diff --git a/.DS_Store b/.DS_Store
new file mode 100644
index 0000000..5f32f73
Binary files /dev/null and b/.DS_Store differ
diff --git a/amr_parsing.py b/amr_parsing.py
index a570738..dafa8e2 100755
--- a/amr_parsing.py
+++ b/amr_parsing.py
@@ -132,13 +132,16 @@ def main():
     arg_parser.add_argument('-s','--start_step',type=int,default=0,help='specify which step to begin oracle testing;for debug')
     #arg_parser.add_argument('-i','--input_file',help='the input: preprocessed data instances file for aligner or training')
     arg_parser.add_argument('-d','--dev',help='development file')
+    arg_parser.add_argument('-a','--add',help='additional training file')
     arg_parser.add_argument('-as','--actionset',choices=['basic'],default='basic',help='choose different action set')
     arg_parser.add_argument('-m','--mode',choices=['preprocess','test_gold_graph','align','userGuide','oracleGuide','train','parse','eval'],help="preprocess:generate pos tag, dependency tree, ner\n" "align:do alignment between AMR graph and sentence string")
     arg_parser.add_argument('-dp','--depparser',choices=['stanford','stanfordConvert','stdconv+charniak','clear','mate','turbo'],default='stdconv+charniak',help='choose the dependency parser')
     arg_parser.add_argument('--coref',action='store_true',help='flag to enable coreference information')
     arg_parser.add_argument('--prop',action='store_true',help='flag to enable semantic role labeling information')
     arg_parser.add_argument('--rne',action='store_true',help='flag to enable rich name entity')
-    arg_parser.add_argument('--onto',action='store_true',help='flag to enable charniak parse result trained on ontonotes')
+    arg_parser.add_argument('--verblist',action='store_true',help='flag to enable verbalization list')
+    #arg_parser.add_argument('--onto',action='store_true',help='flag to enable charniak parse result trained on ontonotes')
+    arg_parser.add_argument('--onto',choices=['onto','onto+bolt','wsj'],default='wsj',help='choose which charniak parse result trained on ontonotes')
     arg_parser.add_argument('--model',help='specify the model file')
     arg_parser.add_argument('--feat',help='feature template file')
     arg_parser.add_argument('-iter','--iterations',default=1,type=int,help='training iterations')
@@ -156,6 +159,7 @@ def main():
     constants.FLAG_COREF=args.coref
     constants.FLAG_PROP=args.prop
     constants.FLAG_RNE=args.rne
+    constants.FLAG_VERB=args.verblist
     constants.FLAG_ONTO=args.onto
     constants.FLAG_DEPPARSER=args.depparser
 
@@ -308,9 +312,11 @@ def main():
         print "Incorporate Coref Information: %s"%(constants.FLAG_COREF)
         print "Incorporate SRL Information: %s"%(constants.FLAG_PROP)
         print "Substitue the normal name entity tag with rich name entity tag: %s"%(constants.FLAG_RNE)
+        print "Using verbalization list: %s"%(constants.FLAG_VERB)
         print "Using charniak parser trained on ontonotes: %s"%(constants.FLAG_ONTO)
         print "Dependency parser used: %s"%(constants.FLAG_DEPPARSER)
-        train_instances = preprocess(amr_file,START_SNLP=False)        
+        train_instances = preprocess(amr_file,START_SNLP=False)
+        if args.add: train_instances = train_instances + preprocess(args.add,START_SNLP=False)
         if args.dev: dev_instances = preprocess(args.dev,START_SNLP=False)
 
 
@@ -330,6 +336,11 @@ def main():
         model.setup(action_type=args.actionset,instances=train_instances,parser=parser,feature_templates_file=feat_template)
         
         print >> experiment_log, "BEGIN TRAINING!"
+        best_fscore = 0.0
+        best_pscore = 0.0
+        best_rscore = 0.0
+        best_model = None
+        best_iter = 1
         for iter in xrange(1,args.iterations+1):
             print >> experiment_log, "shuffling training instances"
             random.shuffle(train_instances)
@@ -338,22 +349,36 @@ def main():
             begin_updates = parser.perceptron.get_num_updates()
             parser.parse_corpus_train(train_instances)
             parser.perceptron.average_weight()
-            #model.save_model(args.model+'-iter'+str(iter)+'-'+str(int(time.time()))+'.m')
-            model.save_model(args.model+'-iter'+str(iter)+'.m')
+            
             if args.dev:
                 print >> experiment_log ,"Result on develop set:"                
                 _,parsed_amr = parser.parse_corpus_test(dev_instances)
-                write_parsed_amr(parsed_amr,dev_instances,args.dev,args.section+'.'+str(iter)+'.parsed')
+                parsed_suffix = args.section+'.'+args.model.split('.')[-1]+'.'+str(iter)+'.parsed'
+                write_parsed_amr(parsed_amr,dev_instances,args.dev,parsed_suffix)
                 if args.smatcheval:
                     smatch_path = "./smatch_2.0.2/smatch.py"
                     python_path = 'python'
                     options = '--pr -f'
-                    parsed_filename = args.dev+'.'+args.section+'.'+str(iter)+'.parsed'
+                    parsed_filename = args.dev+'.'+parsed_suffix
                     command = '%s %s %s %s %s' % (python_path, smatch_path, options, parsed_filename, args.dev)
                     
                     print 'Evaluation using command: ' + (command)
-                    print subprocess.check_output(command, stderr=subprocess.STDOUT, shell=True)
-
+                    #print subprocess.check_output(command, stderr=subprocess.STDOUT, shell=True)
+                    eval_output = subprocess.check_output(command, stderr=subprocess.STDOUT, shell=True)
+                    print eval_output
+                    pscore = float(eval_output.split('\n')[0].split(':')[1].rstrip())
+                    rscore = float(eval_output.split('\n')[1].split(':')[1].rstrip())
+                    fscore = float(eval_output.split('\n')[2].split(':')[1].rstrip())
+                    if fscore > best_fscore:
+                        best_model = model
+                        best_iter = iter
+                        best_fscore = fscore
+                        best_pscore = pscore
+                        best_rscore = rscore
+
+        if best_model is not None:
+            print >> experiment_log, "Best result on iteration %d:\n Precision: %f\n Recall: %f\n F-score: %f" % (best_iter, best_pscore, best_rscore, best_fscore)
+            best_model.save_model(args.model+'.m')
         print >> experiment_log ,"DONE TRAINING!"
         
     elif args.mode == 'parse': # actual parsing
@@ -369,7 +394,8 @@ def main():
         parser = Parser(model=model,oracle_type=DET_T2G_ORACLE_ABT,action_type=args.actionset,verbose=args.verbose,elog=experiment_log)
         print >> experiment_log ,"BEGIN PARSING"
         span_graph_pairs,results = parser.parse_corpus_test(test_instances)
-        write_parsed_amr(results,test_instances,amr_file,suffix='%s.parsed'%(args.section))
+        parsed_suffix = '%s.%s.parsed'%(args.section,args.model.split('.')[-2])
+        write_parsed_amr(results,test_instances,amr_file,suffix=parsed_suffix)
         #write_span_graph(span_graph_pairs,test_instances,amr_file,suffix='spg.50')
         ################
         # for eval     #
@@ -381,7 +407,7 @@ def main():
             smatch_path = "./smatch_2.0.2/smatch.py"
             python_path = 'python'
             options = '--pr -f'
-            parsed_filename = amr_file+'.'+args.section+'.parsed'
+            parsed_filename = amr_file+'.'+parsed_suffix
             command = '%s %s %s %s %s' % (python_path,smatch_path,options,parsed_filename, amr_file)
                     
             print 'Evaluation using command: ' + (command)
diff --git a/constants.py b/constants.py
index d27e896..904a3fb 100644
--- a/constants.py
+++ b/constants.py
@@ -10,8 +10,9 @@
 FLAG_COREF=False
 FLAG_PROP=False
 FLAG_RNE=False
+FLAG_VERB=False
 FLAG_DEPPARSER='stanford'
-FLAG_ONTO=False
+FLAG_ONTO='wsj'
 
 # constants
 NOT_APPLY='_NOT_APPLY_'
diff --git a/preprocessing.py b/preprocessing.py
index f6e143c..f01e4ae 100644
--- a/preprocessing.py
+++ b/preprocessing.py
@@ -52,7 +52,7 @@ def readAMREval(eval_file_path):
     '''
     read in semeval evaluation format (without amr)
     '''
-    eval_file = codecs.open(eval_file_path,'r',encoding='macroman')
+    eval_file = codecs.open(eval_file_path,'r',encoding='utf-8')
     comment_list = []
     comment = OrderedDict()
     #amr_list = []
@@ -145,7 +145,11 @@ def _load_cparse(cparse_filename):
 
 def _fix_prop_head(inst,ctree,start_index,height):
     head_index = None
-    tree_pos = ctree.leaf_treeposition(start_index)
+    try:
+        tree_pos = ctree.leaf_treeposition(start_index)
+    except IndexError:
+        import pdb
+        pdb.set_trace()
     span_root = ctree[tree_pos[:-(height+1)]]
     end_index = start_index + len(span_root.leaves())
     cur = inst.tokens[start_index+1]
@@ -330,7 +334,7 @@ def preprocess(input_file,START_SNLP=True,INPUT_AMR='amr'):
             if 'alignments' in comments[i]:
                 alignment,s2c_alignment = Aligner.readJAMRAlignment(amr,comments[i]['alignments'])
                 # use verbalization list to fix the unaligned tokens
-                Aligner.postProcessVerbList(amr, comments[i]['tok'], alignment)
+                if constants.FLAG_VERB: Aligner.postProcessVerbList(amr, comments[i]['tok'], alignment)
                 #ggraph = SpanGraph.init_ref_graph(amr,alignment,instances[i].tokens)
                 ggraph = SpanGraph.init_ref_graph_abt(amr,alignment,s2c_alignment,instances[i].tokens)
                 #ggraph.pre_merge_netag(instances[i])
@@ -419,7 +423,12 @@ def preprocess(input_file,START_SNLP=True,INPUT_AMR='amr'):
         _add_dependency(instances,dep_result,constants.FLAG_DEPPARSER)
 
     elif constants.FLAG_DEPPARSER == "stdconv+charniak":
-        dep_filename = tok_sent_filename+'.charniak.onto.parse.dep' if constants.FLAG_ONTO else tok_sent_filename+'.charniak.parse.dep'
+        if constants.FLAG_ONTO == 'onto':
+            dep_filename = tok_sent_filename+'.charniak.onto.parse.dep'
+        elif constants.FLAG_ONTO == 'onto+bolt':
+            dep_filename = tok_sent_filename+'.charniak.onto+bolt.parse.dep'
+        else:
+            dep_filename = tok_sent_filename+'.charniak.parse.dep'            
         if not os.path.exists(dep_filename):
             dparser = CharniakParser()
             dparser.parse(tok_sent_filename)
@@ -463,7 +472,7 @@ def preprocess(input_file,START_SNLP=True,INPUT_AMR='amr'):
     
     if constants.FLAG_PROP:
         print >> log, "Adding SRL information..."
-        prop_filename = tok_sent_filename + '.prop'
+        prop_filename = tok_sent_filename + '.prop' if constants.FLAG_ONTO != 'onto+bolt' else tok_sent_filename + '.onto+bolt.prop'
         if os.path.exists(prop_filename):
             if constants.FLAG_DEPPARSER == "stdconv+charniak":
                 _add_prop(instances,prop_filename,dep_filename,FIX_PROP_HEAD=True)
diff --git a/scripts/cmd.test.brown-verb-onto-rne.sh b/scripts/cmd.test.brown-verb-onto-rne.sh
new file mode 100644
index 0000000..d9375c8
--- /dev/null
+++ b/scripts/cmd.test.brown-verb-onto-rne.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --onto onto --rne --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-rne.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.test.basic-abt-brown-verb-onto-rne.log 2>&1 &
diff --git a/scripts/cmd.test.brown-verb-onto.sh b/scripts/cmd.test.brown-verb-onto.sh
new file mode 100644
index 0000000..69beb43
--- /dev/null
+++ b/scripts/cmd.test.brown-verb-onto.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --onto onto --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.test.basic-abt-brown-verb-onto.log 2>&1 &
diff --git a/scripts/cmd.test.brown-verb-rne.sh b/scripts/cmd.test.brown-verb-rne.sh
new file mode 100644
index 0000000..e44a4e8
--- /dev/null
+++ b/scripts/cmd.test.brown-verb-rne.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --rne --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-rne.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.test.basic-abt-brown-verb-rne.log 2>&1 &
diff --git a/scripts/cmd.test.brown-verb.sh b/scripts/cmd.test.brown-verb.sh
new file mode 100644
index 0000000..399631e
--- /dev/null
+++ b/scripts/cmd.test.brown-verb.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.test.basic-abt-brown-verb.log 2>&1 &
diff --git a/scripts/cmd.test.sh b/scripts/cmd.test.sh
new file mode 100644
index 0000000..2d4f930
--- /dev/null
+++ b/scripts/cmd.test.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.basic-abt.log 2>&1 &
diff --git a/scripts/cmd.test.verb.sh b/scripts/cmd.test.verb.sh
new file mode 100644
index 0000000..c9ac6e0
--- /dev/null
+++ b/scripts/cmd.test.verb.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Testing ..."
+/usr/bin/python -u amr_parsing.py -m parse --amrfmt amr --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-verb.m ./data/semeval/test.txt > ./log/amr-semeval-all.test.basic-abt-verb.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-onto+bolt-rne-srl.sh b/scripts/cmd.train.brown-verb-onto+bolt-rne-srl.sh
new file mode 100644
index 0000000..29c3be7
--- /dev/null
+++ b/scripts/cmd.train.brown-verb-onto+bolt-rne-srl.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --prop --rne --onto 'onto+bolt' --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto+bolt-rne-srl -iter 5 --feat ./feature/basic_abt_srl_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto+bolt-rne-srl.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-onto-rne-srl.sh b/scripts/cmd.train.brown-verb-onto-rne-srl.sh
index 6616d33..595e7d2 100644
--- a/scripts/cmd.train.brown-verb-onto-rne-srl.sh
+++ b/scripts/cmd.train.brown-verb-onto-rne-srl.sh
@@ -1,4 +1,4 @@
 #!/bin/sh
 
 echo "Training Model ..."
-/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --prop --rne --onto --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-rne-srl -iter 5 --feat ./feature/basic_abt_srl_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/test.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto-rne-srl.log 2>&1 &
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --prop --rne --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-rne-srl -iter 5 --feat ./feature/basic_abt_srl_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto-rne-srl.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-onto-rne.sh b/scripts/cmd.train.brown-verb-onto-rne.sh
index 25fab4c..504f757 100644
--- a/scripts/cmd.train.brown-verb-onto-rne.sh
+++ b/scripts/cmd.train.brown-verb-onto-rne.sh
@@ -1,4 +1,4 @@
 #!/bin/sh
 
 echo "Training Model ..."
-/usr/bin/python amr_parsing.py -m train --amrfmt amr --rne --onto --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-rne -iter 5 --feat ./feature/basic_abt_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto-rne.log 2>&1 &
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --rne --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-rne -iter 5 --feat ./feature/basic_abt_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto-rne.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-onto-srl.sh b/scripts/cmd.train.brown-verb-onto-srl.sh
new file mode 100644
index 0000000..10d7c03
--- /dev/null
+++ b/scripts/cmd.train.brown-verb-onto-srl.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --prop --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto-srl -iter 5 --feat ./feature/basic_abt_srl_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto-srl.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-onto.sh b/scripts/cmd.train.brown-verb-onto.sh
new file mode 100644
index 0000000..b28e9cd
--- /dev/null
+++ b/scripts/cmd.train.brown-verb-onto.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-onto -iter 5 --feat ./feature/basic_abt_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-onto.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb-rne.sh b/scripts/cmd.train.brown-verb-rne.sh
new file mode 100755
index 0000000..6d5e968
--- /dev/null
+++ b/scripts/cmd.train.brown-verb-rne.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --verblist --smatcheval --rne --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb-rne -iter 5 --feat ./feature/basic_abt_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb-rne.log 2>&1 &
diff --git a/scripts/cmd.train.brown-verb.sh b/scripts/cmd.train.brown-verb.sh
new file mode 100755
index 0000000..6967dba
--- /dev/null
+++ b/scripts/cmd.train.brown-verb.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python amr_parsing.py -m train --amrfmt amr --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-brown-verb -iter 5 --feat ./feature/basic_abt_brown_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-brown-verb.log 2>&1 &
diff --git a/scripts/cmd.train.onto-verb.sh b/scripts/cmd.train.onto-verb.sh
new file mode 100644
index 0000000..e5269c8
--- /dev/null
+++ b/scripts/cmd.train.onto-verb.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-onto-verb -iter 5 --feat ./feature/basic_abt_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-onto-verb.log 2>&1 &
diff --git a/scripts/cmd.train.onto.sh b/scripts/cmd.train.onto.sh
new file mode 100644
index 0000000..a706272
--- /dev/null
+++ b/scripts/cmd.train.onto.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --onto onto --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-onto -iter 5 --feat ./feature/basic_abt_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-onto.log 2>&1 &
diff --git a/scripts/cmd.train.sh b/scripts/cmd.train.sh
new file mode 100755
index 0000000..38b94ae
--- /dev/null
+++ b/scripts/cmd.train.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python amr_parsing.py -m train --amrfmt amr --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt -iter 5 --feat ./feature/basic_abt_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt.log 2>&1 &
diff --git a/scripts/cmd.train.verb.sh b/scripts/cmd.train.verb.sh
new file mode 100755
index 0000000..82d90a0
--- /dev/null
+++ b/scripts/cmd.train.verb.sh
@@ -0,0 +1,4 @@
+#!/bin/sh
+
+echo "Training Model ..."
+/usr/bin/python -u amr_parsing.py -m train --amrfmt amr --verblist --smatcheval --model ./models/semeval/amr-semeval-all.train.basic-abt-verb -iter 5 --feat ./feature/basic_abt_feats.templates ./data/semeval/training.txt -d ./data/semeval/dev.txt > ./log/amr-semeval-all.train.basic-abt-verb.log 2>&1 &
diff --git a/scripts/fix_multi_sent.py b/scripts/fix_multi_sent.py
new file mode 100644
index 0000000..6f41ee7
--- /dev/null
+++ b/scripts/fix_multi_sent.py
@@ -0,0 +1,34 @@
+
+'''
+fix charniak parse file with multiple sentences 
+'''
+from nltk.tree import Tree
+import sys
+import re
+import codecs
+
+def fix_multi_sent(line):
+    tree = Tree.fromstring(line)
+    if len(tree) > 1:
+        newtree = Tree('S1',[Tree('S',tree[:])])
+    else:
+        newtree = tree
+    return re.sub('\n\s*',' ',newtree.__str__())
+
+
+if __name__ == '__main__':
+    old_parse_file = sys.argv[1]
+    new_parse_file = old_parse_file.rsplit('.',1)[0]
+    print '%s >> %s' % (old_parse_file, new_parse_file)
+    result = []
+    with codecs.open(old_parse_file,'r',encoding='utf-8') as f:
+        for line in f:
+            line = line.strip()
+            new = fix_multi_sent(line)
+            result.append(new)
+
+    with codecs.open(new_parse_file, 'w', encoding='utf-8') as wf:
+        wf.write('\n'.join(result))
+        wf.write('\n')
+    
+
diff --git a/temp/feats_gen_basic_abt_srl_feats.py b/temp/feats_gen_basic_abt_srl_feats.py
new file mode 100644
index 0000000..6bd055f
--- /dev/null
+++ b/temp/feats_gen_basic_abt_srl_feats.py
@@ -0,0 +1,122 @@
+#generated by model.py
+from constants import *
+def generate_features(state,action):
+    s0,b0,a0=state.get_feature_context_window(action)
+    feats=[]
+    act_idx = state.model.class_codebook.get_index(action['type'])
+    tx = action['tag'] if 'tag' in action else EMPTY
+    s0_ne=s0['ne'] if s0 else EMPTY
+    s0_w=s0['form'] if s0 else EMPTY
+    s0_lemma=s0['lemma'] if s0 else EMPTY
+    s0_t=s0['pos'] if s0 else EMPTY
+    s0_dl=s0['rel'] if s0 else EMPTY
+    s0_len=s0['len'] if s0 else EMPTY
+    s0_txv=s0['txv'] if s0 else EMPTY
+    s0_frmset=s0['frmset'] if s0 else EMPTY
+    s0_eqfrmset=s0['eqfrmset'] if s0 else EMPTY
+    b0_isarg=b0['isarg'] if b0 else EMPTY
+    b0_arglabel=b0['arglabel'] if b0 else EMPTY
+    b0_dl=b0['rel'] if b0 else EMPTY
+    a0_isprd=a0['isprd'] if a0 else EMPTY
+    a0_prdlabel=a0['prdlabel'] if a0 else EMPTY
+    b0_isprd=b0['isprd'] if b0 else EMPTY
+    b0_prdlabel=b0['prdlabel'] if b0 else EMPTY
+    a0_isarg=a0['isarg'] if a0 else EMPTY
+    a0_arglabel=a0['arglabel'] if a0 else EMPTY
+    s0_isnom=s0['isnom'] if s0 else EMPTY
+    s0_nech=s0['nech'] if s0 else EMPTY
+    s0_c1lemma=s0['c1lemma'] if s0 else EMPTY
+    s0_c1dl=s0['c1dl'] if s0 else EMPTY
+    s0_cpt=s0['concept'] if s0 else EMPTY
+    s0_p1_ne=s0['p1']['ne'] if s0 and s0['p1'] else EMPTY
+    b0_ne=b0['ne'] if b0 else EMPTY
+    b0_w=b0['form'] if b0 else EMPTY
+    b0_lemma=b0['lemma'] if b0 else EMPTY
+    b0_t=b0['pos'] if b0 else EMPTY
+    b0_len=b0['len'] if b0 else EMPTY
+    b0_reph=b0['reph'] if b0 else EMPTY
+    a0_ne=a0['ne'] if a0 else EMPTY
+    a0_w=a0['form'] if a0 else EMPTY
+    a0_lemma=a0['lemma'] if a0 else EMPTY
+    a0_t=a0['pos'] if a0 else EMPTY
+    a0_dl=a0['rel'] if a0 else EMPTY
+    s0_p1_w=s0['p1']['form'] if s0 and s0['p1'] else EMPTY
+    s0_p1_lemma=s0['p1']['lemma'] if s0 and s0['p1'] else EMPTY
+    s0_p1_t=s0['p1']['pos'] if s0 and s0['p1'] else EMPTY
+    s0_p1_dl=s0['p1']['rel'] if s0 and s0['p1'] else EMPTY
+    b0_pathpwd=b0['pathpwd'] if b0 else EMPTY
+    b0_apathpwd=b0['apathpwd'] if b0 else EMPTY
+    b0_rsb_dl=b0['rsb']['rel'] if b0 and b0['rsb'] else EMPTY
+    b0_nswp=b0['nswp'] if b0 else EMPTY
+    b0_pathp=b0['pathp'] if b0 else EMPTY
+    b0_apathp=b0['apathp'] if b0 else EMPTY
+    a0_cpt=a0['concept'] if a0 else EMPTY
+    b0_cpt=b0['concept'] if b0 else EMPTY
+    dist1=abs(s0['id']-b0['id']) if b0 and b0 is not ABT_TOKEN and s0 is not ABT_TOKEN else EMPTY
+    if dist1 > 10: dist1=10
+    dist2=abs(a0['id']-b0['id']) if b0 and a0 and b0 is not ABT_TOKEN and a0 is not ABT_TOKEN else EMPTY
+    if dist2 > 10: dist2=10
+    if [s0_ne,tx] != 2*[None]:feats.append('s0_ne&tx=%s_%s_' % (s0_ne,tx))
+    if [s0_w,tx] != 2*[None]:feats.append('s0_w&tx=%s_%s_' % (s0_w,tx))
+    if [s0_lemma,tx] != 2*[None]:feats.append('s0_lemma&tx=%s_%s_' % (s0_lemma,tx))
+    if [s0_t,tx] != 2*[None]:feats.append('s0_t&tx=%s_%s_' % (s0_t,tx))
+    if [s0_dl,tx] != 2*[None]:feats.append('s0_dl&tx=%s_%s_' % (s0_dl,tx))
+    if [s0_len,tx] != 2*[None]:feats.append('s0_len&tx=%s_%s_' % (s0_len,tx))
+    if [s0_t,s0_txv] != 2*[None]:feats.append('s0_t&s0_txv=%s_%s_' % (s0_t,s0_txv))
+    if [s0_frmset] != 1*[None]:feats.append('s0_frmset=%s_' % (s0_frmset))
+    if [s0_eqfrmset] != 1*[None]:feats.append('s0_eqfrmset=%s_' % (s0_eqfrmset))
+    if [b0_isarg] != 1*[None]:feats.append('b0_isarg=%s_' % (b0_isarg))
+    if [b0_arglabel] != 1*[None]:feats.append('b0_arglabel=%s_' % (b0_arglabel))
+    if [b0_dl,b0_arglabel] != 2*[None]:feats.append('b0_dl&b0_arglabel=%s_%s_' % (b0_dl,b0_arglabel))
+    if [a0_isprd] != 1*[None]:feats.append('a0_isprd=%s_' % (a0_isprd))
+    if [a0_prdlabel] != 1*[None]:feats.append('a0_prdlabel=%s_' % (a0_prdlabel))
+    if [b0_isprd] != 1*[None]:feats.append('b0_isprd=%s_' % (b0_isprd))
+    if [b0_prdlabel] != 1*[None]:feats.append('b0_prdlabel=%s_' % (b0_prdlabel))
+    if [a0_isarg] != 1*[None]:feats.append('a0_isarg=%s_' % (a0_isarg))
+    if [a0_arglabel] != 1*[None]:feats.append('a0_arglabel=%s_' % (a0_arglabel))
+    if [s0_isnom] != 1*[None]:feats.append('s0_isnom=%s_' % (s0_isnom))
+    if [s0_nech] != 1*[None]:feats.append('s0_nech=%s_' % (s0_nech))
+    if [s0_lemma,s0_nech] != 2*[None]:feats.append('s0_lemma&s0_nech=%s_%s_' % (s0_lemma,s0_nech))
+    if [s0_isnom,s0_nech] != 2*[None]:feats.append('s0_isnom&s0_nech=%s_%s_' % (s0_isnom,s0_nech))
+    if [s0_c1lemma,tx] != 2*[None]:feats.append('s0_c1lemma&tx=%s_%s_' % (s0_c1lemma,tx))
+    if [s0_c1dl,s0_c1lemma] != 2*[None]:feats.append('s0_c1dl&s0_c1lemma=%s_%s_' % (s0_c1dl,s0_c1lemma))
+    if [s0_cpt,s0_p1_ne,s0_c1lemma] != 3*[None]:feats.append('s0_cpt&s0_p1_ne&s0_c1lemma=%s_%s_%s_' % (s0_cpt,s0_p1_ne,s0_c1lemma))
+    if [b0_ne] != 1*[None]:feats.append('b0_ne=%s_' % (b0_ne))
+    if [b0_w] != 1*[None]:feats.append('b0_w=%s_' % (b0_w))
+    if [b0_lemma] != 1*[None]:feats.append('b0_lemma=%s_' % (b0_lemma))
+    if [b0_t] != 1*[None]:feats.append('b0_t=%s_' % (b0_t))
+    if [b0_dl] != 1*[None]:feats.append('b0_dl=%s_' % (b0_dl))
+    if [b0_len] != 1*[None]:feats.append('b0_len=%s_' % (b0_len))
+    if [b0_reph] != 1*[None]:feats.append('b0_reph=%s_' % (b0_reph))
+    if [a0_ne] != 1*[None]:feats.append('a0_ne=%s_' % (a0_ne))
+    if [a0_w] != 1*[None]:feats.append('a0_w=%s_' % (a0_w))
+    if [a0_lemma] != 1*[None]:feats.append('a0_lemma=%s_' % (a0_lemma))
+    if [a0_t] != 1*[None]:feats.append('a0_t=%s_' % (a0_t))
+    if [a0_dl] != 1*[None]:feats.append('a0_dl=%s_' % (a0_dl))
+    if [s0_p1_ne] != 1*[None]:feats.append('s0_p1_ne=%s_' % (s0_p1_ne))
+    if [s0_p1_w] != 1*[None]:feats.append('s0_p1_w=%s_' % (s0_p1_w))
+    if [s0_p1_lemma] != 1*[None]:feats.append('s0_p1_lemma=%s_' % (s0_p1_lemma))
+    if [s0_p1_t] != 1*[None]:feats.append('s0_p1_t=%s_' % (s0_p1_t))
+    if [s0_p1_dl] != 1*[None]:feats.append('s0_p1_dl=%s_' % (s0_p1_dl))
+    if [b0_pathpwd,b0_lemma,s0_lemma] != 3*[None]:feats.append('b0_pathpwd&b0_lemma&s0_lemma=%s_%s_%s_' % (b0_pathpwd,b0_lemma,s0_lemma))
+    if [b0_apathpwd,a0_lemma,b0_lemma] != 3*[None]:feats.append('b0_apathpwd&a0_lemma&b0_lemma=%s_%s_%s_' % (b0_apathpwd,a0_lemma,b0_lemma))
+    if [b0_pathpwd] != 1*[None]:feats.append('b0_pathpwd=%s_' % (b0_pathpwd))
+    if [b0_apathpwd] != 1*[None]:feats.append('b0_apathpwd=%s_' % (b0_apathpwd))
+    if [b0_lemma,b0_rsb_dl] != 2*[None]:feats.append('b0_lemma&b0_rsb_dl=%s_%s_' % (b0_lemma,b0_rsb_dl))
+    if [b0_lemma,b0_nswp] != 2*[None]:feats.append('b0_lemma&b0_nswp=%s_%s_' % (b0_lemma,b0_nswp))
+    if [dist1] != 1*[None]:feats.append('dist1=%s_' % (dist1))
+    if [dist1,b0_pathp] != 2*[None]:feats.append('dist1&b0_pathp=%s_%s_' % (dist1,b0_pathp))
+    if [dist2] != 1*[None]:feats.append('dist2=%s_' % (dist2))
+    if [dist2,b0_apathp] != 2*[None]:feats.append('dist2&b0_apathp=%s_%s_' % (dist2,b0_apathp))
+    if [s0_lemma,b0_t] != 2*[None]:feats.append('s0_lemma&b0_t=%s_%s_' % (s0_lemma,b0_t))
+    if [s0_lemma,b0_dl] != 2*[None]:feats.append('s0_lemma&b0_dl=%s_%s_' % (s0_lemma,b0_dl))
+    if [s0_t,b0_lemma] != 2*[None]:feats.append('s0_t&b0_lemma=%s_%s_' % (s0_t,b0_lemma))
+    if [s0_dl,b0_lemma] != 2*[None]:feats.append('s0_dl&b0_lemma=%s_%s_' % (s0_dl,b0_lemma))
+    if [s0_ne,b0_ne] != 2*[None]:feats.append('s0_ne&b0_ne=%s_%s_' % (s0_ne,b0_ne))
+    if [a0_t,b0_lemma] != 2*[None]:feats.append('a0_t&b0_lemma=%s_%s_' % (a0_t,b0_lemma))
+    if [a0_dl,b0_lemma] != 2*[None]:feats.append('a0_dl&b0_lemma=%s_%s_' % (a0_dl,b0_lemma))
+    if [a0_ne,b0_ne] != 2*[None]:feats.append('a0_ne&b0_ne=%s_%s_' % (a0_ne,b0_ne))
+    if [a0_cpt,b0_cpt] != 2*[None]:feats.append('a0_cpt&b0_cpt=%s_%s_' % (a0_cpt,b0_cpt))
+    if [a0_cpt,b0_cpt,s0_lemma] != 3*[None]:feats.append('a0_cpt&b0_cpt&s0_lemma=%s_%s_%s_' % (a0_cpt,b0_cpt,s0_lemma))
+    if [a0_cpt,b0_ne] != 2*[None]:feats.append('a0_cpt&b0_ne=%s_%s_' % (a0_cpt,b0_ne))
+    return feats
\ No newline at end of file