Source Code for Module TEES.ExampleWriters.PhraseTriggerExampleWriter

 1  import sys 
 2  from SentenceExampleWriter import SentenceExampleWriter 
 3  import Utils.InteractionXML.IDUtils as IDUtils 
 4  try: 
 5      import xml.etree.cElementTree as ET 
 6  except ImportError: 
 7      import cElementTree as ET 
 8  import Utils.Range as Range 
 9   
10 -class PhraseTriggerExampleWriter(SentenceExampleWriter): 
11 -    def __init__(self): 
12          self.xType = "phrase" 
13          SentenceExampleWriter.__init__(self) 
14       
15 -    def writeXMLSentence(self, examples, predictionsByExample, sentenceObject, classSet, classIds, goldSentence=None, exampleStyle=None):         
16          self.assertSameSentence(examples) 
17           
18          sentenceElement = sentenceObject.sentence 
19          sentenceId = sentenceElement.get("id") 
20          sentenceText = sentenceElement.get("text") 
21          # detach analyses-element 
22          sentenceAnalysesElement = None 
23          sentenceAnalysesElement = sentenceElement.find("sentenceanalyses") 
24          if sentenceAnalysesElement == None: 
25              sentenceAnalysesElement = sentenceElement.find("analyses") 
26          if sentenceAnalysesElement != None: 
27              sentenceElement.remove(sentenceAnalysesElement) 
28          # remove pairs and interactions 
29          interactions = self.removeChildren(sentenceElement, ["pair", "interaction"]) 
30          # remove entities 
31          newEntityIdCount = IDUtils.getNextFreeId(sentenceElement.findall("entity")) 
32          nonNameEntities = self.removeNonNameEntities(sentenceElement) 
33               
34          # add new pairs 
35          for example in examples: 
36              prediction = predictionsByExample[example[0]] 
37              entityElement = ET.Element("entity") 
38              entityElement.attrib["isName"] = "False" 
39              headToken = example[3]["t"] 
40              for token in sentenceObject.tokens: 
41                  if token.get("id") == headToken: 
42                      headToken = token 
43                      break 
44              entityElement.set("charOffset", example[3]["charOffset"])  
45              entityElement.set("headOffset", headToken.get("charOffset")) 
46              entityElement.set("phraseType", example[3]["ptype"]) 
47              entOffset = Range.charOffsetToSingleTuple(example[3]["charOffset"]) 
48              entityElement.set("text", sentenceText[entOffset[0]:entOffset[1]+1]) 
49              entityElement.set("id", sentenceId + ".e" + str(newEntityIdCount)) 
50              self.setElementType(entityElement, prediction, classSet, classIds) 
51              newEntityIdCount += 1 
52              sentenceElement.append(entityElement) 
53   
54          # re-attach the analyses-element 
55          if sentenceAnalysesElement != None: 
56              sentenceElement.append(sentenceAnalysesElement) 
57