Source Code for Module dataprocessor.output.wmt11tabwriter

 1  """ 
 2  @author: Eleftherios Avramidis 
 3  """ 
 4   
 5  from sentence.dataset import DataSet 
 6  import codecs 
 7   
 8 -class Wmt11TabWriter(object): 
 9      """ 
10      classdocs 
11      """ 
12       
13   
14 -    def __init__(self, data=None ,name="dfki", testset="testset", att_name="rank"): 
15          """ 
16          Constructor 
17          """ 
18          self.metric_name = name 
19          self.testset = testset 
20          self.att_name = att_name 
21           
22          if isinstance(data, list): 
23              self.content = None 
24              self.convert_to_tab(data) 
25          elif isinstance(data, DataSet): 
26              self.content = None 
27              self.convert_to_tab(data.get_parallelsentences()) 
28           
29 -    def convert_to_tab(self, parallelsentences): 
30          """ 
31          Creates an tab for the document an populates that with the (parallel) sentences of the given object. 
32          Resulting tab string gets stored as a variable. 
33          @param parallelsentences: a list of ParallelSentence objects  
34          """ 
35           
36          entries = [] 
37          entries.append("<METRIC NAME>\t<LANG-PAIR>\t<TEST SET>\t<SYSTEM>\t<SEGMENT NUMBER>\t<SEGMENT SCORE>") 
38                   
39          for ps in parallelsentences: 
40              ps_att = ps.get_attributes() 
41              if ps_att.get("testset"): 
42                  testset = ps_att["testset"] 
43              else: 
44                  testset = self.testset 
45              for tgt in ps.get_translations(): 
46                  t_att = tgt.get_attributes()                 
47                  entry = "\t".join([self.metric_name, "%s-%s" % (ps_att["langsrc"], ps_att["langtgt"]), testset, t_att["system"], ps_att["id"], t_att[self.att_name]]) 
48                  entries.append(entry)   
49               
50          #entries = sorted (entries, key=lambda entry: entry.split("\t")[4]) 
51          self.content =  "\n".join(entries) 
52           
53 -    def write_to_file(self, filename): 
54          file_object = codecs.open(filename, 'w', 'utf-8') 
55          file_object.write(self.content) 
56          file_object.close()   
57       
58 -    def write_to_file_nobuffer(self, filename, parallelsentences): 
59          file_object = codecs.open(filename, 'w', 'utf-8') 
60   
61          file_object.write("<METRIC NAME>\t<LANG-PAIR>\t<TEST SET>\t<SYSTEM>\t<SEGMENT NUMBER>\t<SEGMENT SCORE>") 
62                   
63          for ps in parallelsentences: 
64              ps_att = ps.get_attributes() 
65              if ps_att.get("testset"): 
66                  testset = ps_att["testset"] 
67              else: 
68                  testset = self.testset 
69              for tgt in ps.get_translations(): 
70                  t_att = tgt.get_attributes()                 
71                  entry = "\t".join([self.metric_name, "%s-%s" % (ps_att["langsrc"], ps_att["langtgt"]), testset, t_att["system"], ps_att["id"], t_att[self.att_name]]) 
72                  entry = "%s\n" % entry 
73                  file_object.write(entry)   
74               
75          #entries = sorted (entries, key=lambda entry: entry.split("\t")[4]) 
76