Source Code for Module TEES.ExampleWriters.BioTextExampleWriter

 1  """ 
 2  Wrapper for all interaction XML example writers 
 3  """ 
 4  import sys, os, types 
 5  thisPath = os.path.dirname(os.path.abspath(__file__)) 
 6  sys.path.append(os.path.abspath(os.path.join(thisPath,".."))) 
 7  import Core.ExampleUtils as ExampleUtils 
 8   
 9  from EntityExampleWriter import EntityExampleWriter 
10  from EdgeExampleWriter import EdgeExampleWriter 
11  from ModifierExampleWriter import ModifierExampleWriter 
12  from PhraseTriggerExampleWriter import PhraseTriggerExampleWriter 
13  #IF LOCAL 
14  from UnmergingExampleWriter import UnmergingExampleWriter 
15  #from UnmergedEdgeExampleWriter import UnmergedEdgeExampleWriter 
16  #from AsymmetricEventExampleWriter import AsymmetricEventExampleWriter 
17  #ENDIF 
18   
19 -class BioTextExampleWriter: 
20      """ 
21      A generic example writer that automatically calls the correct Example Writer 
22      based on the type of the examples. 
23      """ 
24      @classmethod 
25 -    def write(cls, examples, predictions, corpus, outputFile, classSet=None, parse=None, tokenization=None, goldCorpus=None, insertWeights=False): 
26          if type(examples) == types.StringType: 
27              print >> sys.stderr, "Reading examples from", examples 
28              examples = ExampleUtils.readExamples(examples, False) 
29           
30          # This looks a bit strange, but should work with the re-iterable 
31          # generators that readExamples returns 
32          xType = None 
33          for example in examples: 
34              assert example[3].has_key("xtype") 
35              xType = example[3]["xtype"] 
36              break 
37           
38          if xType == "token": 
39              w = EntityExampleWriter() 
40              if insertWeights: 
41                  w.insertWeights = True 
42          elif xType == "edge": 
43              w = EdgeExampleWriter() 
44          elif xType == "task3": 
45              w = ModifierExampleWriter() 
46          elif xType == "entRel": 
47              w = EntityRelationExampleWriter() 
48          elif xType == "phrase": 
49              w = PhraseTriggerExampleWriter() 
50  #IF LOCAL 
51          elif xType == "um": 
52              w = UnmergingExampleWriter() 
53          #elif xType == "ue": 
54          #    w = UnmergedEdgeExampleWriter() 
55          #elif xType == "asym": 
56          #    w = AsymmetricEventExampleWriter() 
57  #ENDIF 
58          else: 
59              assert False, ("Unknown entity type", xType) 
60          return w.writeXML(examples, predictions, corpus, outputFile, classSet, parse, tokenization, goldCorpus=goldCorpus) 
61   
62  if __name__=="__main__": 
63      # Import Psyco if available 
64      try: 
65          import psyco 
66          psyco.full() 
67          print >> sys.stderr, "Found Psyco, using" 
68      except ImportError: 
69          print >> sys.stderr, "Psyco not installed" 
70       
71      from optparse import OptionParser 
72      optparser = OptionParser(usage="%prog [options]\nWrite predicted examples to interaction XML") 
73      optparser.add_option("-e", "--examples", default=None, dest="examples", help="Machine learning example file", metavar="FILE") 
74      optparser.add_option("-p", "--predictions", default=None, dest="predictions", help="Classifier predictions for the example file", metavar="FILE") 
75      optparser.add_option("-i", "--classIds", default=None, dest="classIds", help="Multiclass class Ids") 
76      optparser.add_option("-c", "--corpus", default=None, dest="corpus", help="Interaction XML file for adding examples to", metavar="FILE") 
77      optparser.add_option("-g", "--goldCorpus", default=None, dest="goldCorpus", help="Interaction XML file with gold elements", metavar="FILE") 
78      optparser.add_option("-a", "--parse", default="split-McClosky", dest="parse", help="Parse XML element name") 
79      optparser.add_option("-t", "--tokenization", default="split-McClosky", dest="tokenization", help="Tokenization XML element name") 
80      optparser.add_option("-o", "--output", default=None, dest="output", help="Output file") 
81      optparser.add_option("-w", "--insertWeights", default=False, action="store_true", dest="insertWeights", help="Write weights for gold elements") 
82      #optparser.add_option("-t", "--task", default=1, type="int", dest="task", help="task number") 
83      (options, args) = optparser.parse_args() 
84       
85      assert(options.examples != None) 
86      assert(options.predictions != None) 
87      assert(options.classIds != None) 
88      assert(options.corpus != None) 
89      assert(options.output != None) 
90       
91      BioTextExampleWriter.write(options.examples, options.predictions, options.corpus, options.output, options.classIds, options.parse, options.tokenization, options.goldCorpus, insertWeights = options.insertWeights) 
92