambda x:int(os.path.getsize( x )) )
+
+log_filename = "learning_curves_exp.log"
+log_file = open(log_filename,'w')
+
+print >> log_file, "%15s %15s %15s | %15s %15s" %("Train. size", "Lefff size", "Prior prec.", "Acc.", "Unk. Acc.")
+log_file.flush()
+ 
+#### set-up experiments
+exp_ct = 0
+for tf in train_files:
+    ftb_size = re.search(r"ftb4\+mergeC\+undocpd\_1\_(\d+)\.",tf).group(1)
+    tags = tag_dict( tf )
+    wds = word_list( tf )
+    for lf in lefff_files:
+        lefff_size = re.search(r"lefff\_(\d+)\.",lf).group(1)
+        file_infix = "tb%s-lefff%s" %(ftb_size,lefff_size)
+        for v in values:
+            exp_ct += 1
+            print >> sys.stderr, "Exp #%s: train size: %s, lefff size: %s, lambda: %s" %(exp_ct, ftb_size, lefff_size, v)
+            # intialize tagger
+            pos_tagger = POSTagger()
+            # load data
+            pos_tagger.tag_dict = tags 
+            pos_tagger.known_words = wds
+            pos_tagger.load_lefff( lf )
+            # train model
+            modelfile = file_infix+'.megam'
+            pos_tagger.train_model( tf, prior_prec=1, model_path=modelfile )
+            # apply tagger
+            predfile = file_infix+'.pred'
+            pos_tagger.apply( infile, outfile=predfile )
+            # eval
+            sink = AccuracySink()
+            unk_sink = AccuracySink() # unkown words
+            compare_files( options.gold, predfile, sink, unk_sink, pos_tagger.known_words)
+            print >> log_file, "%15s %15s %15s | %15s %15s" %(ftb_size,lefff_size,v,sink.score(),unk_sink.score())
+            log_file.flush()
+            
+log_file.close()
+